Hey there! This is Platform Weekly, your weekly look behind the curtain on the platform engineering revolution.

The Five Families of Platform Engineering have met. This is what they said.

On Tuesday, this week we had a godfather style meeting of some of the biggest thought leaders in the platform engineering space at the AWS re:Invent platform engineering executive roundtable. It included my friend and CEO Kaspar from Humanitec, Sean Alvarez from ThoughtWorks, the man himself Gregor Hohpe, and a selected group of practitioners from companies like WesternUnion, Salesforce, IBM, Oracle, GitHub. I don’t think it’s unfair to say that the conversations in that roundtable will shape a lot of the platform engineering world in 2024.

So what was discussed? Let me break it down.

The impact of Internal Developer Platforms is Insane 

It’s now been almost 2 years since the first companies started to put together a modern enterprise-grade Internal Developer Platform. That means we’re starting to see the first real long term data on just how effective those platforms are. And the numbers are awesome.

2 years ago, I was often talking about potentials and hypotheticals about what I felt platform engineering could deliver, and now I’ve had the chance to see the numbers and results firsthand.

The pain and value is being perceived differently by different teams

Every company is different, and every team within those companies are different too. You’ll have teams with very different goals and objectives for their platform, and these teams will have different views for whether the platform is working effectively. That is why it’s so important that Internal Developer Platforms are being set up to best serve exactly what teams actually need, based on insights you’re getting from the actual teams themselves.

You need to start small, and start fast. You can’t be trying to build your final product from the get go

If you want a platform. You’ve got to start now. And you’ve got to move fast. Platform engineering seems to be like water, it will fill the space you give it. If you set a timeline of 2 years, it will take 2 years. If you focus on wanting your 2000 developers to be using it, you won’t see results until the 2000 are.

This is the absolute worst way to approach platform engineering. You want to get it started as soon as possible, even if that means starting small. 

“Pick your pioneer team and get them started. Fine tune, see results, and expand”

The performance of the early platform is crucial to companywide expansion and adoption

Everyone will be looking at that first platform. Many will have preconceived views, but in the end, it’s the performance of that platform that will set the stage for a company's platform engineering initiative.

As much as I hate to say it, there are still many people who perceive platform engineering as a fad. The best way to silence those critics, and make sure there is no slowdown in adoption is to make sure that that early platform works, and works well.

Want to attend things like this yourself? The platform engineering community organizes awesome webinars like this every week

Quick bites

Articles that blew me away:

From the community: