This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2020-06-24
Channels
- # ask-the-speaker-track-1 (423)
- # ask-the-speaker-track-2 (320)
- # ask-the-speaker-track-3 (405)
- # ask-the-speaker-track-4 (68)
- # bof-arch-engineering-ops (6)
- # bof-covid-19-lessons (6)
- # bof-cust-biz-tech-divide (10)
- # bof-leadership-culture-learning (8)
- # bof-next-gen-ops (17)
- # bof-overcoming-old-wow (1)
- # bof-project-to-product (5)
- # bof-sec-audit-compliance-grc (13)
- # bof-transformation-journeys (10)
- # bof-working-with-data (24)
- # discussion-connect-february (1276)
- # games (69)
- # games-self-tracker (3)
- # grc (1)
- # happy-hour (189)
- # help (166)
- # hiring (12)
- # lean-coffee (20)
- # networking (5)
- # project-to-product (4)
- # snack-club (42)
- # sponsors (85)
- # summit-info (274)
- # summit-stories (3)
- # xpo-datadog (2)
- # xpo-digitalai-accelerates-software-delivery (14)
- # xpo-github-for-enterprises (14)
- # xpo-gitlab-the-one-devops-platform (14)
- # xpo-itrevolution (6)
- # xpo-launchdarkly (1)
- # xpo-pagerduty-always-on (1)
- # xpo-planview-tasktop (7)
- # xpo-slack-does-devops (8)
- # xpo-snyk (2)
Really nice examples to show the impact of โdominant architectureโ. Not a term Iโd heard before. Reminds me strongly of Kuhnian paradigm.
I bet @genek is finally learning the mental model by teaching it to others.
I bet @genek is finally learning the mental model by teaching it to others.
Learning is a great Ponzi scheme where you have to become a teacher to get to the master level.
I forgot the slide that mentioned Team Topologies by @matthew and @me1208, which I think is a brilliant book that describes these concepts in fantastic detail.
Lots of great lessons in contrasting the NUMMI plant with other GM plants that only adopted the trappings of the Toyota system without having the culture.
I wish @genek would stop with the book references. My reading list is overflowing
Listening to this it seem like structure is like a consistent Nudge towards a certain kind of behavior.
Listening to this it seem like structure is like a consistent Nudge towards a certain kind of behavior.
It's either a nudge or it ends up being a matter of having to swim upsteam all the time, fighting the current.
I start to wonder whether there is a relationship between war/army and IT organizations, seen al the books refering to US army or historical wars :rolling_on_the_floor_laughing:. Love it!
I start to wonder whether there is a relationship between war/army and IT organizations, seen al the books refering to US army or historical wars :rolling_on_the_floor_laughing:. Love it!
I think it is that war/army provides a documented history of different approaches and outcomes, and then we have IT trying to learn from any where they can.
War metaphors have been common, but in the SRE community there is pushback to "demilitarize" along with a drive to make things work for "normal people" not just heroes. Hero culture is a bad anti-pattern.
I didnโt take @davy.kenisโs comment as about war/military metaphors but rather about taking lessons learned from military literature. (Agree that hero culture is an anti-pattern.)
Correction: itโs not just a thought experiment โย this is what healthcare orgs have done to enable teams to communicate needs. Spookily like Team of Teams, by the way!
Correction: itโs not just a thought experiment โย this is what healthcare orgs have done to enable teams to communicate needs. Spookily like Team of Teams, by the way!
I see recurring characters and themes, for sure. And I'm starting to recognise the theme tune
Any conference behind the scenes is slapstick comedy
Any conference behind the scenes is slapstick comedy
Truth. Did them in the late 1980s when slides were physical. Nothing like having a carousel of slides on the floor ten minutes before a keynote...
I think physical slides made it easier to add a slide while the speaker was already speaking. With a PDF, that's harder to do.
True - although it did take a few days to get slides fixed if there were errors!
Oh I wish it still took days. Would make us mindful about them ๐
There is no kidding amounts of stress getting these videos loaded โย Iโm still havenโt turned in one for tomorrow, causing @erin (and others) no small amounts of stress over the last week. ๐ญ ๐
Have you guys seen Squads from InVision? Best current agile teams exploration lโve seen of late
Have you guys seen Squads from InVision? Best current agile teams exploration lโve seen of late
LinkedIn messaged you with what-little- l have @james839
@genek101 you were talking about feedback, do you think feedforward can be also useful in operating projects with success?
Feedback is analyzing something from the past, see how you can do better, but sometimes we need to look forward and have the end goal in our radar and that's why I wanted to see your perspective.
I think theyโre drip feeding it for good reason as huge production value but interviewed Atlassian, AirBnB, obvs all Spotify brains etc.
Super excited to have Janet and Patrick represent the Nationwide Building Society journey so far and whatโs next ๐
โweโre experiencing the lowest interest rates in literally the last 3,000 years.โ (holy cow)
...yeah - donโt hold back on the threat levels, right?!
The current shock going through the system is a weird enabling factor in some ways (I work in the travel industry - lots of creative problem solving happening...) I shows how much more lean you probably could be if you're forced :thinking_face:
โ20 years as an auditorโ โย I still have my IPPF red book from the Institute of Internal Auditors!
@philipp.boeschen650 exactly that in financial services too -- Nationwide (where I work) has brought forward a lot of change we weren't expecting in all kinds of unexpected ways
Yes Philipp, weโve found the constraints have indeed enabled us to speed up.
Interesting to hear auditing as gathering information in a short burst of work. Reminds me of incident analysis. @janet.chapman, can you say more about how Agile helped you in your work?
Go ahead, name another Chief Internal Auditor in a bank who gets Agile and is called a Mission Leader - lโll wait ๐
Go ahead, name another Chief Internal Auditor in a bank who gets Agile and is called a Mission Leader - lโll wait ๐
what about outside of a bank :thinking_face: ? cant think of anyone either
Sally Clarke, ex-Chief Internal Auditor, Barclays
@janet.chapman is such a refreshing leader to have in a senior and impactful role here at Nationwide - I enjoy hugely our conversations and the role she plays in serving and enabling ๐ฅฐ
@janet.chapman I was trying to find my IIA IPPF red book to impress you. Although, I have no idea if itโs still red. ๐
โweโve been around for 136 years; we have every intention for being around for the next 136 years.โ
@jtf - the role of trust and honest โAgile Conversationsโ (honestly) play a role! Heads-up for @eleanorjtaylor talk later for more detail on Audit story
I recommend you listen to @eleanor.taylor later this morning as she led the work adopting Agile in Internal Audit at Nationwide. We worked closely with a coach who had developed adaptations of Agile for non-technology contexts based on scrum,
^^^ If I understand correctly, it was the audit group that blazed the agile trail in the org, using it to improve flow of work within the audit functionโฆ which is pretty amazing. Did I get this right, @janet.chapman ?
^^^ If I understand correctly, it was the audit group that blazed the agile trail in the org, using it to improve flow of work within the audit functionโฆ which is pretty amazing. Did I get this right, @janet.chapman ?
I think we had a range of parallel โpioneersโ - the Audit journey, coupled with the Digital journey and Change journey have all played a part in reaching the Mission model and WoW support function we are pivoting to
OMG. Thatโsโฆ amazingโฆ. ๐ ๐ ๐ Come on, tell me another one. Next, youโll tell me that theyโre no longer called โwork papers?โ
We had a lot of things on our side, the scale of our team (100 people), the high degree of co-location and an enthusiastic team willing to experiment and learn.
an enthusiastic team willing to experiment and learn. < that doesnโt happen by accident
an enthusiastic team willing to experiment and learn. < that doesnโt happen by accident
It does. 0.1% of the time ๐
It turns out, 0.1% of teams in organizations with <100 teams is not enough.
Agree ๐ฏ % - the role of Janet as a leader and Ellie and Mark as a Coach were pivotal in creating the environment and experimenting safely
How do "member missions" relate to product teams? The latter tend to be more stable and "infinite" as long as the prodct thrieves and is in active development.
@patrick.eltridge really nice insight about assembling all the people needed for the work in one team for a Member Mission. How long does a mission typically last for? Curious if that means you are not pursuing long-lasting teams.
@patrick.eltridge really nice insight about assembling all the people needed for the work in one team for a Member Mission. How long does a mission typically last for? Curious if that means you are not pursuing long-lasting teams.
Missions are enduring - shift from functional alignment to market alignment in value-streams ๐
Last audit joke: From my opening remarks yesterday: Corey Quinnโs tweet is quite funny. It is true, however, that people did cheer when we had auditors from each of the Big Four on stage last year. ๐
It should have taken 9 months; instead they did it in 4 + 4 + 4 days. What a great story of lightning fast problem solving!
โthat took about 4 daysโ < reminds me of @steve773โs discussion of amazing recovery at Toyota at from a fire at a parts supplier. same message: having people prepared ahead of time, aligned on solving the problem, can give amazing results
Amazing stories from @patrick.eltridge on how suddenly things are getting done in โfour daysโ, seemingly impossible months ago!
โWe made the most of the Covid crisis to up the paceโ if everyone could say that instead of justify paralysis
โWe made the most of the Covid crisis to up the paceโ if everyone could say that instead of justify paralysis
That does tell something about the amount of focused work time pre-covid
@genek101 thatโs what they always say now at Nationwide since the Audit team adopted Agile!
@john710 The Missions persist indefinitely. They are definitely long-lived teams.
โAgile is our means, DevOps is our targetโ - great quote.
โanswers not being generated by a โcentral team of experts.โโ ๐
Great message @patrick.eltridge - if you have all the right people in the room, aligned on the most important priority, things are done far faster - and safely - than would otherwise be expected
I have a similar story of lightning fast innovation. On 48 hoursโ over-the-weekend notice that a statewide lockdown order was coming, our local school system managed to distribute Chromebooks and (as needed) Wifi access points to every single child in the district, to enable remote learning.
Nice description of the role of leaders, and the attributes of good leaders.
I am more sceptical. A crisis tends to cut away political crud, no matter the system of organisation. I've seen non 'agile' orgs respond to COVID within the same time frames. I don't honestly think you can use that as an example of systemic cultural change.
I am more sceptical. A crisis tends to cut away political crud, no matter the system of organisation. I've seen non 'agile' orgs respond to COVID within the same time frames. I don't honestly think you can use that as an example of systemic cultural change.
Agree - and yet does show what is the art of the possible if we focus on achieving clear, prioritised, shared organisational goals - it helps with confidence and safety for the organisation to try to improve systemically
IIRC The Goal and The Phoenix Project both start with a similar insight, that a crisis allows people to do exceptional things in exceptional circumstances.
The trick is creating an organization who has that kind of alignment when it isnโt an obvious crisis. I think thatโs the role of mission.
I liked Gene's talk on Steven Spears - no real mention of mission there.
Its fascinating to see how different teams at my organization are recognizing the behaviors they are doing, doing retrospectives, and generating the working agreements/culture to never go back (and also those teams that arenโt)
Say more @nick.jenkins? I took the mission here as the tool used to create alignment. What am I missing?
Oh that's what the NationWide guys are saying but I'm not buying it. I've been in similar orgs, described similarly and the actual culture has been truly awful.
Prob because the mission was lip service.
A common observation is that people are at their best in "Chaos" (see Cynefin). Come together, swarm, experiment, solve the issue quickly. Many orgs are asking themselves post COVID how they can keep the positives from that behaviour without the chaos!
I think alignment is massively difficult to achieve beyond about the arbitrary 150 people limit.
never let a good crisis go to waste
The web of meaning. Mission in this context is not a temporal activity.
Prob because the mission was lip service. < undermining trust this way is a good way to undermine alignment.
@jonathansmart1 how is a long lived value stream manifested? Single, x-functional team, lavishly funded with a product owner?
So the missions at Nationwide have purpose and reflect what members need: so, 'homes and dreams' / 'moments that matter' / 'hassle-free money'
@jeremy.mcgee that's the kind of mission I detest. Hollow words when you're stuck in an endless bureaucratic process, either inside or outside.
each with a small focused central team that feed backlogs in specialist teams elsewhere
@nick.jenkins me too! -- the distinction here is a willingness from COO to cut the red tape
Nested value streams, taking Dunbar number into account (teams, team of teams, team of team of teams)
Nest value streams? I have no idea what that means.
as an insider: there were many more similar stories. Mutual trust and understanding were vital. I think it's the same in many other industries
Anyone capture those characteristics? โadaptive, growth mindset, instinctive coaches, ???โ
Anyone capture those characteristics? โadaptive, growth mindset, instinctive coaches, ???โ
I took a note to go back and capture from the recording. ๐
@nick.jenkins Youโre right Nick - I am using the example internally and here to point to how work can get done safely and much faster with the conditions of having everyone around, and focussing on the most important thing. Itโs a showcase example to the organisation. Itโs far to early to claim systemic cultural change.
โJanet built the most scarily efficient audit department Iโve ever worked with in 30 years. Incredibly effective and visible adoption of agile principles.โ ๐ ๐ ๐ ๐
@nick.jenkins We plan to use the experience of pace and agility we have been able to show over the last few weeks to show the organisation that we can do it. The leadership challenge for Patrick, me and the rest of the leadership team is to make sure we turn our experience in the crisis into permanent culture change.
@nick.jenkins We plan to use the experience of pace and agility we have been able to show over the last few weeks to show the organisation that we can do it. The leadership challenge for Patrick, me and the rest of the leadership team is to make sure we turn our experience in the crisis into permanent culture change.
But how? What's going to change? People can always rise to a challenge, but at what cost? Let's use Steve Spears terminology - what structural changes will you make?
โSenior leaders are responsible for being able to explain why certain decisions are madeโฆ canโt lose that as we empower teams.โ
โSenior leaders are responsible for being able to explain why certain decisions are madeโฆ canโt lose that as we empower teams.โ
I'd be very interested in knowing more about how the UK senior manager regime is being addressed
A term that I now use regularly but I learned a long time ago while testing at NBS was 'appetite for risk' with regards to financial advisors reviewing pensions and other products with members. Often an overlooked dynamic factor that likely has a broad range between smaller startups and larger enterprises but also between different industries offering different types of products.
A term that I now use regularly but I learned a long time ago while testing at NBS was 'appetite for risk' with regards to financial advisors reviewing pensions and other products with members. Often an overlooked dynamic factor that likely has a broad range between smaller startups and larger enterprises but also between different industries offering different types of products.
We use "risk appetite" to describe US Federal agencies' willingness to migrate workloads to cloud: pioneers vs guardians
As a tester I'm also interested in 'appetite for quality' ๐ Many folks expect it should be non-negotiable but that old adage of a cost/time/quality triangle (pick two!) suggests you can deliver something really fast and cheap but the challenge is also to meet all the (functional and non-functional) quality expectations too...
@genek101 Still missing some of yesterday's great keynote sessions in the github repo.
@genek101 Still missing some of yesterday's great keynote sessions in the github repo.
@janet.chapman - Iโm interested in what โInvite over inflictโ means in your WoW principles?
@janet.chapman - Iโm interested in what โInvite over inflictโ means in your WoW principles?
One of our core WoW Principles - relates to a desire to meet colleagues, teams, leaders โwhere they areโ and improving with them in their context and on their own timescales. @jonathansmart1 introduced this and referenced the โdiffusion of innovationโ S-curve to support thinking here...

