I wish we could be in person :)

The real thing

Defo

Used to seeing you on the big stage

Ditto!

Great q

Could you compare loss of developer talent, througput, and a huge stack of debt cost to empowering devs to succeed?

if you care about your customers experience, Why don’t you care about the employees experience who create it?

changing dev experience requires change - a move away from the status quo. It will probably come at the expense of biz features, so I guess it was a tough sell

Biz features without a strong change muscles are destined for irrelevancy

Ok so day 1 was great for your biz feature, whats day 2 look like? 6 months later - why are we keeping this?

amazing, thanks!

Some days I feel like we’re impeding when we make mistakes, most days I say we’re actively empowering IMHO.

I think if you are even asking the question you are more than half way there

Asking the right questions is just as important as solving the right problems

Too right @chill9, asking great questions can unlock so much, being vulnerable enough to try a question and not have it be quite right but you've opened up the channel of conversation is a big step for some people...

I threw that in last minute, glad you liked it

It's important to acknowledge that the new thing we are building is built on the bones of what died before us and to learn from that. I've turned to Inner Source as a techniques to break the cycle of governance and neglect. I'll have to wait for another year or so to know if it's worked.

I fully agree, transformation usually doesn’t give fruit to those who started it :(

If they are incentivized by features and sexy, why even acknowledge the debt?

Tons of push back, but warmed up with being able to articulate cost of delay

Thanks

Well said

I used to want to find out who wrote it and beef up the chest, but I realize perception is gold no matter who it comes from

If you can flip a team culture so bad feedback is a (constructive) norm, especially upwards, then you're really rolling

Great point, also good to look within, do your leaders accept bad feedback for themselves?

Good q - is blame an important thing to reconcile? Or can accountability be shared?

shared accountability leads to no one taking ownership

and it is not about blame it is about owing the end to end experience at a stakeholder level...

and friction is real

I am curious to learn how t-mobile has solved this

If something goes wrong and we’re all the same soup, don’t you think the biggest impacted person/team would be incentivized to fix/revert/etc? It’s a challenge and at the end of the day there are shoulders which it could fall on if the first question is “why didn’t they team take care of that” - it might be the wrong question... and to back up. “Why didn’t we catch this before the problem?”, “how can we be empowered to help more?” Incidents are not always a bad thing if you’re able to change because of them and get a ROI

It’s a huge monolith, a lot of patience and agree on principles first

Oh, wow. I worked on a similar project a few years ago. Most of the monolith is really in the minds. It's not even the monolith per se, it's a lack of expression of intent, understanding of the real functionality, and no doubt, lack of meaningful automated tests. Ok, I may be projecting here (ha!) but I'm betting it's not too far. Best of luck and thanks for the talk!

Nice to hear from you Ken! Hope all is well!

there's also a prequel: https://pickl.eu/blog/getting-out-of-quicksand-with-devops/

So true Paula - we really have to fight hard to be a software company supplying our products to automotive rather than an automotive company doing software.

The typical cycle time in aviation is 8 years. Not a typo. I don't know what it is in the auto industry, but by guess the same. And yet how much code runs trains-planes-automobiles these days? millions. Its a hard culture shift

definitely... even though for automotive, cycle times are "only" about 4 years 😏

we are investing in test hardware and making testing easier. at the same time we invest heavily in simulators / virtualisation.. because we will never have enough hw

gary gruver discusses this in the hp firmware story. he said that there will never be enough printers to test everything 😄

Thanks! I will reach out to you on LinkedIn to take this offline

This is so true. Especially in the physical world. Even if you can test, machining a test part costs major $$ and time. Its painful enough there's a bit of a culture not to do it too much (my experience). And sometimes you can't replicate a 20 year old machine...

well, basically we are still stuck with yearly budgeting cycles that give us little flexibility to react to changes in the market... so we bet on each of our products at some point in time rather than acting in an agile way

ah... have you tried to propose a better way?

we've gone big on budget in a similar way to what you describe

I'm searching for a better model to propose that supports enterprise agile / value stream flow, but also gives feeling of control/reassurance to senior leadership

we are actually currently working on a proposal that does exactly what you are looking for as well

unfortunately we don't have anything yet that I could claim will most likely work for us

(and that senior management will agree too...)

good luck - keep in touch!

