You Build It You Run It sounds great... but it won’t work here!
You work in an enterprise organisation. You’ve got lots of delivery teams, and a central operations team that does deployments and production support. You need daily or weekly deployments. You need 99.9% reliability, and less than 1 hour to restore. You need a growth mindset, and learning culture.
The good news is you can have it all with You Build It You Run It. It’s an operating model in which on-call product teams build, deploy, operate, and support their own digital services.The bad news is... well, for some, it’s a bit controversial. It’s a 15 year old idea that flies in the face of traditional IT, and still isn’t well understood in enterprise organisations.
- How You Build It You Run It can be proven to protect more revenue than it costs
- How developers can be encouraged to try, and even enjoy, being on-call
- How to clearly establish who to hold to account for reliability
- How to build an effective, repeatable incident management process
- How developers can spend most of their time on product features, not fires
- How to balance demand and availability of specialist skills
This talk demystifies You Build It You Run It fears, drawing on experiences from multiple enterprise organisations who’ve successfully implemented on-call product teams. It explains how to simultaneously achieve frequent deployments, high availability, and a high rate of learning across many teams and digital services.
Outline/Structure of the Talk
This talk is inspired by Jez Humble’s 2016 talk ‘Continuous Delivery sounds great but it won’t work here’ https://vimeo.com/193849732. It includes:
- What is You Build It You Run It and its forebears
- What’s involved in deployment throughput
- What’s involved in service reliability
- What’s involved in learning culture
- It’ll cost too much!
- Developers won’t want to do it!
- Our Ops manager has to be accountable!
- There would be no incident management!
- We would drown in BAU maintenance work!
- We would need 20 DBAs for our 20 delivery teams!
You Build It You Run It is a hybrid operating model, which can co-exist with a central operations team:
- On-call product teams can achieve a level of throughput, reliability, and learning culture that working with a central operations team cannot
- Run costs need to be weighed against revenue protection and opportunity costs
- Making one person accountable for the reliability of everything isn’t necessary
- Incident managers can help on-call developers with an effective incident response
- Developers can balance product features and BAU maintenance work themselves
- Highly skilled, in demand staff like DBAs can happily remain in a central team
All conference attendees
schedule Submitted 7 months ago
People who liked this proposal, also liked:
Jeff Patton - The Mindset That Kills Product ThinkingJeff PattonAuthorUser Story Mapping
schedule 2 months agoSold Out!
We all know what a product is. We buy and use them all the time. But, what does it mean where you work? Why is there so much resistance to creating great products where you work? This talk focuses on how to recognize the mindset that gets in the way of effective product design and development. You may be guilty of the mindset yourself. I know I am. We’ll talk about how to recognize when you, your leadership and the processes you use have slipped into anti-product mode. And, give you ideas on what you can do to fight it.
Ben Conrad / Gerald Benischke - Securing digital services at HMRC DigitalBen ConradHead of Product, MDTPHer Majesty's Revenue and CustomsGerald BenischkeAppSec LeadHMRC Digital
schedule 5 months agoSold Out!
HM Revenue & Customs is the tax collection authority for the United Kingdom government. The department is responsible for the collection of revenue (taxes and duties) from all UK taxpayers, be they citizens or businesses. We work together with the Government Digital Service (GDS) and other departments to ensure the services we provide are built to common strong standards.
This session peels back the covers on what it is like to secure HMRC’s digital tax platform, which is built on AWS and comprises 1000+ microservices built by 100 teams with ~1500 deployments a month. Security incidents such as Log4Shell and news reports of data leaks are always a risk to digital services, and at HMRC Digital we’re in a position to react quickly and confidently to incidents as they occur.
We’ll share some insights into how we’ve secured the microservices that run on the platform, including:
- identifying vulnerabilities prior to live deployments
- increasing buy-in from teams for service security
- leaning on an opinionated tech stack to boost our security position
- using a service catalogue and async chat comms to power security collaboration
We will also provide some recommendations on what you can do to get started with your own AppSec programme.
Sneha Prabhu / Archanaa Ravikumar - It's a jungle out there, but we are your tribe!Sneha PrabhuMarket Delivery PartnerThoughtworksArchanaa RavikumarHead of EngineeringAutumn Life
schedule 5 months agoSold Out!
Which woman in IT does not dream of being the next Sheryl Sandberg? Of having wielded our shiny swords and vanquished every demon that crosses our path to that dream leadership role? Of putting on our superhero capes and whizzing past misogyny to glory?
If we were given a rupee for every time we became the default note-takers in a meeting, we’d be vacationing on our private island right now. Instead, we’re here, putting together a talk to share the hurdles we’ve faced in leading tech teams, and how we overcame them. We are now in a mindspace where we can look back and laugh at the memories.
The two of us worked together when we were bright-eyed young graduates, fresh out of college and dreamy-eyed about our future. We went on to take different paths - one as a tech leader and one as a business leader - despite which we hit several similar challenges. We found solidarity in each other, and our shared experiences helped us grow stronger and gave us impetus to push ahead.
The journey to get there in each case was equally hard, and we learnt and grew tremendously. We hope that our learnings will help the rest of our tribe get past the same challenges with ease and much less pain
Archanaa Ravikumar / Sneha Prabhu - Don't just build the rocket, take it to the moonArchanaa RavikumarHead of EngineeringAutumn LifeSneha PrabhuMarket Delivery PartnerThoughtworks
schedule 5 months agoSold Out!
There is no dearth of great product ideas. There is also no shortage of people who are driven towards making those ideas a reality. With plenty of access to talent, capital and resources, dozens of startups are emerging every day, eager to take their world-class idea to fruition, smartly and quickly.
The 0 to 1 journey is a much documented story of our generation. We’ve heard the stories of countless startups who have hit the ground running, and garnered attention and funding through early validation.
Much less is written about the journey beyond validation. Ideas fizzle out, execution fails to scale, and capital starts drying out. The best companies are those that react to the market while continuing to evolve, scale and grow.
When considering scaling, companies are often faced with problems around conflicting priorities, constrained capacity, an evolving culture and market pressures. The same old processes cease to work and process optimisations can only reap so much benefit. So, what happens next? There's a need to take a step back and restructure the teams, to focus on the right priorities and invest in nurturing the right culture.
A startup is deemed successful when they’ve found product-market fit. However, we believe that that is the start of the real journey. That is the time to load up on ammunition, chart out a clear strategy, and win the many battles ahead. As Victor Hugo once said, no army can withstand the strength of an idea whose time has come.
Kumar Iyer - Mind your MetricsKumar IyerPrincipal ConsultantThoughtworks
schedule 6 months agoSold Out!
While many software delivery teams and organisations do pay attention to Engineering metrics, they often fall trap to measure what is convenient. Test coverage is a stellar example of an overly (ab)used metric. This talk aims to put forth a few unconventional, interesting yet useful engineering metrics ranging from code-level metrics to Devops metrics and ways to measure them. It also brings in the notion of a hierarchy of metrics where a collection of low-level metrics are known to influence higher-order metrics and so on, ultimately resulting in the greater good of the software product.