@karl.marfitt Thatโs right Karl, we seek to adapt controls and policies to fit the risk of the experiment in question. This approach depends on a culture of transparency and collaboration to build trust and safety though...
> Changed the language (to suit the context) and kept the principles Love it, @janet.chapman ๐
> Changed the language (to suit the context) and kept the principles Love it, @janet.chapman ๐
@janet.chapman do you have some examples of the kind of language you pivoted to using? Would be really cool to pinch be inspired by some of your ideas.
@kenny It means bringing people to the party because they want to be there and see why there is value for them in being there (Invite). Making Agile something that โhappenedโ to the team, something that was forced on them (Inflict) wasnโt going to be successful.

The Mortgage Payment Holiday journey was a brilliant example of the ways-of-working, practice and the tech capability we have worked hard to introduce and mature
The Mortgage Payment Holiday journey was a brilliant example of the ways-of-working, practice and the tech capability we have worked hard to introduce and mature
isnt that the DevOps Enterprise Summit tag line?
Yup - check back on last yearโs London talk from @jonathansmart1 too - great summary of BVSSH
@janet.chapman would be interested to discuss the reframing of terminology to overcome agile vs process with audit. We just got our FedRAMP P-ATO and this was a significant barrier that forced us to make certain decisions and modifications to how we operated.
@janet.chapman would be interested to discuss the reframing of terminology to overcome agile vs process with audit. We just got our FedRAMP P-ATO and this was a significant barrier that forced us to make certain decisions and modifications to how we operated.
Would be interested in learning about your journey to getting the FedRAMP @brian.martin
@kenny It means that we are not imposing a standard pattern of business agility, centrally designed and rolled out. Rather, we are starting with individual team, talking about the principles, and supporting teams as they figure it out locally, in the work.
@kenny It means that we are not imposing a standard pattern of business agility, centrally designed and rolled out. Rather, we are starting with individual team, talking about the principles, and supporting teams as they figure it out locally, in the work.
The BVSSH work is honestly so key for us - incredibly fortunate to partner with @jonathansmart1 and a genuinely shameless plug for his upcoming book...just brilliant ๐:star-struck:๐
Can you provide some figure on your deployment frequencies (for non-regulated / regulated cloud / non-cloud parts of your systems) @patrick.eltridge @janet.chapman
Are you going to look at remote teams in conjunction with the office based hires?
Really inspiring stuff @patrick.eltridge and @janet.chapman - looking forward to hearing more
โWe are hiring! We are likely the only financial services org that is opening a new building right now.โ โ @patrick.eltridge If people are interested in these 1,000+ open positions, what should they do? cc @richard.james
โWe are hiring! We are likely the only financial services org that is opening a new building right now.โ โ @patrick.eltridge If people are interested in these 1,000+ open positions, what should they do? cc @richard.james
https://devopsenterprise.slack.com/archives/CB0JGT2NQ/p1592905733125300
TFW a COO says "containerization is going to be really important for us" ๐คฏ
TFW a COO says "containerization is going to be really important for us" ๐คฏ
Yeah - amen to that - brilliant to have a Board sponsor who โgetsโ this stuff on both what and how and why!!!
Got to remember that @patrick.eltridge is a technologist - believer in both how we work and what we work with
Did I hear the time's up alert buzzing in the background right at the end?
(@sam and @patrick.debois256 changing the movie reels in the back of the projection room. ๐
(@sam and @patrick.debois256 changing the movie reels in the back of the projection room. ๐
@brian.martin. We changed terminology to be less of a turn off to our auditors and shaped some of the principles to fit how the works needs to flow in an audit, including levels of review and approval that we needed in place to satisfy regulatory expectations. But what we built was recognisably Agile underneath the words.
Thank you @patrick.eltridge and @janet.chapman โย so wonderful!!!
@patrick.eltridge great insight into the Nationwide story. What's the timeline that you have put around the hiring of the 1000 engineers. when thinking about building the development capabilities will low-code platforms also have a role to play? thanks!
@peter.fassbinder Itโs highly variable at this early stage. But ultimately, in a mature DevOps culture, with significant controls automation, Iโd expect there to be little difference between regulated and non-reg changes. We def have some pockets of regulated change that have release cadences measured in days rather than months.
@peter.fassbinder Itโs highly variable at this early stage. But ultimately, in a mature DevOps culture, with significant controls automation, Iโd expect there to be little difference between regulated and non-reg changes. We def have some pockets of regulated change that have release cadences measured in days rather than months.
Reasonably it depends more on the tech and team maturity in balancing speed and safety than on whether more/less highly-regulated - there is some brilliant work in-flight to embed Intelligent Control into our CI and Jira tooling to support/accelerate here too
very excited to be part of a great team and exceptionally ethically minded organisation. devops / cloud / containerisation is a great space to be in at this time with Nationwide
Thanks nationwide for this interesting talk, underlying the importance of top level management sponsorship to move organizations faster.
@jthomson Weโve been hiring for the past 9 months - we have on boarded about 400 great people so far, and still looking. @richard.james - Iโm not sure we do much low-code yet - is that right?
@jthomson Weโve been hiring for the past 9 months - we have on boarded about 400 great people so far, and still looking. @richard.james - Iโm not sure we do much low-code yet - is that right?
Weโve got a pocket of low-code in the investments space but, no, not (yet at least) a major area for us.
On engineering - there is a major growth in cloud-native, open-source practitioners for Platform and Stream-aligned teams
fabulous insight thanks - would love to learn more in terms of the cloud native strategy
I was wondering how Army and Navy fits with the coopetition/OSS model (Team of Teams and Dr. Spears) They are conflict-focused organisations where cooperation sits at the different level would like to hear your thoughs @genek101
โamong bar staff, we started arguing about what corkscrew was the best. what really is the value derived from corkscrew? open, pour, deliver, no cork leftโฆ this argument less relevant [thru eyes of customer]โ ๐
Thanks to DOES and Gene for inviting us and for all the great comments and questions. This has been fun! (And check out our jobs #hiring)

Thanks to DOES and Gene for inviting us and for all the great comments and questions. This has been fun! (And check out our jobs #hiring)

We will 110% be back - multi-year journey and emergent every quarter - challenge from @genek101 was for a CEO to talk...well, weโve had the COO this year and I love a challenge ๐ฅฐ๐ฅฐ๐ฅฐ

โI love doing complex mathematic equations by hand; born too late for this jobโ ๐
Many thanks to @genek101 for the conference and inviting us. Was a hoot!

Many thanks to @genek101 for the conference and inviting us. Was a hoot!

