Last quarter an engineer told me: We do not own churn → Pricing, sales, onboarding, support, all of it affects churn more than our code.
He was right.
So we stopped asking engineers to own churn.
We asked them to own what causes it.
Here is the mental model we use now:
- Passive awareness → the why: Engineers know the business metrics. Churn, retention, NRR, expansion. We review them quarterly. They have context and purpose.
- Active ownership, the how: Engineers own product metrics linked to their work. Adoption rate of a new feature. Time to complete a task for the user. Error rate in a workflow. Success rate of an onboarding step. They define, instrument, and track these every week.
This shift made ownership tangible. It also made us faster.
Efficiency is the new growth. In 2025 you win by increasing impact, not headcount.
How we run it in practice
- Start with a user problem and a hypothesis. Product requirements are hypotheses. Example. Clinicians abandon claim submission because eligibility checks are slow and brittle. If we cut the steps and improve validation, more claims will be submitted on the first try.
- Pick two or three leading indicators the team controls, eg:
- Adoption. Percent of active customers using claim submission weekly
- Flow efficiency. Median time on task user want to do from start to submit
- Quality. First pass success rate, client side and server side error counts
- Instrument before you build. Definition of done includes events, dashboards, and alerts. No ticket closes without the metric in place. We use simple dashboards the team updates, not a BI request that ships next month. Copilot helps wire events fast, v0 speeds the UI spike. Keep the loop tight.
- Set a clear target in your OKR. SMART beats vague. Increase claim submission adoption from 42 to 70 percent by end of Q1. Cut median time on task from 6 minutes to 3. Lift first pass success from 78 to 92 percent.
- Review weekly. What moved. What did not. Ship another small change. Remove blame. Keep score.
- Connect the dots quarterly. When a team moves adoption and quality, retention and expansion follow. Show the chain so people see their work climbing up to the business number.
This is how you build product oriented developers. You do not make them responsible for magic. You give them a scoreboard they control, and you tie it to the mission. They become partners in product, not code vending machines.
It also changes behavior:
- Engineers interview users more.
- They propose experiments.
- They spend fewer cycles polishing work no one uses.
Our values come to life:
Outcomes over outputs. Technical excellence in service of users. Direct feedback without blame. Team first.
You can still track DORA, roadmap contribution, and your investment mix. Keep those. But make the day to day scoreboard about user value. Adoption. Time to value. Success rate. Errors. Recovery.
Set context, and make measurement cheap. Do the plumbing so teams can see their impact in days, not quarters. Tie goals to SMART OKRs. Trim ceremonies that do not help. Use AI to cut toil. Protect focus.
Your devs do not need to own churn.
When they own what causes it, the business takes care of itself.