DevRel as a Service
Unbundling The Full Stack DevRel, and building Infrastructure for scaling Developer Relations beyond "the DevRel team"
Search for a command to run...
Unbundling The Full Stack DevRel, and building Infrastructure for scaling Developer Relations beyond "the DevRel team"
This idea of devrel as an enablement role makes total sense. Rank and file devs are the true subject matter experts so its their voices we want to hear.
However, devs usually have enough on their plate with their job, and generally prefer to code rather than create content, so it would need to be incentivized by management.
Another challenge is that content creation is really a skill on its own and requires confidence. Even with enablement from a devrel, companies would have to invest in educating their devs on how to blog, speak etc.
"incentivised by management" only goes so far tho. we can heavily encourage, set time apart, offer resources and bonuses, but ultimately the person has to be motivated or else the content wont be good.
some companies have cultures where they hire people who are intrinsically motivated. those people naturally want to share, for bragging rights, for sharing what they encountered, or because they feel the strong connection to building tech credentials for the company (for revenue or hiring puproses). thats the ideal
once people require extrinsic motivation to do content then you kind of already lost.
swyx here: I recently hosted the inaugural AI Engineer Europe and Matt Pocock was a standout speaker, crossing 1m views across talks and workshops in <2 weeks. I asked him to write a post on how he sp
Thariq Shihipar's (Anthropic, Claude Code) Devrel/Writing Process — building implicit knowledge and then making it explicit.
It was a pleasure to rejoin Jack Bridger for a third time on Scaling Devtools, where I usually try to cram as much devrel metathoughts per minute as I can. Nominally the pod was about how DevRel is So Back, but informally I’ve been sharing a bunch of...
How to create extraordinary dev videos by going against the grain

and we're not sure why
Translations: Japanese
One of my favorite jokes about DevRel is that the First Commandment of DevRel seems to be, upon getting the job, you must then immediately turn around and create a self congratulatory blogpost/podcast/Twitter Space about "How To Get Into DevRel". Surely everyone wants to be in DevRel, yes? By 2030 no engineers will be left, tech will just be all devrels devrelling their $19 ebooks and $10 Substacks and $4.99 Super Follows to each other and we shall have achieved Universal Basic Devrel.
On a more serious note, I was awakened to the limits of DevRel by an anecdote from a previous manager, who was told we couldn't have more headcount because "hey, you're already 10% of the company! Do more with what you have!"
Which is a very good point - there's probably an optimal proportion of devrel team sizing, similar to the now-accepted wisdom that there should be between 5-8 engineers to every engineering manager, with parallel product management, product marketing, and product design (and, if you're really fancy, product analytics) people attached in various ratios.
CommonRoom's DevRel Comp Report noted the surprisingly wide distribution of devrel team sizes:

But my ex-manager's manager's point stands; at some point, DevRel needs to stop growing.
Sidenote: In fact, you probably want DevRel headcount to scale sublinearly to both the number of your engineers at the company and to the number of customers, if you want any hope of improving profit margins in future.
Contrast this with the other empirical observation that DevRel marginal productivity declines dramatically as teams grow; the first person does a lot, the 10th person tends to do more of the same of what the first 9 did, by the time the 50th person comes along you barely know what they did all year. In other words, DevRel average productivity (however measured... that's a topic for another day) tends to decline with increasing team size, at the exact moment when you need it to go up (to scale headcount sublinearly).
So. How do you scale DevRel?
Note: this is a rhetorical question that has a longer answer than can fit in this essay; for other scaling dimensions, please see my discussion on InfraEng on going from singleton programs to language/geo/vertical split programs.
One of my favorite examples to trot out is that GitHub, Snowflake and Stripe had no DevRel for their first few years of existence. They did fine, didn't they? Sure, for every one of those examples, you can find a Twilio or a Plaid where the founders strongly attribute their success to establishing a good devrel/evangelism program. But for the places without, DevRel basically was everyone else's job. Stripe was famous for the Collison Installation where Patrick and John would show up at prospective customers' offices and do the integration work for them, removing all objections and getting hands on feedback. TPW and Chris Wanstrath spoke at a bunch of Ruby meetups to get GitHub going in the early days. And so on.
To scale DevRel, DevRel should turn from an Ownership role (the buck stops here, we do everything related to DevRel) to an Enablement role (we enable everyone else to do DevRel!). I started writing about this dichotomy last year).
This seems like a shirking of responsibility until you consider other "enablement" roles, like HR. Consider that HR departments don’t actually do the hiring for individual teams. Instead, most companies have individual line managers become hiring managers, who basically have the final say. Instead of owning hiring, HR departments enable hiring managers to do hiring as part of their jobs, by building infrastructure for as much of the job as possible - buying ATS software, determining benefits, sourcing candidates, organizing onboarding, and so on.
So - to scale DevRel, we need to identify "devrel infrastructure" and provide it as a "service" to the rest of the company, and potentially our champions in the community.
The classic "it's just me" First DevRel Hire (our previous thoughts here) tends to be "full stack" - writing, speaking, workshopping, coding, interviewing, editing, planning, they do it all. This tends to be the setup for teams up to about 8 or more people - each devrel tends to be a mostly autonomous lone wolf, with perhaps natural affinities to certain user personas/communities, and natural strengths to certain kinds of content channels or deliverables within the company, but usually taking their projects from ideation to production to publication to syndication all the way from start to finish. The expectation is that if you are a good DevRel, you will have a polymathic command of all skills from editing YouTube videos to organizing the company conference to writing jokey social media memes for Fellow Kids.
But of course, no media or professional education organization operates like this, and as companies realize that they're really building a media-company-within-a-company, specialized roles start to emerge.
The simplest way I've found to think about this is by analogy to the division of labor found in the bank trading floors of my past: Front Office, Middle Office, Back Office (readers of my book will find this familiar).
The usual power dynamics is that Front Office tends to be more glamorous and better paid, and status decreases the further back you go, but when things go wrong - when a mistake has been made or a limit violated, then the power inverts, and the Back/Middle Office can really hold the Front Office accountable for their mistakes.
You could adapt this 3 tier architecture for the DevRel responsibilities:
The details may differ based on your chosen medium or usecase, but once you've split out the "full stack" of DevRel like that, you can really see that the "Front Office" can be expanded to more than just the people with "Dev Advocate" in their job title. If you build up the right DevRel Infrastructure, you make it easy to provide "DevRel as a Service" to the rest of the company, and even enthusiastic volunteers in your community.
This is an incomplete list, but here are some DevRel Infra items that I have maintained in the past:
What else makes sense to add to this "infrastructure" list?