@patrick.eltridge @janet.chapman very awesome story. so happy to hear about the london expansion!
โโฆand we started having arguments of Tableau, PowerBIโฆ so we did both. and it was fun, comparing the tools; but it delayed shipping of info to customer, and cost too much.โ ๐ ๐
โwhat happens when you get senior technologists in a room together; more talk about tools, more politics.โ ๐
^^ nodding furiously - too much tools & tactics, not enough visualization & strategy
Every company does employee satisfaction and forget fundamental things like how people experience work day to day.
@simonw Definitely factoring in remote working in our hiring - weโve learned a LOT from the last three months.
@simonw Definitely factoring in remote working in our hiring - weโve learned a LOT from the last three months.
Shameless plug for one of our former team members new org: http://remote.com - a new startup they've founded to help with remote hiring. We can make an intro if you would like cc @nlomas
We (GitLab) have been all remote since we started. Happy to chat further about how we make it work (1,300 team members now in 68 countries) and introduce you to our Head of Remote.
^ reminds me of early days of CITCON or Devopsdays
@andrea.hausmann @duncan.lawie how did you boil down information/decission from the devops enterprise workgroup towards the lower teams?
@andrea.hausmann @duncan.lawie how did you boil down information/decission from the devops enterprise workgroup towards the lower teams?
Also targetting teams who were interested or making progress, and getting them to show how they were moving forward
it sounds like a jokeโฆ but going into the emergent consequences would actually deserve it.
(@allspaw has some interesting views on Conwayโs Law, it would certainly pep things up)
That's correct @nuwayser ๐ It was a great experience for us as authors
โฆ i just think about the potential outcome of these conversations on toolsโฆ well none of them actually meet our needs, so lets build our ownโฆ 9 months later, we have built a new tool that has some parity, still no value shipped
Yep, in my experience processes are always addressed last when moving to DevOps (unfortunately)
In follow up to the excellent talk by @patrick.eltridge @janet.chapman cc: @richard.james https://www.nationwide-jobs.co.uk/tech-talent/
@janet.chapman @patrick.eltridge Excellent presentation and thank you for sharing and being so open.
This talk is really resonating so much with me, I have a couple of unusual (non IT!) tools related stories that definitely shaped my early career before IT. The best thing I can say is that it's a poor worker that blames their tools, especially when very high quality tools are quickly available at very low cost!
This talk is really resonating so much with me, I have a couple of unusual (non IT!) tools related stories that definitely shaped my early career before IT. The best thing I can say is that it's a poor worker that blames their tools, especially when very high quality tools are quickly available at very low cost!
True - and we are trying to focus on outcome rather than tools, but the balance between letting all have their own tools and being able to share skills and methods
Mhm as soon as people identify with the companies mission and vision then putting that success before personal growth becomes more common I think :thinking_face:
Fascinating insight about how inherently political technology, or tied to ego, choices are.
Fascinating insight about how inherently political technology, or tied to ego, choices are.
We had tech sitting in our company for ages because someone was so proud of the deal they stuck and forced everyone to use it ๐
Oh, youโre going to LOVE the next presentation from @scott.prugh!!! ๐
Golfcourse software gives me severe anxiety ๐
100% we've had many discussions around tools with phrases like "Ugh tool A is rubbish I prefer tool B", I sometimes think it's a learning curve issue, i.e "I've used tool B for years, why should I have to learn tool A"
Is the person using his/her ego to serve the project or the opposite, is usually quite eye opening on intrinsic motivations....
โWe looked at our metrics, and we had either chosen those metrics because they made us look good, or because we had learned how to make ourselves look good by using those metricsโย :rolling_on_the_floor_laughing:
โThe discussion about tools became more and more politicalโ - yup. Iโm all about moving away from โcultureโ and โorganisationโ talk though and towards the only actionable bit - changing behaviours to increase Psychological Safety at every level- obsess with Impression Management and stopping it in every team
โThe discussion about tools became more and more politicalโ - yup. Iโm all about moving away from โcultureโ and โorganisationโ talk though and towards the only actionable bit - changing behaviours to increase Psychological Safety at every level- obsess with Impression Management and stopping it in every team
Our pyschological safety is typically great at team level, poor at senior manager level
Tech people live in constant fear of their hard earned experience being made irrelevant
Tech people live in constant fear of their hard earned experience being made irrelevant
Exactly - and we tie a lot of that โexperienceโ to tools.
When in reality weโve learned how to use tools (generally) to solve problems.
so, actually, the ability to solve problems (regardless of used tools) is the added value of an individual
Oh, I remember the first change resistance when having to move from a safe VMWare virtualization to this suspicious Hyper-V when changing jobs. Well, it's not that different.
Also, a great ride where instead of replacing with new flashy tools, we just maintained the existing Powershell deployment scripts and got down to releasing at will.
Check out this โconfessionโ about the effect of being a SME on on the โoldโ tools (start about 2:00 in): https://www.youtube.com/watch?v=ocWc6blSj4A
It is easy to install cool tools and a cool toolchain, and announce that you are 'doing agile' or you 'have DevOps' now.
It is easy to install cool tools and a cool toolchain, and announce that you are 'doing agile' or you 'have DevOps' now.
How many times have I heard this... the real horror is when the leadership do this, fold their arms and STOP as they are now DONE ๐ก
The context setting power of "Accelerate" has been so important in our org too. We've handed out 6 cases to senior leadership to establish a common language.
The context setting power of "Accelerate" has been so important in our org too. We've handed out 6 cases to senior leadership to establish a common language.
I gave a reading guide, starting with the metrics chapter and then a bit of "choose your own adventure" as to which bits to read next
I put a flyer in it with why they should read and and information about the DevOps Dojo. Then I followed up to see which copies had dust on them. Our COO's copy is worn and has tones of book marks, so she get as much help as we can give her.
We also built a DORA dashboard and said, "these are the metrics" If they want different metrics, they can just build their own. Control the information and you control the narrative. Don't be evil.
@duncan.lawie Great use of the Accelerate metrics. @nicolefv Research is a fantastic set of work.
@duncan.lawie Great use of the Accelerate metrics. @nicolefv Research is a fantastic set of work.
It really has helped - but we also need to provide enough supporting measures so people can see the steps along the way
in my organisation I see huge differences between groups and locations some are top notch, others laggards the latter typically own what is perceived more business-critical set of applications
in my organisation I see huge differences between groups and locations some are top notch, others laggards the latter typically own what is perceived more business-critical set of applications
I also see DevOps lag with teams supporting large, mission critical, monolithic applications.
The 24 capabilities in Accelerate are a great baseline to align teams and leaders on.
The link to the metrics chapter in Safari is part of our core learning path.
My company (Siili Solutions) has a lot of Clojure people and combined with the Love Letter to Clojure from @genek almost had me spend some time learning Clojure. However, the parenthesis triggered a PTSD (PostTraumatic Scheme Disorder)
Where I do think Lisp can be awesome, it has too much "syntactic vinegar"
โMake sure people are on a trajectory of improvement rather than making unreasonable comparisons between organizations.โ < This is one theme in our talk later today as well.
we should do a poll in this slack channel, how devops are you?? Love the question!!
we should do a poll in this slack channel, how devops are you?? Love the question!!
You'd be surprised on how many companies are actually measuring their agility as a percentage...45% last month, 55% this month...
if they are just looking at metrics in a view of months they already falling behide.
โthe easiest way to learn what a person thinks is to ask themโ < +1000