I'm @PhilipDayUK on twitter

Thanks philip 🙂

Thank you Philip

do you mean general dashboard frameworks? I'm afraid most open source solutions are in a rather sorry state as the dashboard craze seemed over.

there are some great saas solutions afaik

we currently use the westrum model to measure it quarterly in my team. and ENPS

what do you use?

So I work at a place that makes a product that evaluates PS over time - I took the final slide as you wanting to do more in that area. Could also be me misreading 😄

As in if you're happy with what you've got, I won't be a pest.

i think atm the google/microsoft form is enough 😉

Thank you Andrew for your contributions! We have made so much progress in our Salesforce space. Appreciate you!

Hi Misty! Time/space varies based on local context; however, we have incorporated learning into our planning model. For example, when we got the pilot areas onto the DevOps journey, we planned for 10-15% of dedicated learning, so our team members would have the space to balance learning with delivery.

We haven't put a % on it, but we carve off Fridays for my organization and have deemed them #futurefridays, our teams have even changed the 3 questions they ask on Fridays to be around learning 1. What learning did I accomplish this week? 2. How do I believe that will help me contribute more broadly to the team? 3. What learning am I committing to for the next week?  This shows we are willing to invest the time to foster learning, but also helps align learning to the team needs.

I love how you all have reframed the questions in a learning context. That's a great idea!

Was thinking about AA's keynote yesterday and your organization's approach to investment decisioning - that's how we've approached DevOps in this space. When we started the Engineering Capability Model pilot a year ago, we proposed it to our senior leadership team in 'investment terms', which is what led us to be explicit about dedicating 10-15% over 12 months for training, practices adoption & growth, so our senior team could make balanced investment choices without compromising the space/time/sustainable-work-pace for DevOps adoption.

@nazia.ali are you all working on any tools to help people participate remotely in conferences from the comfort of their autonomous vehicles? 😉

@abd3721, touche'

@fernando.b.rodriguez. - Yes! We have a sample in the presentation and happy to provide additional context! @abd3721 - love it!

Mary! Wow, so awesome to see you here! Thank you so much for attending. Hope you're enjoying DOES20!

Thank you so much Bryan! Hope you're doing well. Cox Auto misses you for sure! Thank you for your contributions and thought leadership. Much appreciated!

Great to hear from you Bryan! Hope all is well on your end!

Thanks for carrying the torch and pushing the envelope! I'm glad to see you sharing your journey, it's inspiring and impressive!

Shout out to @chris656 who introduced me to VSM. And @damon for all the additional support 😉

Hey, whatever works 😂

1. Request access to Cld subscription

1. Request access to Cloud subscription via manual Jira to spin up experimental infra 2. Get an email asking how many VMs I want.... 3. :man-facepalming:

A precondition to "the conversation" is probably trust and psychological safety. But.... I think other lean tools are good here too. Five whys, etc are good.

it is! the whole process is hard. But don't let perfect be the enemy of the good.

Yes! I think I have an example later (I can't remember) where the goal was quality/defect escape. Though turns out the side effect is usually better efficiency either way.

@aasantos They might benefit more from a problem management discussion that delves more deeply into why they have stability issues... but the part of value stream mapping that talks about evaluating options that could eliminate the stability issues would definitely apply

@paula.thrasher @denee.ferguson Thanks for the input! In my org there have been some heated discussions on whether VSM should be applied only where time to market has been explicitly identified as a source of dissatisfaction, or more broadly as a tool to build a common understanding of processes, stimulate conversation and improve collaboration. It seems the latter woult better align to your comments. Really appreciate it!

