Craft Platform Engineering Change "Baby Steps"
Part of a "Software Engineering Enchiridion": Platform Engineering
(Image: My brilliant nephew, Teddy, making his own valuable baby steps in the world these days)
“First say to yourself what you would be; and then do what you have to do.” — Epictetus
You don’t need to boil the ocean. You need a kettle. Heat the water. Pour the tea. One cup. Then another. That’s how platforms are built. That’s how adoption spreads.
Big-bang platform rollouts? They’re the soufflé that collapses before it leaves the kitchen. Nobody trusts them. Nobody wants them. They’re slow, heavy, bureaucratic. By the time you’re done, the world has moved on.
A baby step is different. It’s sharp. Focused. It solves one pain in a way so obvious and useful that people adopt it before you’ve finished your pitch.
But here’s the catch: the baby step must be valuable in two directions.
For platform curators, it must simplify, harden, and sustain.
For platform users, it must make them feel like they’ve just levelled up. Not just “oh, neat,” but “hell yes, I can ship faster and look smarter doing it.”
This is where Kathy Sierra’s idea of “badass” comes in. Don’t build features. Build confidence. Build swagger. Give a developer the feeling they’ve just pulled off a magic trick. That’s what sells. That’s what spreads.
Aphorism
The smallest step, if valuable makes users capable, more confident, more effective, is the greatest leap for adoption.
Practices
Keep the steps small. Smaller than your instinct tells you. But frame them high. Frame them as transformations, not chores.
Look to tell two stories at once:
Ops says: “It’s secure by default.”
Dev says: “It’s one line of config.”
Then sell it. Sell it in your weekly forum. Sell it in hallway conversations. Sell it by demoing it so simply that your audience thinks, “I could have done that. Hell, I just did.”
And then measure the right thing. Don’t ask “Did you like it?”, ask “Do you feel more effective now?”
Pitfalls
Don’t make the curator’s toy that developers see as busywork. Don’t build the user’s candy that leaves curators with a stomach ache. Don’t dress up a whale and call it a minnow. If it’s too heavy to trial, it’s not a baby step.
And never assume good technology sells itself. It doesn’t. Ask Betamax.
Checklists
Does this make a developer feel sharper?
Is it small enough to test in a sprint, but big enough to matter?
Can I pitch it in one breath?
Does it make life better for both sides?
Do I have a story so compelling it spreads without me?
Examples
The one-click secure storage module is a baby step. Developers feel like instant security ninjas. Curators sleep well because compliance is baked in.
The logging pipeline that only ops understands is not. Developers shrug. They don’t care.
The flashy tool with no governance is not. Curators get wrecked cleaning up after it.
The API gateway you can spin up in minutes is badass. It sells itself.
Here’s the truth: adoption isn’t earned with mandates. It’s earned with badassery. The smallest, most sellable step that makes someone feel more powerful will outpace your architecture committee every time.
Build steps that make people better. Build steps that make them faster. Build steps that make them want to tell their friends.
That’s the job when you're evolving a developer platform.
And you are always evolving it as there's always one already, but that's the subject of a different maxim.
Further Reading
Kathy Sierra & Bert Bates, “Badass: Making Users Awesome”
Gene Kim et al, “The Phoenix Project”
Kent Beck, “One at a time”
Marty Cagan, “Inspired”