โthe easiest way to learn what a person thinks is to ask themโ < +1000

though to be fair it is even easier to guess and then just assume youโre rightโฆ ๐
I discovered "The power of the questions mark" via Agile Conversations ๐ ๐ Pratical example: dropped a mail this morning to our CEO/CTO, containing only questions... just need to ensure they understand it's to understand better, because there could be expectation that only answers should be given to them,
sounds like you have an opportunity to be transparent on why youโre being curious. ๐
When typing the above, realised I missed that transparency part and assumed that would be clear without mentioning, now I "feared" my purpose was not clear... guess a small follow-up mail will be send out soon ๐ Thanks for helping me crawling up after hitting the ground again ๐
โA leader of a 1,000 person org said, โI donโt need surveys; I already know what my employees think.โโ ๐
โA leader of a 1,000 person org said, โI donโt need surveys; I already know what my employees think.โโ ๐
we hear โwe already asked them something last yearโ every day
Yes, common toolchain! So important for training amd metrics observability
Yes, common toolchain! So important for training amd metrics observability
Also on top it reduces friction between teams, if you have a common stack, helping other people with their bugs gets A LOT easier
it is a piece of the puzzle we have this BUT some tools do not integrate (and we are adding an integration tool to the mix for this purpose)
Yesterday I spoke a bit about how we're approaching this at William Hill too, so its good to here we're not alone in this approach
Yes - our toolchain is Inner Source, to enable more integration - extension to additional languages, for example
but I see a huge difference in teams' maturity regardless of toolset in use
We offer flexibility as long as: we get pipeline metrics, security and compliance gates, and capitalization data.
We also made a few core choice and then left things like CI pipelines deliberately as a suggestion - we offer a central option but are not forcing people into it
But we created APIs so people can integrate their CI pipelines into our reporting metrics
@sim.l.williams we have a questionnaire that teams have to fill in periodically we are just beginning pulling out data from tools at the enterprise level
Would love to hear more about how you are getting data from the tools to support it @gvian
long story our major issue is that backlog and bug are in Rally which has no simple integration with Azure DevOps where we have source code so we are adding an integration tool like ConnectAll, OpsHub, or TaskTop to sync data and start getting an end2end picture
โmoney doesnโt grow on trees, not even at banks; and nor do resources. in game theory terms, goal is to maximize benefit for most people, not for one personโ
With central toolchains this vital, are central teams operating and caring for them as products equally vital?
With central toolchains this vital, are central teams operating and caring for them as products equally vital?
ie "that part of the built environment that is only noticed when it stops working" (Paul Graham Raven)
โ35% of our funding ensures that we be customer-focused; we user inner-sourcing; anyone can put their extentions into toolchain, as long as they donโt break things.โ
IBM has company wide toolchains as well. Make it easy for teams to adopt. Also give them the option to choose something different.
IBM has company wide toolchains as well. Make it easy for teams to adopt. Also give them the option to choose something different.
Carrot-not-stick is such a good meta-pattern
@andrea.hausmann - You used the word โgrowโ when suggesting teams can add what they want to the toolchain in order to โgrowโ better. What does grow mean in that context?
@andrea.hausmann - You used the word โgrowโ when suggesting teams can add what they want to the toolchain in order to โgrowโ better. What does grow mean in that context?
Improve the toolchain that it really helps them for their own need --> this should help them progress or work as good as possible wherever that is within their own business.
Gotcha - I like it. Not just develop software, grow in their own business. Thanks!
โpractitioner to practitioner to share expertise across; org; and communities of practice to broaden their understanding of what other people do; earlier in career, we encourage broadening understanding of what other people do.โ
Re: politics: This is a great book for understanding the impact of politics on innovation. It's written by a physicist turned business leader - who brings a physicist pov to understanding organizational culture.ย https://www.amazon.com/Loonshots-Nurture-Diseases-Transform-Industries/dp/1250185963
@duncan.lawie good point that the lead time for changes metric is easier to measure in teams who already have a fairly streamlined deployment pipeline. Thatโs one thing I love about the 4 key Accelerate metrics: they encourage the right behaviour (moving to a more streamlined pipeline in this case), unlike most metrics . .
The purpose of IT is almost always supporting something, if we're building IT to build IT systems rethink :thinking_face:
The purpose of IT is almost always supporting something, if we're building IT to build IT systems rethink :thinking_face:
Tech-for-tech is not necessarily a bad thing
The opposite is not good - the shoeless cobblersโ children is so common
Amazing tech for customers but internally things strung together with sticking plasters
People are the most challenge of the three pillars for DevOps. Tools you learn, you experiment. The process you implement, define, refine. But people have a lot of background, believes and culture that is a challenge to change.
People are the most challenge of the three pillars for DevOps. Tools you learn, you experiment. The process you implement, define, refine. But people have a lot of background, believes and culture that is a challenge to change.
The same thing can work with people! You can experiment with your conversations and end up changing the culture. Normally though people arenโt aware of the need to improve their conversations. They assume that a better culture means other people behave differently.
Oh I agree! I think people is actually the point of highest leverage. It is a shame then that people donโt study people skills in the same way that they study tools and process. (Disclosure: Iโm one of the co-authors of Agile Conversations and this is a big topic of me!)
I'm new in DevOps world, but I see how important is Agile, so I would say that you hit the right subject with the book. I'll look after it.
https://www.conversationaltransformation.com/agile-conversation-book/
@jtf this way of thinking hadn't come to my mind before, but suddenly it feels natural. I tend to try to adapt my way of communicating with others to try to get to a common ground and understanding.
Thanks for sharing about the book, sounds very interesting. Is there a PNL root on the ideas? I'll sure read (well, probably listen to) it!
The academic origins are in the work of Chris Argyris: https://infed.org/mobi/chris-argyris-theories-of-action-double-loop-learning-and-organizational-learning/
Oh, sorry! I meant NPL, Neuro-linguistic programming. PNL is in Portuguese! My bad, it might be waking up early for couple of days and brain not fully shifted to English! lol
Not connected to NPL as far as I know. Though from what Iโve seen lots of communication approaches share the fundamental insight that you need to actually pay attention to what the other person is saying. ๐
โas a plus it helps them to improve their empathyโ yes! Growing empathy should be a huge effort in the People Practice we all should toil on
@cncook001 right, sounds to me some difficult balance to manage a such big toolchain and flexibility to the users so they can get quickly new features and/or new tools on demand
I'm sorry. I'm going to sound like the black hat yet, but does no one else have a problem with a centralised tool chain? What happened to let the practitioners choose their own tools?
I'm sorry. I'm going to sound like the black hat yet, but does no one else have a problem with a centralised tool chain? What happened to let the practitioners choose their own tools?
Accelerate has a good paragraph on this - let me see if I can find it
Its a balance. Let teams choose their tools, but not all tools.
itโs also about the dynamics that get created among the teams. Team Topologies talks a lot about that.
For us, many are happy to accept the tools given. The "top performers" are free to integrate
Yeah we also see that the bulk of teans just want something working - teans who really want to control it themselves are much rarer
And good point earlier that any centrally provided toolchains should invite usage, due to adding value
This matters because cross organizational context understanding and reuse gives you lots of scale. Also make sure that centralized toolchains are self service
this reminds me of themes that Dan Cundiff talks about in the 2018 Vegas talk on Crowdsourcing Technology governance: https://www.youtube.com/watch?v=4XYIs-6ZMjk specifically Danโs section
I consider the life cycle of in-house toolchain a Shu-Ha-Ri journey for in-house users.
I don't have a problem with centralised tools, assuming that they are easy to use. Centralised platforms can provide a degree of economy and efficiencies but only if they meet the user needs. There is a lot to get get right and IMHO the most important is funding. get the funding model wrong then you will have a terrible platform. In terms of security, it's the deployment pipeline that is the risk, not the rest of the tooling. A centralised set of tools will give you much better visibility of what is going BUT and it's a very large BUT, devs will eb better if they choose their own tools. How to resolve the paradox? Have a few tools as "enterprise" (Source code, Artifacts, Knowledge, Work tracking) and then have default recommendations for each tooling capability. Let the devs chose and the best ones will grow, when they get to a certain size, then can be considered central. It's all balance, the yin and yang. Let 600 teams pick every tool of their own and it's going to be an expensive bill to maintain.
We have been using Inner Source as a method to allow teams submit their own changes
100%. At Barclays we had the concept of โcranesโ and โspannersโ
Workmen donโt bring their own cranes to the building site
Given the pace of change in tools and languages I can't see how any one can advocate for one tool set with a straight face. Thank you @rohrersm for the references. I'm with Nicole on this.
But they can absolutely bring their own spanners
And definition of cranes and spanners in a dev chain ?
It depends :) As little as you can get away with to achieve your goals!
@matthew.cobby not necessarily expensive. If you have SaaS tool chains, open source etc. If everything becomes an enterprise licence then yes you are going to struggle to manage multiple licences better than your vendors.
As an example - we were (when I left) getting to the stage where we did want common metrics about lead time - but it was OK to use whatever tool you wanted as long as things were published.
yeah, so don't mandate a tool - mandate a principle. like visible work.
Yes, and a plug-in point to get the metric out of the tool
So as long as the tool publishes specific metrics, thatโs ok
Certain things like SonarQube we really did mandate across the board
Though purely to visualise rather than a mandated gate
And some things you don't want to build yourself. Fair enough. But even a single choice there could lead to tunnel vision.
I am a big believer that you have to treat the โrecommendedโ tools like a product. Assume your customers (the engineers) have a choice. Understand why they are choosing different and make the โrecommendedโ tool the easy and obvious choice. As leaders, we can never mandate as it will create resentment. Leaders need to also know when to evolve what is recommended.
Yes, itโs balance.
The nirvana is the Amazon-like one: give people balanced outcome metrics (better value sooner safer happier for example) and let them decide how to hit their KPIs/OKRs/etc.
Also with centralised tools you tend to get centralised teams - to control/support the tool - and you have a bottleneck in your process again.
Have a minimum viable tool platform. Another example is open source scanning - you probably donโt want to pay for 20 different licenses and let one team get BlackDuck, another WhiteSource and yet another Sonatype
@nick.jenkins not always. I built the first version of the enterprise tool chain for the bank and I worked very hard to make it all self service. Didnโt get to all of the capabilities but enough to get access and to get moving. Treat it as a product then let users self service. I will acknowledge that in many companies itโs jobs forever for tool admins
If you let everyone store their source code where ever they want, you are taking a huge risk on your years of investment
This is an alignment/autonomy discussion
The trick is making the first 60% want to move, the rest will more easily follow after that - making people want to move takes excactly the mentioned product view of your platform/service to make it attractive and solve the right problems ๐
If the teams within their own autonomy choose to go to the central chain - then you're winning, that takes a lot of product effort though
Iโve been brewing a blog post on this for a while, time to write? I learnt a lot on the journey from running tools on virtual box to an outsourced data centre to taking the first material workload to cloud for a bank in Australia.
@matthew.cobby a model we've used for source is to let people have their own (secure) repos and then replicate to a central copy, with git centric repos it's easy. And it still gives the teams autonomy and speed. Nothing is insoluble.
@nick.jenkins youโve mandated git there :)
Weโre having problems getting people off TFSVC :-/
(And in Barclays we had that, plus SVN, plus Visual Source Safe...)
Thereโs nearly always a minimum viable <something> you need to mandate. The trick is to get it as minimal as possible.
Tool variation, different forms to provide, alignment vs autonomy, it is a highly context-based choice.
I remember in Project to Product, this issue is mentioned, the author thinks the more important to think about is how you can keep the information distributed in those tools connected as value stream map.
Yes, thatโs specifically around work management tooling - where @mikโs company provides Tasktop Hub for connecting tooling and Viz for extracting valuable metrics from those tools.
And thatโs a small subset of the whole dev/ops concept-to-cash tooling space
Connections are also important in the rest of the dev/ops tooling space too - but theyโre more and different connections than the VSM-related ones.
(E.g. connections between the CMDB and code quality tooling to enable ecosystem roll up & dice & slice of metrics)
some related discussion in this thread. https://devopsenterprise.slack.com/archives/C015DUDNPLM/p1592942509302700
There is a management overhead to any tool chain. Why would you want that multiple times? What value does it bring? Focus the teams on creating your own IP
To use the cranes and hammers analogy - try telling the next plumber that comes to your house he has to use your spanners.
But on a similar axis you could imagine it being the question of should all the plumbers from the same company have different tools? :thinking_face:
But I don't know enough about plumbing to be of much use in that discussions to be fair ๐
The difference is that in that analogy you would be asking every household to have their own tool chain. I donโt think any case the analogy works that well.
yeah the more I think about it the more confusing it gets ๐คท
I think the way the โcentral tooling/toolchainโ is managed is as important as the tools themselves. In many cases the toolchain is managed in a โwaterfallโ way without the required feedback loops and experimentation.
As a provider of in-house toolchain, I consider the centralized tool chain to meet needs of part of the in-house users. Never consider it to be the only choice or standard. You are lucky to have those keep using it, but just be happy when they leave yours to find more proper ones for their own needs.
Related question: do you all have a standardised ops/service management tool chain - e.g. ServiceNow? How much of that is federated / up to the teams?
