hardScenarioCase StudySeniorStaffLead
Your DevRel programme reports growing impressions, follower counts and packed conference sessions, but finance is asking what it returns. How would you show whether the work is driving adoption, and how would you defend the budget when you cannot attribute cleanly?
5 min readUpdated 2026-08-15
Answer
Short answer
Split your numbers into reach, engagement and behaviour, and only ever claim the behavioural ones as results. Where attribution is impossible, defend the programme with cohort comparisons, holdouts and counterfactual costs rather than an invented revenue figure.
In a strong developer relations answer, the proving devrel is working without attribution detail should connect the decision to a visible consequence.
Was this answer helpful?
Likely follow-ups
- Your best-performing tutorial by traffic correlates with no measurable increase in successful API calls. What are the plausible explanations, and which would you test first?
- Finance offers to fund either two advocates or one advocate plus a paid conference booth. What evidence would make you choose the booth?
- How would you set up a holdout so that withholding developer content from a segment is defensible rather than reckless?
- If your only usable metric were time-to-first-successful-call, what would you stop doing?
Related questions
- You are taking questions after a talk and a developer says your API's error handling is a mess and cost their team two weeks. They are partly right, and the fix is not yours to promise. What do you say, and what do you do afterwards?hardAlso on developer-relations and developer-advocacy5 min
- Your advocacy team can either build things - an SDK, sample apps, better reference docs - or publish things: posts, talks, videos. Both compete with the product roadmap for the same engineering time. How do you decide the split?hardAlso on developer-relations5 min
- Leadership has asked for 20 per cent growth in activations next quarter. How do you decide whether that is the right target?hardAlso on metrics5 min
- Your A/B test was configured for a 50/50 split but the data shows 53/47 across two million users. The result is significant and positive. What do you do?hardAlso on metrics5 min