@paula.thrasher Curiously the process was not unfamiliar to me... a previous VP was a big fan of "The Goal" and had us all map out a process and find the bottlenecks and identify ways to optimize it.... I just didn't know that was what value stream mapping meant....

It's a great tool, but I've been seeing some resistance to do it because it is seen as complex and expensive activity, as you need to have many people in the room for several hours. So I'm trying to articulate benefits beyond waste reduction in order to garner some more buy in.

@aasantos Consider having a small subset of folks work on drafting the flow chart in advance. Leverage the bigger session to get feedback on accuracy/missed items. That can reduce the time required... I've found the process flow to be eye opening for senior leadership - they don't always understand the pain their teams deal with to get things done.

So I absolutely have experienced the push back you have about this being a big expensive tool. I was at a large manufacturing firm where this was a three day minimum exercise. To my point on the magic middle ground: If you give people homework/prep time, you can have a good and effective conversation in 2-4 hrs. Thats a much easier sell.

To answer the question on when its not the right tool - this isn't for newly created/re-organized teams. There is not enough forming - storming - norming yet to map in that case. Second, if you have a glaring hole you don't necessarily need VSM to find it. That said - sometimes you need the visual and "the conversation" to get consensus.

@paula.thrasher @denee.ferguson These are great insights. Thanks a lot!

I think this is why I ask teams to gage the broadness of the improvement before addressing it. ideally its single team but if its multiple teams the owner has to be higher in the organization? if that makes sense?

Like a Product Owner in a Product VS?

We’ve used Miro and been really happy with it

I have also heard some good things about Lucidchart but I've never used it. Keep the tool ideas coming - I'm open to learning too!!

Thanks. Have tried Lucid Chart and it has some good capabilities. Haven't tried Miro. Keeping teams engaged on remote VSMs can be challenging

I've used google drawings, powerpoint, http://draw.io, keynote, lucidchart, Google sheets, Miro and Mural, and my favourite is Mural easily

The advantage of the collaborative whiteboard is having everyone inside the meeting and contributing, and mixing collaborative input, voting, brainstorming etc drives engagement

@nepobunceno @priyadarshini.balacha

Thanks for sharing Paula! I use Mural as well and love it (along with the teams I work with - they all start using it for most meetings after our sessions)

I wish I had Mural. 😢

We don't tool shame here 💪, anything collaborative will work really well, and I still get the same outcomes with a screenshare (with some extra engagement effort)

I agree. I wish I had a collaborative tool for this. 😞

We just tried, failed, tried, failed less, etc.

What do you consider failure?

Failure is OK!

I think the one skill you may want to tap is a good facilitator. That can be an existing facilitator (agile coach, etc). But otherwise - I mean you have to start somewhere so find a starter project and try

Or success?

Listened to examples from @paula.thrasher @chawkladyand an article from Microsoft.

We just dove in. Even having the conversation is beneficial.

(in my opinion)

Failure in this context was that it felt ineffective the first time. So, we value stream mapped our process for creating improvement playbooks. Ate our own dog food and did instant retro as we did it. We were better the next time. We continue to refine it.

Thanks this give me the courage to "try this at home"

@kim.darzi Ditto here

@kim.darzi @dmaillet63 I truly hope this helps you run your own session. My goal is to make this a more common practice! Feel free to use/reuse the handouts and slides and I am happy to talk more if you have questions.

I can attest that it's a really powerful tool that teams always appreciate.

Most often there isn't one, but it's a natural match for business analysts, scrum facilitators and even product owners. Some leading organizations have a role like the value stream architect and dojo staff like Walmart, but it's rare at the moment.