@duncan.lawie How does credit Suisse organise COP in the time of Covid? I am assuming that the teams are remote.
@duncan.lawie How does credit Suisse organise COP in the time of Covid? I am assuming that the teams are remote.
Many of our teams are cross-region or cross-site. We are setting up sessions in "America-friendly" and "APAC- friendly" sessions
And the well established "site-based" CoPs are still meeting, as they have the personal relationships in place to keep working
So London and Zurich can pick, whilst Singapore and NY have something that works for thme
Oh okay. It's something we were doing but it lost interest. Have you got any tips on making it an inviting space?
Work in Progress - using the results of Satisfaction Survey to find topics and presenters
Fantastic to hear this, my superiors are still convinced that context is everything, engineering is the easy bit that can be outsourced
@andrea.hausmann @duncan.lawie congrats on getting that amazing testimonial from someone who matters!!!
Fantastic! Hiring engineers with no obstacles - that's sounds like a miracle
Not quite Chapter - โspotify modelโ Chapter is formalised, CoP is less formal and โlaw of two feetโ (i.e. you can choose to leave)
But similar horizontal grouping of skills or practices
We use the terms slightly differently. Guild is something you belong to based on your skills. Community of Practice is open to anyone who has an interest in the topic. Guilds are focused on improving skills and improving the life of their members. Communities are very organic and a shared experience.
e,g. a DevOps CoP is a community of people who want to deliver software better. Not a meeting of people with DevOps in their title.
Context is key! All these types structures might be useful :)
(and avoid copying + pasting the โspotify modelโ which even spotify donโt use) https://youtu.be/cwbiSCgiZNA
Communities of Interest are underused too
CoPs are (ideally) for actual practitioners to learn from each other
Whereas CoIs are for newbs too or for those who just want to watch rather than participate
I hadn't used the CoP/CoI before. We have broadcast events (talks, slides, demos) that are often more CoI in this context. I then use Workshops, Deep Dives, Lean Coffees for peer to peer learning
@andrea.hausmann @duncan.lawie Thank you so much for this amazing presentation, resonated amazingly! ๐
I liked the question about how you discuss about happiness in organization @duncan.lawie!
@andrea.hausmann @duncan.lawie good presentation! Core vs context is important. How do you define engineer when hiring (easier hiring compared to others)?
@andrea.hausmann @duncan.lawie good presentation! Core vs context is important. How do you define engineer when hiring (easier hiring compared to others)?
yes exactly. So we built in obstacles to overcome if you are not hiring an engineer. ore approvals, more reasoning, etc.
Yes I like that idea! But how do you define an engineer? It can be narrow lika one with a master in computer science or wide like any engineering degree etc.
Can you share some more thoughts on how you have started to measure the DevOps transformation and how you set the priorties
Thank you @andrea.hausmann @duncan.lawie!
be a good person, shape a better world
Vielen Dank @andrea.hausmann and @duncan.lawie for this talk which resonated with some similarities here...
It represents the consumption of our APIS from front end applications
@scott.prugh i love the focus on the business-oriented metrics, supported by the IT-oriented metrics.
PS: I am endlessly fascinated that despite being SVP of Software Engineering, @scott.prugh still self-identifies as primarily an architect. (which given my fascinating with structure/dynamics, maybe itโs not a surprise that leaders need a strong architecture skills.)
Great presentation - thanksย @andrea.hausmannย @duncan.lawie for the insights
What a graph โย 33% in customer growth; but # of transactions increased by 800%. โWe could not increase our costs by 800%, and thus the story of how we broke that curve.โ
Context of the presentation: ones where you spend more time in audit and compliance with them vs. using them for engineering; and 500% price rises come renewal
Large monolithic database vendor default sales strategy? ๐
Also watch for PE firms acquiring tools and middleware
I feel so young... goes off and googles SLBOS ....
I feel so young... goes off and googles SLBOS ....
My googling didn't help me much apart from now I know SLBOS is also a bracket... https://www.bhphotovideo.com/c/product/734643-REG/Chief_SLBOS_SLBOS_Universal_Interface_Bracket.html
@duncan.lawie Have you seen this? https://engineering.atspotify.com/2014/09/16/squad-health-check-model/ We use it quarterly to checkin with squads.
@andrea.hausmann @duncan.lawie That was an excellent presentation, congratulations. The point where you mentioned the importance of finding common ground, for instance, the metrics of Accelerate, really resonates. It makes totally sense!
met @scott.prugh yesterday and he has awesome โbeen thereโ wisdom
Funny. I just read Will's book 2 weeks ago. I have been meaning to reach out to him as we share some of the same thoughts.
It's actually a plaque for our SLBOS port. However...several leaders at CSG, including Scott and me, are currently reading that
@kboth_does You will see more details about the plaque comming up
โGolf course software; software selected higher up in the org, and then inflicted on the teams. Look how happy the CIO and salesperson is!โ ๐
so, "designing work so that problems can be seen (@steve773)" before hurtling down Microservices Avenue...
โDonโt do this to your people, let them pick their own toolsโ ๐:skin-tone-5:
I've heard of Golf Course Features before... I guess it can be an entire system too
I love how @scott.prugh says โlow TPS density.โ Which I think is polite for โruns like a pig.โ ๐
OMG Golf Course Software. YES. How many developers have groaned under the pain of using tools that โdonโt require much developmentโ. Great, instead we have to become SMEs in configuring some arcane tool.
OMG Golf Course Software. YES. How many developers have groaned under the pain of using tools that โdonโt require much developmentโ. Great, instead we have to become SMEs in configuring some arcane tool.
My first job was just like this. A small org where I was the only dev. The deputy MD would stealth choose all tech and expect it to get implemented the following day.
I was a Maximo developer for a while. I had to learn how to specify UIs entirely in XMLโฆ and all database changes had to be specified as a change script. WHY???
If all of our products are built on the same golf course software as the underlying database, what would some first steps I can take as an engineer to help us get off of that software?
If all of our products are built on the same golf course software as the underlying database, what would some first steps I can take as an engineer to help us get off of that software?
@scott.prugh You might've missed this with all the other excitement in the channel.
If the middleware(A) is the target pick a good opensource alternative(B). Pick a transaction. Test it. Record results. Port the code. Test that new one. Now figure out how to deploy. This is where you need some way to switch from A to B. For that, I recommend a software router. You can write one and deploy it in front of the old middleware. Have the middleware switch from A to B for a subset of trans and further validate fidelity.
@jonathansmart1 Look for hostile vendors swooping in to buy testing tools that enterprises have built all their test strategies on and then outsourced!!
@jonathansmart1 Look for hostile vendors swooping in to buy testing tools that enterprises have built all their test strategies on and then outsourced!!
100% what has been happening with testing for years.
like the traps which are mentioned. I do experienced some another trap I would like to share: Technology invention trap: fancy cool super modernized technology but no idea how we can use it or how the organisation benefits from it
@scott.prugh did you expect the legacy systems to change and pick up the new api/event layer or did you shim that data into the existing message bus/integration patterns?
Itโs what weโve committed to here as well here at Saxo. If youโve got a sizeable legacy, and youโre not prepared to stop the world to modernise, 5+ years is what it takes.
Wouldn't say the vendors are hostile, it's just their business model. Either you change their mind and bring them on-aboard, or hire more employees of your own.
"The Gardening Chanel with James Prigioni" quote about the best time to plant a tree <3
OMG. This is true. That scene in Unicorn Project came straight out of a story from @scott.prugh!!! And I had no idea that 8U servers were THAT heavy! (itโs been 25 years since Iโve touched one, and I donโt think I ever had to move one.)
OMG. This is true. That scene in Unicorn Project came straight out of a story from @scott.prugh!!! And I had no idea that 8U servers were THAT heavy! (itโs been 25 years since Iโve touched one, and I donโt think I ever had to move one.)
i've never carried servers that big but once i've carried about 160 UPS's all by miself :))
Technical excellence takes time and 5+yrs is clearly the kind of commitment we need to make it a reality. In the race for speed, most of the time we attract tech debt and when not handled in a timely fashion, we become part of legacy. ๐
Technical excellence takes time and 5+yrs is clearly the kind of commitment we need to make it a reality. In the race for speed, most of the time we attract tech debt and when not handled in a timely fashion, we become part of legacy. ๐
The challenge is that the typical lifetime of a CIO is << 5 years
Go as high as you can for commitment.
For us itโs at CEO level and weโre trying to get it even higher at that, to the board
(see leadership discussions yesterday, needs both :))
I am really looking forward to that @steve.robert.barr
One of my first jobs was scrapping old Mass Spectrometers, nothing like taking a large hammer and smashing up old machines.
(Strange memory. I remember moving a 19" Hitachi monitor in 1987 up some stairs. It was almost 100 lbs, and I almost dropped and destroyed it. I canโt imagine how heavy some of the 8U things are, @scott.prugh!)
(Strange memory. I remember moving a 19" Hitachi monitor in 1987 up some stairs. It was almost 100 lbs, and I almost dropped and destroyed it. I canโt imagine how heavy some of the 8U things are, @scott.prugh!)
I was told they are close to 300LBs. Like a refrigerator.
I used to have a double 8" floppy drive that 3 people had hard time to carry as it was 1m3 of metal.
This was the newer โlighterโ model. I cannot find a picture of the one I had. http://www.ghb.cz/storage/201701020243_2016_11_17-28_GHB_vystava_pocitacu_24.jpg
A related tree quote that I love is, 'a society grows great when it's old men plant trees whose shade they will never know'. Forgive the gender bias. It's from ancient Greece.
Nice presentation, todays keynotes are killing it. this might be the tech, but the biggest thing that stood out so far was I never thought about using an ALB like that with flags in the header for routing. that gives me an idea using the AWS ingress controller with a kubernetes stack on doing deployments, and just switching tags on deployments when we are ready for routing. im loving all these stories :thumbsup:
Nice presentation, todays keynotes are killing it. this might be the tech, but the biggest thing that stood out so far was I never thought about using an ALB like that with flags in the header for routing. that gives me an idea using the AWS ingress controller with a kubernetes stack on doing deployments, and just switching tags on deployments when we are ready for routing. im loving all these stories :thumbsup:
Feature switches are a key architectural construct to lower the cost of activation from deploy
โyouโll note that the assembler and Java are isomorphically equivalent.โ ๐
โyouโll note that the assembler and Java are isomorphically equivalent.โ ๐
"We built a cross compiler" -> indeed, this is a great example of a home built investment to modernize. Love it.
I love this talk because it shows what engineering greatness looks like, and the impact to business is difficult to overstate!
I love this talk because it shows what engineering greatness looks like, and the impact to business is difficult to overstate!
he makes it sounds so simple and easyโฆ. when this took a great vision, architecture, and ruthless rigor to make happen. its inspiring
Adding 'reserved for future use' feature-switches everywhere now seems like a sensible approach for any as-yet-unknown potential future strangulations. Although also wondering how much wasted data there might be with network protocols that include unused reserved bits when uncompressed.
Adding 'reserved for future use' feature-switches everywhere now seems like a sensible approach for any as-yet-unknown potential future strangulations. Although also wondering how much wasted data there might be with network protocols that include unused reserved bits when uncompressed.
You have to be careful about what happens in the "unused" branches and what might happen if your switching controller fails.
@scott.prugh: @jessicam just said to me, โI love how hard you laugh watching these talks, even though youโve seen them before.โ I told her, โItโs because these talks are SO GOOD!โ cc @jeff.gallimore (Weโre on a Zoom call all morning long during the plenary session.)
Love how many of these metrics tie very neatly into the whole principle of evolutionary architecture!
Love how many of these metrics tie very neatly into the whole principle of evolutionary architecture!
Just discussed this with my team and we will implement most of these measures in our new greenfield project
It speaks to my (lack of) experience when working on similar things on my teams. Great talk
Loving the level of tech chops that @scott.prugh and his team demonstrate - great role model
Itโs important to share stories of long-time employees doing amazing innovation work too.
@scott.prugh - thanks for sharing your mainframe transformation approach and journey - really helpful!
โno version control, no unit tests, no functional tests. no horizontal scaling.โ
โno version control, no unit tests, no functional tests. no horizontal scaling.โ
Exactly like riding a sport bike on the back roads of Arkansas in the rain naked.
Kudos for showing the pics of the teams that achieved all these things
@scott.prugh what do you mean with end to end functional testing? How do you define end to end?
Ah!!! The entirety of this speech from @steve.robert.barr was actually in an early draft of Unicorn Project!!! (It was with much anguish that I had to cut it to hit word count target.) ๐ญ (Steve, Iโm not sure if you knew that. I actually listened to this video many times, manually transcribing it!)
Ah!!! The entirety of this speech from @steve.robert.barr was actually in an early draft of Unicorn Project!!! (It was with much anguish that I had to cut it to hit word count target.) ๐ญ (Steve, Iโm not sure if you knew that. I actually listened to this video many times, manually transcribing it!)
@scott.prugh where can I get better info on MF testing and CI tooling?
@scott.prugh where can I get better info on MF testing and CI tooling?
We had a string API layer in front of the mainframe. A lot of the test automation was done in Gherkin and SpecFlow
@steve.robert.barr Hip Hip Hooray!
@steve.robert.barr Hip hip hooray!!! Kudos to your entire team. So amazing. Many scenes in Unicorn Project was based on seeing that so many years ago!!!
HHH! No wonder they run fast
@scott.prugh did I miss it, how long did this "mainframe to Java" migration take?
I'm getting a fan of these "forcing functions", as discussed during bof yesterday with @scott.prugh and @jonathansmart1
I'm getting a fan of these "forcing functions", as discussed during bof yesterday with @scott.prugh and @jonathansmart1
Some mentionned: โข Frequency driven โข WIP driven (minimize WIP) - batch size of ONE โข ROD
Two "enabling constraints" (1) Time (2) Limit work in progress (and move to pull based system)
(Holy cow, @scott.prugh โย THANK YOU for being up to take questions during your talk. Itโs 4am in Central Time Zone!!! I totally forgot about that!)
(Holy cow, @scott.prugh โย THANK YOU for being up to take questions during your talk. Itโs 4am in Central Time Zone!!! I totally forgot about that!)
4 words, HUGE impact "batch size of one"
This talk will get moving many in the industry. Legacy is not scary anymore: it can be modernized.

