Post "Accelerate": Why are we still failing at adopting DevOps in the enterprise? | Rasmus Lystrøm

Thank you. I'm Rasmus Lystrøm, and this is Post Accelerate. Or is it really? So six years ago, we got this amazing book from - Nicole Forsgren which irrefutably proves - that not only does DevOps work, - it actually works a whole lot better than anything else we've tried - in the industry. So we ask ourselves the question, - why is it that enterprises keep failing at DevOps? The enterprise will be crashing. So to understand that, I think we have to go back in time to, - I think, - 67 or something, where a - smart person said that enterprises are bound to produce systems which are - constrained by their communication structures. So you ask yourself, what is an enterprise's communication structure, - which has been covered time and again today, - because it's decovered at silos? We have a bunch of silos, - each with their own management hierarchy, - their own budget, their own built-in competition and rivalry, - and chiefs are fighting everyone else in the other silos. And it's always someone else's problem. It's always the other guy. If you don't know what a silo is, by the way, - it's an enterprise-grade security device created to prevent the accidental - transfer of knowledge from one department to another. Thank you, Heidi, for that one. So the enterprise fixes this with a C-Suite 2.0. So we have a problem with communications, so we're going to hire - a chief communications officer whose only role is to bridge the conversation - between silos. And when DevOps tells us that what we have to do to build not software but - products for our customers is to unite all these things into a common team, - this is completely impossible. As the VP once told me, in order to do this, - even with a small team, I need board approval. So what do enterprises do? Well, we resort to siloed DevOps. Let's form a DevOps department. That's an ops department which uses infrastructure - as code in Azure DevOps. That's gonna fail in a couple of years, - so we're gonna form a cloud center of excellence, - do the exact same thing. Now when that fails, we're gonna do a platform engineering department. And two years later, your fantastic platform - And you're going to hear operations say, - boo-hoo, the developers don't want to adopt our platform. Why can't we make them? A platform is a set of tools the IT organization makes for itself. Nobody asked for your frigging platform. Enterprises say we have a culture of you build it, you run it. Who's ever been successful at making a plumber do a carpenter's work? Why do you think that works if you tell developers to do operations? So when we ask for - individuals and interactions and working software and customer - collaboration, what do enterprises do? Give us this. They give us processes and tools, - requirements for documentation and contracts, - and following a plan, and you have to do it via ticketing systems. And get approvals from boards. When we ask to bring developers and business people closer together, - like we said a long time ago, - what do enterprises do? We outsource operations and we offshore development to the Far East. Apparently you're still here. We ask for loose coupling and microservices. What do enterprises do? They give us enterprise architects. Enterprise architecture. And the enterprise service boss. I think it's time we get rid of the IT department. It made sense in a bygone era when technology was separate from - the business, now it just hurts both. We need to evolve the organization, - bring people together, break down the silos, - insist that we work together instead of creating work for each other. We have to take the fight to the CEO. We have to make the organization form a new strategy where we empower - feature teams to ideate and create cool stuff together, - not for the business, but as the business. Now, this is going to feel risky and uncomfortable for most enterprises, - but hey, if you're entirely comfortable with your strategy, - there's a strong chance it isn't very good. Thank you.