The "Field of Dreams" Fallacy
In the 1989 film Field of Dreams, a farmer hears a voice say "If you build it, he will come." He builds a baseball diamond in his cornfield, and players appear.
Platform teams sometimes operate on the same assumption: build a great tool, and developers will naturally use it. This is almost always wrong.
Developers are busy. They have existing workflows, existing tools they know, and existing habits. A new platform tool — no matter how technically superior — has to compete against the comfort of the familiar. If adoption is low, you have a product problem, not an engineering problem.
Mandatory vs. Voluntary Adoption
Before tracking adoption metrics, you need to understand your adoption model:
-
Mandatory adoption: The organization requires all teams to use the platform. In this case, 100% adoption just means 100% compliance — it tells you nothing about whether the platform is actually good. You need satisfaction and DORA metrics to understand value.
-
Voluntary adoption (the "Paved Road" model): Developers can choose to use the platform or build their own path. In this model, adoption metrics are a direct signal of product-market fit. If developers choose your tool over the alternative, it means they find it genuinely valuable.
Most mature platform engineering organizations aim for the voluntary model, because it creates a healthy feedback loop: the platform team must continuously earn their users' trust.
Key Adoption Metrics
-
Monthly Active Users (MAU): How many distinct developers used the platform's tools, APIs, or self-service interfaces in the last 30 days? Track this as a trend — a growing MAU indicates the platform is spreading organically. A flat or declining MAU is an early warning signal.
-
Retention Rate: Of the developers who tried a new platform feature last month, what percentage returned and used it again this month? High retention (above 60–70%) indicates the tool is genuinely useful. Low retention (below 30%) suggests developers tried it once, found it lacking, and went back to their old approach.
-
Golden Path Market Share: What percentage of new services, pipelines, or deployments are being created using the platform's recommended templates versus being built manually? This is the most direct measure of whether the platform is becoming the default way of working.
-
Time to First Value (TTFV): How long does it take a new developer to go from "I want to deploy a service" to "my service is running in production" using the platform? A shorter TTFV means lower onboarding friction and higher likelihood of adoption.
Reading the Signals
If your platform has strong DORA impact but low adoption, the platform works — but developers don't know it, don't trust it, or find it too hard to use. The fix is product work: better documentation, better onboarding, developer advocacy, and UX improvements.
If your platform has high adoption but poor DORA impact, developers like using it but it isn't actually making them more effective. The fix is engineering work: improve the underlying capabilities.
If both are strong, you have a successful platform.