Wow. Great keynote! Such great stories - so much to take away! Thank you @scott.prugh ๐๐
๐thanks @scott.prugh I particularly liked the honesty about vendor relations
Oh, feature tetris. My old nemesis.
@scott.prugh I am amazed by your story! Congrats! This presentation was the best so far
Maslach Burnout Inventory (MBI): 15 minutes, $15 USD, major wake-up call
@genek101 that was exactly how I felt last year when we decided to shutdown our company
@genek101 that was exactly how I felt last year when we decided to shutdown our company
๐ (Amazingly difficult to find an emoji to express how I feel.) Congratulations on finding your path forward, Patrickโฆ
chilling story. It is surprising how many founders get to a point where they feel powerless.
Yes. Yes, it is. I hear it a lot. As leaders, we need to be super sensitive to it in our teams โ and in ourselves.
@scott.prugh Great Presentation, really like the way you mapped the KPI's of Accelerate in your team
really great to hear this story of someone having the courage to have this needed conversation
Scary. Was just saying a minute ago the one job lโd always reapply for in our company is Product Owner - anything else CEO, founder, etc lโd happily have others do as lโve burnt in them before - much harder to keep the โwhyโ in those
Thank you for sharing your personal story and experience @genek101
Thanks for reading this to us, and thank you or sharing your personal experience. @genek101. Very thought provoking.
Awesome story @genek101 Thank you for sharing and adding your personal experience!!
"getting people to see what I see" - how many of us can relate to that!
I have a semi-functional pattern for fighting feature tetris with individual teams: 1. Invest in backlog refinement activities to raise vision horizon from toe tips to actual future. This way you affect the plans before they are dictated. 2. Fix WIP to max to your real bottleneck to become a trustworthy partner to business by really delivering what you promise. 3. Fix dailies to address the team's biggest problem instead of individual round of reporting to refocus every day. This works... sometimes. What's your pattern?
I have a semi-functional pattern for fighting feature tetris with individual teams: 1. Invest in backlog refinement activities to raise vision horizon from toe tips to actual future. This way you affect the plans before they are dictated. 2. Fix WIP to max to your real bottleneck to become a trustworthy partner to business by really delivering what you promise. 3. Fix dailies to address the team's biggest problem instead of individual round of reporting to refocus every day. This works... sometimes. What's your pattern?
When the PO or business or whoever takes a step back and "empowers" the team, the team has to enter that space willingly or it will be undone. I remember one very specific moment when the PO showed vulnerability in a situation like that after the three step pattern had been executed for the team and their business peers think of it as the crown jewel in their tribe now.
"You'll have to back me up on this from now on"
this is very similar to what we talk to people. This is how we address the first point. 1. Have backlog grooming by the leads to set priorities, and add details. Then review with the team during backlog refinement for the team to understand what needs to be accomplished, more details, and hopefully improve the estimation based on the breakdown (but it is still a challenge).
Oh, I am very specific not to have it by the leads alone because it leads to less ownership
A rotation of who prepares the next feature, however, achieves the same with wider ownership.
So, a "learning together" event with a rotating subteam as much as necessary and then a weekly workshop with the entire team to sync with everybody
One other thing I am yet to find the team to try out would be mob programming some of the key areas before there is enough feasible tasks for everybody
In my experience it works very well with newly formed teams. About 3 years ago people got shuffled around, and a team with very different backgrounds came together. Mob programming set the bar to understand what we were doing without judgment. It got dissolved because 5 people working independently deliver more than 5 people working on one thing. This was one of my learning experiences for gathering metrics. @pgibbs1587 Any other things you want to add?
100% agree Mob programming works very well and levels up the team. Metrics are the key and this type of programming looking at Integration , Deploy freq and leadtime arenโt the only data to look at. You have to look at the team and realize that everyone on the team actually becomes smarter about the product and process and gains more ownership of that product. There may be enough feasible tasks for everybody to work alone that would keep people away from mob programming but the knowledge the team will learn about the whole product and system by mobbing outweighs that by alot. In my experience.
I see that anything architecturally significant is good to be mobbed. Going back to the high and low context idea, that part of the code should be low context and high quality in order to let everyone work with it efficiently far into the future.
Welcome to Q&A @samgu! (And at 238am his time) Rockstar!
@genek101 that fearlessness of outcome you stated when we talked on the phone is something I tell people is so true. It takes a certain level of accepting the consequences when you are driving change. It also takes resilience and the ability to be encouraged by the smallest motion forward. You need to be a wildly optimistic, pragmatic cynic.
As a martial artist I learned that crisis means danger and opportunity and have always loved it
@bryan.finster โwildly optimistic, pragmatic cynicโ yesโฆ
Quarantine feels good to an introvert, who always wanted to work from home full time ๐
Quarantine feels good to an introvert, who always wanted to work from home full time ๐
@bryan.finster the good news is that courage is trainable - we can pull others into fearlessly speaking up too
@bryan.finster the good news is that courage is trainable - we can pull others into fearlessly speaking up too
And leaders need to create safety for people top speak up. This is very key.
And it counts that we are religiously empathic and forensic about the types of fear too -I.e. impression management- not just tell people to have balls
Hard but a great feeling and makes it easier for me to sleep at night
This is why l always return to Googleโs Aristotle findings - Psychological safety (courage to speak up and not impression manage) is first but you have to have Clarity and Structure and Purpose too
Yes. A clear message of a path forward, even if they disagree, is far more effective "we need to make things better!"
For me once i got ahold of DevOps in my brain and with hard facts and metrics. Fear left me and purpose drove me. And these events reassures me that the thought process on my team is correct.
so true @samgu "Leadership crisis" I see in many countries and companies
so true @samgu "Leadership crisis" I see in many countries and companies
So true. Its time to retrospect how we choose leaders, promote and share leadership..
Microsoft Teams has been amazing... COVID forced our global switch-over to be accelerated,
The Teams usage and scaling story has been stunning. That team really responded amazingly.
Didn't that graph say there are more minutes between pushes? So less frequent? Or maybe I misread it?
@samgu re: GitHub pushes, is there a trend to what type of projects are being worked on?
Having a technical talk from the teams team on how they tackled that scale would be super interesting :thinking_face:
We get a lot of questions around 'how do I return to the previous governance state, pre-COVID. How do I reel back the leash'...
So what do you all think about the time to merge pull requests going down? Is that all good, or are we going to be cursing the shortcuts six months from now? Or both?
Iโve seen PRs get merged that never would have passed code review a couple of months ago. So there WILL be negative consequences. Some technical debt is building up there (for example, disabled tests).
Iโm a true believer in getting value out the door - just pointing out that there will be costs later. ๐
One of the advantages of getting the value out the door is that (given that nothing is certain and that therefore everything is an experiment), itโs only when itโs out the door that you can really understand how much (if at all) the code will actually be used. Doesnโt mean that we should just ship anything, but thereโs definitely a balance and many developers end up over-optimising prematurely.
Working for a 100% remote company has certainly been one of those life-changing moments for me; now I've done it, I'd be very surprised if I ever go back to non-remote work.
โemployees post-pandemic want to work from home 40% more; employees interested at half that rate.โ
GitLab is such an interesting case for ages on how 100% WFH can just be amazing if pulled of right
GitLab is such an interesting case for ages on how 100% WFH can just be amazing if pulled of right
We hired a Head of Remote last year - he's produced some great content https://about.gitlab.com/company/culture/all-remote/guide/ is a good starting point
If you want to talk to us about 100% remote working and how we do it, we're always keen! Our way of working and handbook is something we're incredibly proud of! #xpo-gitlab-complete-devops
Oh that is an interesting conversation ! The only downside with you guys is there is too much content to follow sometimes ๐
We've tried to distil down what we produce. Having everything public can make it noisy
yeah but it's so valuable because it distils for different target audiences already
@genek101 Thank you very much for sharing that impressing and emotional confession before. Will it be available for re-reading?
Same here, working 100% from home, even before covid. The time saved from commuting can improve life quality
@samgu are you also going to cover conversations between employers and employees about locality changes?
โMIT published research paper: WFH in non-high-tech firms had better financial returns; surprising, that it was those firms, not just high tech firmsโ (!!)
So timely, one of the few bits remaining in my book is the chapter on the post-pandemic work and l was dreading looking for the stats. Owe @samgu a virtual beer
And the saved time from commuting also allow us to push more code! ๐
I mean, just from my (personal) perspective; as a new dad I get to spend more time with my wife and daughter. I get to be there for the "firsts" no matter when they happen, she's only across the hall!
So many high quality signals here and so little noise, it's still somewhat overwhelming and a bit intimidating ๐
https://www.dropbox.com/sh/oamszbiycv9xiyj/AAD96ss9uk920DPe2F54UzX9a/Day%202%20-%20June%2024?dl=0&subfolder_nav_tracking=1 when they are ready
First speaker I can whole-heartedly agree with.
I cant see WFH not being the new way. I think this exercise gives the data that most companies were to scared to experiment with and after looking at this there are too any benefits not to.
Kill the zombies! < great message :dagger_knife:
Kill the zombies! < great message :dagger_knife:
@me1342 there is an academic conference we're hosting early August on this .
@me1342 there is an academic conference we're hosting early August on this .
Bloomsbury wants the manuscript way before that sadly but l can sneak an interview in this chapter if you have time
still laughing because of this: "who drove your company digital transformation: 1)CEO, 2)CTO, 3)COVID-19"
Yes!! Tackle your technical debt! Don't be a victim!
I did WFH for six years 2003-2009. The only tools that really changed: Around 2010 Google Docs commoditized massive multiplayer text editing. Around 2020 Miro and such commoditized MMO Canvas. Otherwise... it is pretty much the same.
I did WFH for six years 2003-2009. The only tools that really changed: Around 2010 Google Docs commoditized massive multiplayer text editing. Around 2020 Miro and such commoditized MMO Canvas. Otherwise... it is pretty much the same.
More stable than Etherpad too ๐
100 days of Scrum of Scrum notes in the first version of that thing crashed the server.
Word365 is reaching the level of Google Docs any day now ๐
Haven't used it in a few years now, but it was definitely starting to get there ~2016
Still failing to sync changes
WFH - there is a great thread on twitter, was linked somewhere on here yesterday (sorry can't remember by who) https://twitter.com/sytses/status/1264341436138270720 - argues hybrid is hard and remote will get blamed, and rolled back
WFH - there is a great thread on twitter, was linked somewhere on here yesterday (sorry can't remember by who) https://twitter.com/sytses/status/1264341436138270720 - argues hybrid is hard and remote will get blamed, and rolled back
Sid is the CEO of GitLab and the architect of our WFH culture. But he's 100% right, hybrid approaches tend to have drawbacks, because some communication & decision making inevitably isn't documented, this will hurt you, longer term.
Say no more!! I recall @dominica saying: "No is a full sentence!!"
Get clean and stay clean on tech debt. Love that. Is there a 12 step process?
Shifting right as well as left ๐ฏ
@philipday I know the thread...I don't believe that necessarily. If you work to support remote first, hybrid is entirely reasonable.
"don't be dependent on humans to tell you how machines are doing" -- that's fantastic
We also created a Remote Playbook for anyone interested to discover more on WFH culture https://about.gitlab.com/resources/ebook-remote-playbook/
We would argue that work we have done in incident response has increased resilience in our people. This helped us adjust to covid. We even ran a release deferral with incident command smoothly
If I had told our CEO that we should become a remote first company in January he would have send me to psychologist to examine my brain. Last month he was reading exactly this remote playbook from GitLab.
If I had told our CEO that we should become a remote first company in January he would have send me to psychologist to examine my brain. Last month he was reading exactly this remote playbook from GitLab.
Unfortenately, my company is doing everything to "restore as before". It's insane.
I love the IBM work from home pledge. Very human-focused: https://www.linkedin.com/pulse/i-pledge-support-my-fellow-ibmers-working-from-home-during-krishna/
So good, @samgu โย @pdmoore will give an awesome talk tomorrow about Zone Management tomorrow, about how to do more Core, do less Context.
as a leader it requires quite a lot to keep engagement and team spirit up, when people are isolated at home.
@scott.prugh do you guys measure Flexibility and Resilience separately? We do. Happy to expound why if you liked
Interestingly here in Denmark, we were in many ways lucky - lockdown was short & weโre all now back in the office as if nothing changed. But maybe we miss out on having long enough away to rethink working patterns.
๐ช:skin-tone-2: @samgu
*Warren Buffett*ย once said, "The difference betweenย *successful people*ย andย *really successful people*ย is thatย *really successful people*ย say no to almost everything."
awseome
On the topic of WFH, we've been using Topaasia for online retrospectives. Your truly also created an "Effective Remote Work" deck in the game based on 17 years of remote work in various kinds of environments. https://topaasia.com/en/
On the topic of WFH, we've been using Topaasia for online retrospectives. Your truly also created an "Effective Remote Work" deck in the game based on 17 years of remote work in various kinds of environments. https://topaasia.com/en/
The Topaasia game is greatly influenced by [Dialogue and the Art of Thinking Together](https://www.amazon.com/Dialogue-Thinking-Together-William-Isaacs/dp/0385479999). I saw an immediate connection with Agile Conversations. ping @jtf
Isaacs is a colleague of organizational learning guru Peter Senge <well thereโs a pretty direct connection
Thanks to Covid, I heard the founder of Topaasia finally finished reading the Fifth Discipline ๐
And I may have accidentally name dropped your book to him
What's the best way to build the line of communication for that packet of information? ๐
โฆstand by, allโฆ. @jeff.gallimore closing talk before break is comingโฆ. cc @patrick.debois256
โฆstand by, allโฆ. @jeff.gallimore closing talk before break is comingโฆ. cc @patrick.debois256
looks like it never quite played - BTW, really nice job on the framework and execution
Great talk @samgu really loved some of these phrases, great points too. Looking forward to this break out now!
If we miss the โDevOps Confessionsโ talks, will they be available in the library at some point?
@genek101: going way back to your opening remarks on Structure do you have a top 3 / top 5 reading list to understand the research in the area?
@genek101: going way back to your opening remarks on Structure do you have a top 3 / top 5 reading list to understand the research in the area?
Can you do me a favor? Can you DM or email @annp (<mailto:annp@itrevolution.com|annp@itrevolution.com>) and letโs get some time together in July? There are times when Iโm embarassed Iโve spent weeks talking about MIT Beer Game with Steve Spear. Either weโre onto something, or weโre two people who shouldnโt hang out together, because we make each other dumber. Hahaha Looking forward to head session!!!!
@jtf @genek101 It hasnโt helped that weโve turned the beer game into a beer game. Shots each time we say beer game. So, itโs entirely possible that everything weโve said has been gibberishโฆ
@nicole.bryan @cdeardo #ask-the-speaker-track-4 thanks for a good presentation! What is the difference between light and dark bars in the charts? Like dark and light green (feature) in flow load.
This is @steve773 hanging out in Slack today if anyone wants to chat. Cheers.
@steve773 @jtf this book seems to be in a similar vein as some of the discussions yesterday (just stumbled across it while following Twitter threads): https://www.amazon.com/Power-Experiments-Decision-Making-Data-Driven-ebook/dp/B084P2SD7Y
@steve773 @jtf this book seems to be in a similar vein as some of the discussions yesterday (just stumbled across it while following Twitter threads): https://www.amazon.com/Power-Experiments-Decision-Making-Data-Driven-ebook/dp/B084P2SD7Y
Another one for the list! The fallout from this conference is going to last a long longer than the conferenceโฆ ๐
So glad youโre here, @david627!! Looking forward to your talk! cc @jessica.reif @jessicam
@david627 @genek101 Hey David. Good morning. Looking forward to seeing the preso. Cheers. Steve
Highlight of my day! Really looking forward to it @david627
Do you mean on the DOES stage or by any other book about teams :white_frowning_face:๐
https://www.amazon.com/One-Mission-Leaders-Build-Teams/dp/0735211353 is pretty good...
@david627 when we started our journey in 2014, our SVP referenced your book constantly. Really formational to our thought processes.
@david627 when we started our journey in 2014, our SVP referenced your book constantly. Really formational to our thought processes.
Likewise, the exec committee at my previous place of work all read it โ๏ธ
Thanks @bryan.finster would love to hear more about your journey incorporating some of these lessons at Walmart!
Thanks @bryan.finster would love to hear more about your journey incorporating some of these lessons at Walmart!
@david627 you mentioned the philosophy that rivalry between organizations gets better outcomesโฆ was the thinking that competition brings out the best in them?
@jeff.gallimore Exactly. The competition between units for Jobs would drive us to constantly improve. That was the thinking pre 9/11

