“What makes a good DevRel?” usually gets a personality answer: curious, empathetic, good on stage, writes well, ships code, loves community. Those traits help. They are not a definition of the job.

Good DevRel is a system for helping developers succeed with a technology — and helping the organization that funds that work survive contact with reality. The system looks different if you sit inside a product company versus an open source project, but the craft underneath is shared.

This essay is a practical bar. Use it to assess yourself, hire better, or decide whether your project actually needs “DevRel” or just better docs and a release process.

First: name the context

Industry language helps here. Developer First (B2D) companies sell primarily to developers — Twilio, Stripe, MongoDB, Unity. Developer Plus companies sell to businesses or consumers and also offer APIs, SDKs, or platforms (Apple, Salesforce, banks with developer portals). Open source projects may power either, or stand alone as community infrastructure with a thin commercial layer.

Context changes your scoreboard, your politics, and how much time you spend educating internal stakeholders versus external developers. Ignoring context is how good people get rated on the wrong game.

Product company DevRel Open source DevRel / maintainer-as-DevRel
Primary job Adoption + successful usage of a product Healthy contributors + successful users of a project
North-star examples Activation, monthly active developers, PQLs, expansion influence Active contributors, issue quality, release adoption, ecosystem integrations
Main risk Becoming expensive marketing cosplay Burning maintainers / extractive “community”
Credibility source Product truth + peer trust Contribution history + fair process
Internal work Heavy (Sales, Product, Marketing alignment) Heavy (maintainers, foundation/company, CODEOWNERS)

The shared craft: five capabilities that always show up

Whether you wear a company badge or a maintainer hat, the people I respect share these capabilities. Not “soft skills.” Capabilities you can practice and evidence.

1. Technical empathy with receipts

You can reproduce a developer’s failure mode. You know what “works on my machine” actually means in your stack. You do not hand-wave around setup pain. Empathy without technical depth becomes cheerleading. Technical depth without empathy becomes gatekeeping.

Evidence: a debug session write-up, a fixed sample app, a docs PR that removed three common errors, a workshop where 80%+ of attendees finished the path.

2. Teaching that creates independence

Good DevRel reduces the need for DevRel. Your content, samples, and office hours should make the next developer faster without you in the room. That is the difference between a hero support act and a scalable program.

Prefer paths over posts. Prefer runnable examples over abstract architecture diagrams. Prefer “here is the failure and the fix” over “here is why our product is exciting.”

3. Two-way advocacy

Outbound: put the technology in front of the right developers, in the right depth, at the right time. Inbound: carry developer reality into the people who can change the product or project.

Most teams over-index outbound because it is public. The mature ones protect inbound bandwidth and track whether feedback ships. If nothing ever ships, you do not have a feedback loop — you have a suggestion box.

4. Taste for channels that match the half-life

Not every message belongs on every platform. Social posts die in hours. Blog and docs compound for years. Workshops create depth. Conferences create proof and distribution if the talk is good enough to watch later. Good DevRels choose channels with intent instead of filling a content calendar for its own sake.

5. Measurement without self-betrayal

You can translate work into language leadership understands without turning relationships into pure pipeline. That means keystone metrics, leading indicators (time-to-first-value, activation, contributor first PR), and qualitative stories that explain the numbers — not vanity dashboards that impress nobody who funds you.

What “good” looks like inside a product company

Company DevRel fails when it becomes either a roadshow or a support ticket sink. It succeeds when it tightens the loop between product quality, developer activation, and durable trust.

A good company DevRel program usually includes:

Company bar check Can you point to developers who reached first value faster because of your work, and to one product decision that changed because you brought signal in? If neither, you are performing DevRel, not practicing it.

Company-specific skills that separate seniors

What “good” looks like in open source

Open source DevRel (or maintainers doing DevRel work without the title) fails when “community” means extraction: more users, more issues, more drive-by demands, same number of burned-out maintainers. It succeeds when more people can contribute safely and users can succeed without opening a crisis issue.

A good OSS DevRel posture usually includes:

OSS bar check Did someone become a regular contributor because of how you designed the on-ramp? Did a release land with fewer “it broke for me” threads because of how you communicated? Stars do not answer those questions.

OSS-specific skills that separate seniors

The anti-patterns (both worlds)

Anti-pattern What it looks like What good replaces it with
Conference tourism Travel-heavy calendar, thin follow-up 1–2 high-quality talks + recorded assets + local depth
Content mill Volume without paths or maintenance Fewer pieces, longer half-life, kept up to date
Support cosplay DevRel as free L2 forever Patterns → docs/product fixes; escalation rules
Metric theater Stars, members, impressions only Activation, contribution quality, DRQLs, retention
One-way evangelism All outbound, no product loop Tracked feedback that ships

A simple self-assessment

Score yourself 1–5 on each. Be cruel. Inflation helps no one.

  1. I can take a new developer from zero to first success on our golden path without live screenshare heroics.
  2. I ship teaching assets that stay correct for more than one release cycle.
  3. I maintain a written list of top developer pains with evidence, not vibes.
  4. At least one product/project decision per quarter moved because of signal I brought.
  5. I can explain my work to a CFO-level audience without only citing impressions.
  6. Developers trust me as a peer, not a billboard.
  7. I protect deep work and have explicit non-goals for the quarter.
  8. (OSS) New contributors can land a meaningful first PR without private DMs to me.
  9. (Company) Sales/Product know what I own and what I refuse.
  10. I am getting better at one specialty (content, community, product/DX, education) instead of being mediocre at twelve.

Under 30: you are early — pick one path and go deep. 30–40: solid operator. 40+: you should be mentoring others and writing the standards your org lacks.

What it takes (the short version)

Good DevRel is not “being social online about technology.” It is the disciplined practice of making developers successful at scale, with enough organizational skill that the practice still exists next budget cycle.

In a company, that means activation, trust, and product loops under a clear charter. In open source, that means contribution health, user success, and maintainer sustainability under fair process. The costume changes. The craft does not.

If you want the field to be taken seriously, raise the bar past personality. Hire for systems. Practice like the developers you serve are watching — because they are.