Before we got married, my wife and I visited Teotihuacan near Mexico City. We climbed the massive Pyramid of the Moon and had our picture taken with the even larger Pyramid of the Sun behind us. The enormous complex was built over hundreds of years by a civilization that preceded the Aztecs. Archaeological and historical evidence tell us the place where we stood was previously used for animal and human sacrifices. It's now a tourist destination.

Pyramids have been built around the world by many different cultures. Edifices constructed at great expense in order to glorify a king or deity remain impressive hundreds and even thousands of years later. In a twist of fate, however, the identity of whatever or whoever they were built to honor becomes a bit of trivia shared by tour guides. And that's assuming the knowledge has survived. Lacking a complete understanding of the cultures that produced the pyramids, we're naturally left with awe and mystery.

When founders pass a community onto the next generation, the new owners sometimes fail to understand what they are receiving. Obviously there's revenue in the form of advertising or services provided. But how a community functions and why people participate seems shrouded by a cloud of confusion. Too often people assume that since the company owns a community platform, they must enjoy the right to change the community to suit their needs. This is the source of frustration for all involved.

I cribbed many of my ideas about community platforms from Alex Komoroske's concept of platforms as gardens.¹ Communities can function like software platforms and, with a little care, they can be the foundation of a great business. This particular post is a case-study of Stack Overflow, which largely failed to match its promise as a community platform. Future articles will cover more successful platforms and my hopes for College Confidential.

Communities always form around a purpose. People (including volunteers from the community) built Stack Overflow so that there would be canonical answers to all the boring questions people kept asking on programming forums. To accomplish that goal, Stack Overflow's developers designed custom-built software that eschewed the trappings of a forum in order to encourage disciplined questions and answers.

Some founders chose to stay focused on a niche community. Stack Overflow, however, expanded into other subjects where canonical answers to questions could be useful, which is pretty much any subject at all. It renamed itself Stack Exchange and spun up hundreds of communities covering everything from writing to database management. Around the time it accepted funding from Andreessen Horowitz (and not coincidentally) the company began to lose interest those topics and narrowed its mission to software developers. It fits with an investment story of "software eating the world".

That turned out to be a pretty good investment since the company sold for $1.8 billion a few years later. But it's been a slow-rolling disaster for the Stack Exchange platform and community. Stack Exchange had a good track record of expansion, but leadership decided "Nah. We like being in the technology niche." Specialization also made Stack Overflow extremely vulnerable to technological disruptions. It might not maximize profit margins and growth, but healthy community platforms search for a wider audience.

I was the community liaison for the Documentation project and I thought it might be useful for sites other the Stack Overflow. When I pitched the idea internally, the response was tepid at best. The goal was to expand Stack Overflow. Work on other sites on the network, including technical sites like Server Fault and Unix, would not accomplish that goal. I'm not sure the outcome would have been different, but I wonder what would have happened if the project had been tested on a smaller community than Stack Overflow.

It was the same story with Jobs, Stack Overflow's career service. While it was possible to find job listings outside of the narrow "developer" focus, it always felt like an oversight. For instance, the line between developer and data scientist can be fuzzy, so sometimes a data scientist job listing would slip in. There was no consideration given to integrating the job board with Cross Validated, Stack Exchange's statistics community. Stack Overflow was king, so the scope of the Jobs service never had the chance to expand to mathematicians, engineers, astronomers, sound designers and so on.

When Jobs was shut down, the reasoning was:

Developers, as you all know, don’t have a hard time finding job opportunities. The problem is often finding the right opportunity and job boards and sourcing are ineffective solutions. The effort it would take us to truly differentiate in this space is not one we could justify.

From one perspective this is absolutely correct. I even argued that cutting Jobs would "result in fewer distractions from improving the Stack Overflow platform". What I didn't understand at the time was Jobs made Stack Overflow a better platform. The existence of job listings service offers a small, but meaningful incentive to stay active on Stack Overflow in the off chance of landing a better programming job in the future.

When I lost my job, I found LinkedIn almost worthless because it has all the jobs.² Much better were niche job listings attached to community management communities. They don't really even compete with general job boards because the listings are just posts in Slack. Everyone involved is part of the same community already so there's a much greater chance of like-minded employers and applicants finding each other. Stack Overflow Jobs never needed to compete on features. It could afford to be slightly wonky as long as it connected the types of people who are attracted to Stack Overflow with the types of organizations that value those people.

I could go on to talk about how long it took us to figure out organizations would gladly pay for what eventually became Enterprise and Teams. Since the company's vision didn't include expanding in that direction, we didn't see it as a serious possibility. Now Stack Overflow makes something like half its revenue from Teams. Chat was another missed opportunity. When I first join the company, it was assumed there was no market for a chat service. After Slack proved that assumption wrong, the excuse was that Slack had already won the chat market and who wants to play second fiddle?

Community platforms strive to meet the needs of their community even if it's not the most profitable use of resources. Platforms must be farsighted to see better things. Many companies, especially those driven by venture capital, can't afford to think years ahead. Shortsighted people don't build pyramids.

Imagine what a priest who practiced sacrifices from the top of the Pyramid of the Moon would think about my wife and I enjoying an afternoon on that very spot. He'd probably say we missed the point. maybe I'm as out-of-touch with modern life as that priest when I write about what has become of Stack Overflow. It's hard to argue with a company that was purchased for nearly $2 billion³ just 12 years after it started.

And yet that financial success came at a cost. First it cost nearly two hundred Stack Exchange communities features that were limited to Stack Overflow alone. Second it removed financial diversification that might have prevented the disastrous attempt to expand into AI. Third it cost community trust that seems to be running critically low. Finally, it prevented Stack Overflow and Stack Exchange from becoming an enduring platform for so many communities.

To understand what we missed out on, I'll need to write about a successful community platform: BoardGameGeek.

---

1. I believe I independently analogized communities to rose gardens. ↩

2. Turns out many "community manager" jobs are basically property management. To filter those out (and because I plan to never have a daily commute again) I only looked for "remote" positions. ↩

3. Disclosure: I got $237,549.76 from selling my Stack Exchange options. ↩