@jeff.gallimore Exactly. The competition between units for Jobs would drive us to constantly improve. That was the thinking pre 9/11

โฆand it also reinforced stovepipes and hindered collaboration, right? because for me to win, i have to be better than YOU.
To @jeff.gallimore and @david627 points. My team vs your team can be productive if itโs in the spirit of friendly rivalry and if (thereโs a big and if) and if our work is not interdepedent.
However, if our work is interdependent, that sort of metric is terribly corrosive. Imagine if on a basketball team, everyone got paid by individual points scored rather than some element of games and trophies won? Jack Welsch ran such a zero sum system at GE. Fortunately, he had decades of riches from GE Finance to hide that GE everything else wasnโt terribly competitive. Raises the question, if our leading business schools have a forced curve how does that reconcile with creating leaders able to facilitate collaboration and shared gain? I was once subject to such a grading system. Easy for me as an educator. Didnโt matter what I did. I just had to distribute I IIs and IIIs to the ratio, and I hit my numbers. Far harder to teach where A=mastered the material, B=gained proficiency, and C=showed familiarity. Iโm now on the hook. One, Iโve got to pick the right material for the audience that mastery is possible (why would I want to teach where the goal is โfamiliarityโ). Two, Iโve got tow rok on my teaching because who doesnโt want to earn an A (well, some donโt but you get my point), and three, Iโve got to be adaptive to alter who I teach how so my teaching matches their learning. In a forced curve system, II and III means you didnโt try hard enough to beat your teammate. With an absolute scale, (assuming motivation all around), B means I didnโt do my job.
Founded in the year 1635, predating the invention of paper currency. You can imagine how highly evolved the bureaucracy was there โย evolved over centuries! ๐
@jonathansmart1 @bryan.finster - had they read the Projects too?
@jonathansmart1 @bryan.finster - had they read the Projects too?
@genek101โs ๐
The SVP was pushing Phoenix Project hard in 2014. I'm waiting for Unicorn to hit Safari, but many are reading it.
I think one of the most underestimated things in the Knowledge Economy was that the single worker in the factory has now a much better education. During Taylors time that was not the case.
PS: Iโm incredibly jealous that @jonathansmart1 has a glowing quote from David Snowden on his book cover. ๐
One of my fave quotes from @jonathansmart1 new book โSooner Safer Happierโ was the fact that forty languages were spoken on the Ford assembly line 100 years ago โ showing how little collaboration was actually required. did I get that right? cc @steve773
One of my fave quotes from @jonathansmart1 new book โSooner Safer Happierโ was the fact that forty languages were spoken on the Ford assembly line 100 years ago โ showing how little collaboration was actually required. did I get that right? cc @steve773
It didn't matter, as people didn't need to speak to each other, doing reptitive, specialised tasks
However, their batch size and quality process was horrendous. https://medium.com/itrevolution/5-minute-devops-organizing-for-failure-6eec67b86f82
@rene.lippert so correct. The democratization of education in the middle of the 20th century was a significant driver for the knowledge econmy
@david627 I love the use of cynefin to discuss how people and orgs are stuck thinking in the wrong quadrant...
First rule of the cynefin fight club is you never tell the customer about cynefin. The second rule is, when someone asks you how to make accurate predictions... THEN show the cynefin"
First rule of the cynefin fight club is you never tell the customer about cynefin. The second rule is, when someone asks you how to make accurate predictions... THEN show the cynefin"
Was having that conversation last night about estimates. ๐.
Ironically I have used the cynefin analogy comparing ITIL to agile and devops processes
I am not sure if a lot of DevOps is in the complex domain. Most of it is just assigning stupid repetitions to computers.
Sure, there are the complex parts but especially the tooling part is moving complicated things into obvious space.
I find that it really works with orgs. Especially Head of HR, Head of Procurement, Head of Finance who have been told by someone that they need to be 'agile'. The domains of work conversation is a relief
Right approach for the type of work. Ordered and knowable or unordered and emergent
I have found that Cynefin is often a confuser too ๐
Like metrics. Only to be shared with those who know how to use it.
That's like Catch 22. The top management should be able to use cynefin but they don't and can't be told...
Can't afford a 5-year ease ๐
@scott.prugh i was talking to @steve773 about this recently. the advocated tactics TOTALLY change depending on where you believe you are.
@scott.prugh i was talking to @steve773 about this recently. the advocated tactics TOTALLY change depending on where you believe you are.
the big idea from that conversation was whether you could THINK your way into the right answer or EXPERIMENT (act) your way into the right answer. Very few problems these days in the former category.
If experiment is the right tactic, then that seems to align into the complex
domain. Of course that also matches with considering people as non-linear first order components
(h/t @jtf)
And I argued that modern cloud based saas is more like "warfare". There are emergent threats: cyber, competitors. Rapid change. Dangerous operating environment.
โshared consciousness is the first thing that is compromisedโ --> as a side effect of structure love that phrase: โdecision making space gets pulled upโ โย that a great phrase for thatโฆ phenomenon. ๐
theme from the book: push authority to information, not information to authority.
@david627 I want to present this to my organization at Kessel Run, I think it will really resonate!!
@david627 I want to present this to my organization at Kessel Run, I think it will really resonate!!
We built teams around problems as well. Pipelines, testing, etc. Those teams were ephemeral, but impactful to the org.
I like this as a leadership mantra: Eyes on, hands off
โthe structure [before/after] remained unchangedโ (i.e., Navy SEALs still reported thru Dept of Navy, Army Rangers thru Dept of Army, etc..). still functional orientation, but teams matrixed into mission teams? @david627
Ah, thatโs the phrase. โeyes on, hands offโ vs. โdecision space being being pulled upโ
Iโm doing the mental mathโฆ 15,000 people times a 90 minute meeting every day.
Iโm doing the mental mathโฆ 15,000 people times a 90 minute meeting every day.
I like to use $100/hr as a rough estimate for US professional resources. $2.25 million per meeting, each day. Dang.
They were going to spend that money anywayโฆ the question is how to get value for the money spent.
I thought this was fascinating: this horse-trading between leaders allowed better allocation of resources (vs. centralized planning).
I thought this was fascinating: this horse-trading between leaders allowed better allocation of resources (vs. centralized planning).
This reminds me of David J Andersonโs description of Kanbanโs replenishment meetings in his blue book. He describes how replenishment meetings require multiple stakeholders to horse-trade with each other, and that leads to better results than having one person coordinating.
one of the awesomest examples of โradical transparencyโ lโve seen is a leadership team with a public Trello board for their management Kanban - took forever but once they worked out they havenโt lost their halo and no sky collapsed it increased everyoneโs trust
@ann.marie.99 lot of time. But it reduced overall meetings by 20% and the speed and quality of operations went up 10X
@ann.marie.99 lot of time. But it reduced overall meetings by 20% and the speed and quality of operations went up 10X
I think a government is the only entity that could spend that much on a meeting every day. ๐
they werenโt spending the money on the meeting, they were spending it on the results
"Your C-suite is where the idea came from"... umm... nope
"Your C-suite is where the idea came from"... umm... nope
Since smoking cigars in the corner office was banned for health reasons, what are they actually doing?
Being a big cat myself, I see how that can be a full day job
Okay, he didn't say "good idea" came from C-suite. They have ideas, granted.
@david627 in your Navy SEAL days, what people tended to be the most influential, ala that graph you just showed? (specific rank, skillset?)
their version of deploy frequency = raid frequency
their version of deploy frequency = raid frequency
over time rank mattered a lot less. It was how active you were in the network
over time rank mattered a lot less. It was how active you were in the network
were there any surprising characteristics of who tended to be most influential in the network?
Oh, this is very true. Influence and pay grade are loosely coupled.
A very neat aspect that leads to how it is the reign of the impact you provide to the organization. The influence and reputation become more important than skill, but all are more important than title.
The speed and quality was an output of changing our operating rhythm and specifically that daily O&I or Scrum of Scrum.
@steve773 The daily operational 90 min review in TOT reminds me of Capability 3?
"Since it's a Bureaucratic org I can pretty much be passive aggressive and there is nothing you can do about it." -- OUCH! and true... ๐
@david627 do you remember the Flintstone's episode where Fred becomes an executive at the quarry? He was told all he ever had to do was say 3 things. Who's baby is that? What's my line? I'll buy that.
Doesn't it seem that a lot of the transformation ideas, or at least energy ,in this summit community's context originate sort of bottom up?
We're gardeners... @helen.beal was on to something in talking about Kingley Vale
Reminder: Donโt miss an opportunity for live Q&A with David Silverman at 7:00pm London time as a follow-up to his closing talk today. Join the discussion at https://global.gotomeeting.com/join/359896061
I loved that comment: โmaybe the best thing to do is to just separate two people who donโt work well togetherโ
I loved that comment: โmaybe the best thing to do is to just separate two people who donโt work well togetherโ
Not on my team. ๐ I told a manager I was firing one of the team for malpractice. He said I didn't have that authority. I said I revoked his source access, so apparently I did. ๐
"commodity" vs "soft skills" - one of the biggest challenges i've had in a tech community
โself awarenessโ : ability for your team mates to predict what youโll do, even in stressful situations
Ohh boyโฆ whole month salary spent on booksโฆ ๐ STOP GIVING SUCH GOOD ADVICE
<!here> continue the conversation with @david627 https://sched.co/cbkL
So is โself awareโ and โconnectingโ - โEQโ and โempathyโ @david627?
On the topic of military leadership, our :flag-fi: local approach has 3/4 of the corner stones of transformational leadership and the fourth "Idealized Influence" is replaced with "Building Trust"
Thank you so much to @david627 โย he will be available for live Q&A session later today!!!
Hosting Q&A later for those that want a live discussion. Thanks for having me.

