DevRel Without Metrics Fails. Every Time.

There's a romantic idea in developer relations (DevRel) that I used to find sympathetic: our work can't be measured, you just have to believe in it. The trust, the community, the developer who walks up after a talk and says your session years ago is why they're here. No dashboard captures that, so why pretend?

Here's what I've learned watching this field for years: every DevRel team that operates on that idea fails. Not sometimes. Every time. Not because the belief is wrong, the best work really does resist counting, but because of what happens around a team that brings no numbers at all.

Failure one: someone else's dashboard

A team without metrics doesn't get judged less. It gets judged by whatever numbers are lying around, and the numbers lying around belong to marketing: views, clicks, leads, impressions.

So the unmeasured DevRel team gets measured anyway, by a ruler built for a different job, and by that ruler it always loses. It looks like an expensive content team with bad output discipline and a large travel budget. When the hard year comes, and it always comes, that team is the first line item cut, and leadership walks away convinced DevRel doesn't work. The vacuum gets filled. Never by you.

Failure two: you can't steer blind

The budget problem is the visible failure. The quiet one is worse: a team with no metrics can't learn.

Without any feedback signal, every activity is defended by feeling. The conference we always go to. The video series we enjoy making. The format we've done for years. Was any of it working? Nobody knows, so nothing stops, and nothing improves. The team drifts toward what's pleasant instead of what's effective, with full conviction and zero evidence. I've seen passionate, talented teams spend years this way. The passion was real. The direction was random.

Failure three: "we can't measure it" sounds like "we don't know what we do"

Put yourself in the chair of a chief financial officer (CFO) for one minute. Across the table, every function explains what it produces and how they'd know if it stopped working. Then comes the team that says: trust us, it's unmeasurable.

That answer doesn't sound principled in that chair. It sounds like a team that hasn't thought hard about its own work. And here's the painful part: the CFO is right.

Saying "it can't be measured" and stopping there is not a philosophy. It's an unfinished thought.

The twist: the unmeasurable core is exactly why you must measure

Now the part the romantic view gets backwards.

Yes, the heart of DevRel resists counting. Trust compounds over years while budgets judge quarters. Your best outcomes are disasters that didn't happen. The developer who saw your honest talk in 2023 and whose company signed in 2026 is a chain no attribution system connects. All true.

But that unmeasurable core doesn't survive on faith. It survives under the protection of everything you can measure. Measure the footprints of trust: who came back, who stepped up from attendee to speaker, whether an agent completes your quickstart, whether the models describe your product correctly. Call them what they are, proxies, honestly labeled. Collect the stories with the discipline of data: the person who walked up after the talk, written down, named, dated. One anecdote is a nice moment. Forty dated anecdotes in a year is evidence. And agree on the frame before any number is named: what DevRel is for, in writing, so your proxies get read in your frame instead of marketing's.

Do that, and something changes in the room. When a team shows up with honest proxies, documented stories, and an agreed frame, it earns the right to say about the rest: this part has no number, and we do it anyway, because we believe it's why the measurable things work. That sentence, from a team with a track record of honest measurement, is leadership. The same sentence from a team with nothing is a shrug with conviction.

Companies fund unmeasurable things all the time: brand, culture, judgment. But only from people who've proven they can count what's countable.

The right to be unmeasured is earned with measurement.

The one-line version

DevRel without metrics fails every time, and not because the metrics people were right about what matters. They're usually wrong about that. It fails because no numbers means someone else's numbers, no learning, and no credibility to protect the work that really can't be counted.

So measure everything that can be measured, honestly, and label the proxies as proxies. Not because the dashboard is the job. Because the dashboard is the fence around the part of the job that will never fit on one.