Whatever facilitator you have available? I don't have a single answer there. I've done them with outside facilitators, with agile coaches as facilitators, I've done it (but not for my own team), I had the HR team facilitate once (No really! they were great!)... I think the key is to have someone neutral do the Socratic method thing and ask questions, and also to mediate conflicts. Just not someone on the team because (1) you can't facilitate and participate at the same time effectively and (2) you can't mediate conflict if you are a stakeholder

Yes 100% this ☝️ Psychological safety and facilitating skills are more important than understanding the domain or context. The team brings that and uncovers it during the process.

This is an awesome way to get started without higher level buy-in or a dedicated effort, it pays off in the short term via better self understanding but also gives you an awesome artifact to share with others and spark some progress

Thanks! There are things to like about the virtual conference format that I can't do in person 😀

it is defiantly one of the positive aspect that came out from do everything virtually. I think it will be interesting to see how we as a community will adopt to this new environment and how much we will keep future interactions

Yes!! Thank you Paula and thank you for your excellent talk!

I would love this. I learn something new with each group I talk to. https://calendly.com/paula-thrasher and lets find a time to chat!

Great, I'll schedule some time for me and Ashley to meet with you. Thanks!

that's what's missing in any transformation 🙂

yes!

That was another system that was built separately, but integrating with it was a huge win for our framework

can you share tools ?

hmm I see job postings for it, so I guess it's not a secret haha. We use ServiceNow and there's a completely custom tool that integrates with it

ha ha 😄 then we are going in the right direction

We do the same

it's an API layer on top of ServiceNow and CI/CD frameworks can integrate into it

this is one of those undifferentiated heavy lifting that every org is doing, but hasn't been solved by any one org for everyone yet

how many CI CD tools you use ?

it's mostly based around Jenkins

It's currently an internal product - we're hopeful it could become an open source product 🤞

any chances to share with a fellow practitioner? If not, maybe the architecture diagrams and overview of what composes the architecture so I can build on top of that?

@roderick.randolph

I can say that we use a lot of open source tools (including terraform, ansible, docker, etc) 😄 - let's connect on http://linkedin.com/in/roderickrandolph to keep this discussion doing

taking you on the offer @roderick.randolph, just sent you an invite 😉

I'm sorry to say, but I actually haven't had a chance to watch @tapabrata.pal's session... is it online?

if he's online he can chime in, we work together now 😄

Yes. @matthew.cobby this is a part of the implementation of what I shared in terms of software clean room

It's nice to see the evolution of the ideas

I think it was one of the best things we did for the Innersource project hehe

and yeah for sure, can get a lot of designers or devs with a design bent engaged

Pretty much, but you'll see a bit later some of the things we did to try and help with the organic growth

Eh!

:flag-ca:

Same thing at WM

oh yeah? Very cool 😄! I'm in the Toronto area myself 🙂

I'm in Arkansas, but some of our highest performing teams are in Toronto. Must be the LaBatts

hahaha

Inferiority complex? That's my excuse...

Hey @steveelsewhere !!

Hey Arthur! Thanks for once again repping the north with 🔥🔥 along with Roderick - it's so awesome to see you folks here, with so much progress every year

Haha thanks @steveelsewhere, thanks to you @roderick.randolph and I had our first duelling presentation :)

At WM, we control the platform, they control their configuration. I've no idea ow anyone could control them centrally.

Same at USAA

We inject security scans and compliance rules. Don't want anyone to forget those. :)

We have a feudal system. Every Service Owner has and runs their own CI/CD. It's chaos at scale. We've been rolling out a common CI/CD pattern but it's still very distributed. One central Jenkins with central config is a definite anti-pattern

Same with us, mostly. Our China branch kind of control it centrally though. They basically give the standard pipeline script to all the teams, so basically everyone follow the exact same pipeline. I do believe each team can build on top of it and customize though...

We have something similar, basically the original incantation was Jenkins Shared Library that teams could put together as lego pieces to build their own pipeline. The above question about automating the change control, that helped focus the organization around a set of required stages that had to be run

so in order for teams to qualify for automatic change management and onboard to Continuous Delivery, they needed to conform to a certain set of stages

and yes, as Michael Pai said, that was the next iteration that fixed the stages and allows developers to configure the stages as needed. This allows us to ensure, to @bryan.finster's point, that certain security and other stages always run.