The captions were baked into the presentation - ITRev had nothing to do with it
https://blackswanfarming.com/cost-of-delay/ great page. Ran out of oxygen several times telling people exactly this. The Cost of doing nothing is HUGE.
Great talk @david627.That was like a conference in a talk :rolling_on_the_floor_laughing:
A couple of notes I forgot to mentionโฆ My coworker Jabe Bloom is getting his PHD in Transition Design from Carnegie Mellon School of Design.ย A lot of the design ideas in my presentation come from Jabe - https://design.cmu.edu/user/856
@genek101 weโre wondering the title from the red book in your bookshelf. Help us out :rolling_on_the_floor_laughing:
@david627 Great Presentation and thank you for your Service to all of us. Expected nothing less from a FrogMan.
Oh! That book on my shelf is โDynamic Manufacturingโ, recommended to me by @steve773. Some mind-blowing case studies in there.
Oh! That book on my shelf is โDynamic Manufacturingโ, recommended to me by @steve773. Some mind-blowing case studies in there.
โa PhD in transition designโ #goals
I was typing out a question about what do failures look like and then theres a slide of it!
Product = builds things that donโt matterโฆ the reason why i moved from dev to product ๐ฐ
^^ we hired a prodmgr who make the same move - sick of building useless features
I love how @jwillis can get really charged when he gets passionate โค๏ธ
From seeing safety as an absence of negatives to seeing it as the presence of a positive capacity to make things go right. A focus on safety and risk should become a focus on resilience.
From seeing safety as an absence of negatives to seeing it as the presence of a positive capacity to make things go right. A focus on safety and risk should become a focus on resilience.
But to be sureโฆthereโs no law preventing us from doing it again @jwillis, right? ๐
Someday... Hopefully next year we can start the new normal. We are long overdue for a jam session.
So much โamen!โ
i love @ricookmdโs paper on all of this: how complex systems fail. blows my mind every time i read it.
@jeff.gallimore I was so super delighted to see @ricookmd at the @allspaw Q&A session yesterday!!! (Two people with bow ties there! ๐

@jeff.gallimore I was so super delighted to see @ricookmd at the @allspaw Q&A session yesterday!!! (Two people with bow ties there! ๐

@jwillis do you reckon that the reason you accomplish teaming and therefore Psychological Safety with these people (eventually) it is because theyโve been reading you for long enough to be in awe and be โpreloadedโ with trust?
@jwillis do you reckon that the reason you accomplish teaming and therefore Psychological Safety with these people (eventually) it is because theyโve been reading you for long enough to be in awe and be โpreloadedโ with trust?
I suspect he is really good at employing the trust conversation
from Agile Conversations
This talk has got me thinking, is there any books that cover a range of IT / Company issues and how those things happened? i.e we got hacked, this is how it happened, here's the lessons learned.
This talk has got me thinking, is there any books that cover a range of IT / Company issues and how those things happened? i.e we got hacked, this is how it happened, here's the lessons learned.
This idea was the origination of the DevOps confessional two years ago. There is value in talking about failure (it just turns out, that its hard to get your marketing departments to allow that in public)
@ricookmd last night said something to that effect - no-one publishes the truth about incidents
Yeah I imagine so, but there must be a wealth of learning potential wrapped up in these incidents that could truly benefit the community.
Yes - part of the genesis was the happy hour chat at 2017 after @ricookmd @steve778 and sidney dekker talk (https://www.youtube.com/watch?v=CFMJ3V4VakA)
> Thinking on the complexity diagram a bit. It is more unnerving to have a system that has actions and things being done without known connections or processes well known, than having one that is super complex yet known. Wonder if it is akin to "devil you know" versus the "devil you don't"
An 31-year-old EA felt proud of being one at the young age. Didn't respect my congratulations on being a young dinosaur ๐
@jwillis how would you describe the โ90's model of architectureโ?
@jesse.cafarelli That's one of the greatest compliments. Goldratt and Gene have both got that compliment... Thank you.
@jesse.cafarelli That's one of the greatest compliments. Goldratt and Gene have both got that compliment... Thank you.
You really did hit the nail on the head. Trying to figure out how to fit EA's process they have into a more devops, flexible development cycle that does not hinder development efficiency has been a challenge
@jeff.gallimore That's part of the reason for the 5E. It's like Chinese medicine... balance theory...
When trust is high, I think many developers want increased visibility (i.e. open source). Trust is not always high in organisations, of course.
Governance that feels helpful and protective of clarity, structure and stability
Constraints and freedom resonates well with what @schmark talking about bureaucracy.
hint hint... in 2020 you must have a platform strategy for the next decade. Not PaaS but a Platform as an Interface...
@jwillis ahhh, thereโs deming. i was waiting for that from you.
Where can I learn more about Platfrom as an Interface? I dont know if I know how to design that right now.
Where can I learn more about Platfrom as an Interface? I dont know if I know how to design that right now.
Precision in operational definitions is critical to data quality
The Three Economies is a good place to start. The cheat sheet about has Jabe's two blog about 3E
Metrics-Based Transformation. figure out how to measure non-vanity metrics and use that to constantly evolve those metrics and inform decisions
Metrics-Based Transformation. figure out how to measure non-vanity metrics and use that to constantly evolve those metrics and inform decisions
Use devops proxy/vanity metrics to glue/align your software delivery capability to your business outcomes, IMHO
outcomes becomes the key. not just measuring an output. While important, it should inform how well you achieve the outcome, not be the key measure
but I do agree that leveraging what you can currently get easily can lead you to figuring out what those metrics of outcomes should be
Turns out one can kill it after ToT :woman-shrugging:
@jwillis good point about getting management to learn Taguchi. Learn, change, grow, get better, orโฆ
@jwillis good point about getting management to learn Taguchi. Learn, change, grow, get better, orโฆ
Although TPS(Toyotao Production System) is quite famous because of Taguchi, many Japanese will argue actually a lot of ideas of Lean came from TDS(Toyota Product Development System)
Good post about platform as interfaces ... they don't describe it this way.. but the ideas are in the paper. https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/43438.pdf
@jwillis I saw a presentation from Doug Kirkpatrick a few months go about org structures. Are you aware of his work? The No-Limits Enterprise: Organizational Self-Management In The New World Of Work
It is certainly not a traditional org. Sounded like a great way to run one though. Based on relationship agreements, not management.
One helpful way Iโve found to talk about platforms is to ask people to think about the definition of the word โplatformโ (the generic, non-tech definition). What is a platform? Itโs something you put things on. Does that tally with the platform as an interface concept, @jwillis?
I was hired as CGO for a giant banking software company to build their platform and then they asked why we were trying to make one in a fun board meeting a year later ๐
I was hired as CGO for a giant banking software company to build their platform and then they asked why we were trying to make one in a fun board meeting a year later ๐
Thanks a lot @jwillis
Thanks @jwillis . Had vibrant memory of your talk at does Beijing, not disappointed by this one as well.
(I find it fascinating that platforms is a very hard structural separation, but truly enables dev productivity. Iโm so fascinated by structure these days, as maybe you coudl tell from my openin remarks! @jtf)
@john710 I find that lexicon is one of the worst enabling problems in large organizations.. not just platforms. However, platform is one of the worst. I've heard crazy names for platforms in an org.
Diversity is a strength when people stop seeing differences as a threatโฆ
Diversity is a strength when people stop seeing differences as a threatโฆ
โฆ which means people need to learn how to do that!
@jtf my Sensei and Coach says that it is not the differences that harm teams, it is the judgements of those differences.
Very nice link between that and Non Violent Communication. It took me years and years to get around to reading the book and learn that by violent he meant judging. (โThat was a great job!โ is a violent statement.)
To actually change behavior Iโm a big fan of Chris Argyris. (We use his models in our book Agile Conversations.) This article gives a good flavor of how he used his approach to surface and discuss differences appropriately: https://hbr.org/1986/09/skilled-incompetence

@nickeggleston I will. However, I double booked by accident. I have a 2pm at another large conference today. I won't mention them they are not a sponser. But I'll be around before and after.
Interesting to see that ethnic diversity seems to have an even greater positive effect on a company than gender diversity.
Diversity across many domains (of thought, culture) as well...
Diversity across many domains (of thought, culture) as well...
This made me think of one of our old IBM mottos to 'treasure wild ducks' https://www.ibm.com/ibm/history/ibm100/us/en/films/wild_ducks.html
So, the science exists. When facts don't convince people to follow the science, what to try next? @jwillis
Effective collaboration means productive conflict. Thatโs one of the key values of diversity.
Effective collaboration means productive conflict. Thatโs one of the key values of diversity.
Productive conflict between ideas rather than unproductive conflict between people.
Years ago I worked for GE and was able to go to Crontonville .. back in 1993 One of GE's core values was Diversity. One of the best classes I've ever taken called Manager Modeling.
Thank you for the opportunity to do this. I appreciate being able to share my passion.
diversity should be done for justice not for economic reasons financial, economic gains should be considered a nice side effect
diversity should be done for justice not for economic reasons financial, economic gains should be considered a nice side effect
Itโs great when the right thing, also does so much good overall, including financial results
I think thatโs the real point right there. ^ ^ Do the right thing, even if it costs money and time, even if itโs hard, because in the end, itโs worth it.
Hi Josh!! I had such a visceral reaction when I saw this talk โย I find โdevops shamingโ incredibly frustrating.
Loving this topic. creating a positive experience and guiding folks is what brings folks on the journey. Reminds me of the first follower video.
I use DevOps very cynically. If we are the "DevOps" group, then we must know what we are talking about because "DevOps Engineers" are paid more. All part of being a wildly optimistic, pragmatic cynic.
I use DevOps very cynically. If we are the "DevOps" group, then we must know what we are talking about because "DevOps Engineers" are paid more. All part of being a wildly optimistic, pragmatic cynic.
Slide about the race car reminded me that Infrastructure teams are just another feature team
Not everyone has adopted agile yet... hear hear. Selling DevOps as the "next step from obsolete agile" is another weird idea.
Not everyone has adopted agile yet... hear hear. Selling DevOps as the "next step from obsolete agile" is another weird idea.
That will happen at somepoint. so someone can buy a bigger boat.
Newsflash: The term DevOps is NOT appealing outside IT when they are touched by our transformation. Why not sell it as Agile or Lean?
I sell it as faster, higher quality delivery that requires partnership from the business to deliver that helps improve our bonuses.
Who needs a Toyota when you have BMW? ๐
This is a great presentation by @patrick.debois256 that gives great isight to how the movement evolved and why Devops https://www.youtube.com/watch?v=cNdb8FaTaOQ
would love to hear more expansion on "Being a Journey Guide". It resonates with me. As a facilitator of workshops it is also akin to how you guide the group through THEIR conversation and not be an answer-person.
be a good person, shape a better world and give feedback
THANK YOU @sam and @patrick.debois256 for landing the plane on Day 2 without mishaps!!! Yโall are awesome!!!
Interesting echos of @jwillisโs talk about platforms in @david627โs Q&A right now: by enabling the platform which disseminated the information, multi-team agility and velocity was enabled - moving from differentiation to scope economy