@matthew.cobby: we have a Jenkins as a Service. Basically we run a K8S cluster with one shared Jenkins for smaller teams and different namespaces for standalone Jenkins. We build the Masters from the same image, helm chart and jenkins configuration as code file, so every master in the cluster starts the same. Once the master is created we give a sub-set of the users admin rights (with limitations though, somethings we dont want them to change)

we did always generate a base Jenkinsfile as part of the tooling, though many devs take it and run with it

@rodriguessemensati.e - A central operations teams for a Jenkins cluster is a good way but let the teams control their own abilities

Central CD platform org is key, IMO.

we do, once we create the Master we give it to them (mostly, as I said somethings are mandated centrally, like security configs, IDP integrations, etc)

Fragmented platforms cannot be made secure and compliant.

If you mob it, this is built in

and usually it's you that reads it haha

so often it's helping future Matt 🙂

💯 it bring more transparency to how the software actually works!

mobbing has been a huge benefit to onboard teams, expand knowledge, and solve complex problems.

many IDEs have pair programming feature built-in - super valuable in this virtual world!

Day 1 of covid guess who showed up in our dojo? Woody Zuill

More of the former, I haven't seen Dojo as a concept yet at Capital One

our workshops would last for a couple days, but not an immersive multi-week Dojo

We've been doing the same as you. Engineering Bootcamp over 4 days. Day 1 is intro to basic tools, Day 2 is intro to app dev, Day 3 is hands on our microservice template, Day 4 is our UI application template

yes! So these bootcamps are a bit different, they were more around onboarding new teams to work with us on the CI/CD framework

to your point, I helped introduce a 5 day bootcamp in Canada for new hires that guides them through pretty much the same as you mention @matthew.cobby, good to hear we're approaching it the same way 😄

the first day actually spends time on getting new devs familiar with the business, how Capital One makes money, etc.

This really came from the top, our leadership recognized that technology is going to be a competitive advantage for us

then came the actual transformation, which was a huge amount of work on the dev side of course

but it really does require buy-in from leadership

it didn't start there though, I know that the org was dabbling with AWS much earlier. Regardless though, transformation at the entire org level requires CEO and exec buyin

I see why Capital one has the name "Tech Company" rather than "Financial Company". That's great to hear that leadership is with such great mindset

oh yeah? How are you finding it

I've been thinking of using that as well, but there's so much choice

It’s old school - node js

I like it

yeah I remember using it back in my Angular days

There’s also backstage from spotify

Cookie Cutter looks pretty good too: https://github.com/cookiecutter/cookiecutter

Which I hear is cool

Yea - AA uses cookie cutter with backstage

ohhhh! I just listened to a Changelog podcast about Backstage, it sounds super interesting

https://asciinema.org/a/5ucnJpJFpSTpzCUV9UcbgO79v?t=16&size=medium&rows=20&style=tango&speed=3&poster=npt:40

That’s yeoman

yeah they have a fancy CLI UI 🙂

ahh that's your tool? Very nice

funny enough, we're using https://github.com/CITGuru/PyInquirer to do a very similar interaction 🙂

💯 agree - docs must evolve with the software

GitHub Pages? or using another static site generator for software docs specifically?

Mark down content and sphynx generator...

It allows us to review content of docs side-by-side with code... e.g. PR comment: "You changed the way this works, but didn't update the docs!"

We use use vuepress with a company template and push to GitHub pages

(this is beyond just code-level doc gen, but more of the big-picture and support documentation)

we also included a PR review template that reminds you to update the documentation as you go 🙂

PR templates with a checklist are awesome 🙂 and yes @jeanetienne.lavallee I mean markdown files to generate documentation sites, similar to what GitHub does with their own documentation (which has been recently open sourced by the way)

Spot-on