YOUR KNOWLEDGEBASE IS JUST A COMMUNITY WITH GOOD FILING

Week 9 · Blog Post 15 · Web 2.0-Based Learning and Performance · Summer 2026

AI-Generated Image | ChatGPT

Listen to this post:

7:05

The archive builds itself. That's the problem.

I have a folder. You have a folder. Every instructional designer I know has some version of it — a bookmarks bar, a Notion page, a Slack channel named #resources where fifteen people have posted and two hundred have scrolled past. We call these knowledgebases. Most of them are landfills with a search bar.

What would it actually take to build a knowledgebase on social media — not a library I maintain alone and nobody visits, but a living archive a community fills, uses, and keeps honest?

I went looking, and the first thing I read knocked the legs out from under my assumptions. Hu and colleagues studied early-career teachers curating math resources on Pinterest, treating the platform as a "living data archive" of what teachers actually value when nobody's grading them. The finding: teachers overwhelmingly pinned resources that asked students to perform standard algorithms — low cognitive demand, high visual appeal. The archive built itself beautifully. What it archived was the easy stuff.

That reframed the whole problem for me. The failure mode of a crowd-built knowledgebase isn't emptiness. It's mediocrity at volume.

PERSONALITY IS INFRASTRUCTURE

The instinct, when a shared repository stalls out, is to fix the container — better tags, better folders, a better platform. Young and Rossmann tried something else. Their academic library shifted its social media from broadcast marketing to deliberate community-building: hand-crafted posts with an actual voice, real replies, personality treated as a design requirement rather than a lapse in professionalism. Their Twitter following doubled in a year and interactions climbed 275 percent.

Note the order of operations there, because I keep getting it backwards. They didn't build a resource hub and then go looking for engagement. They built a place where people were already talking, and the resources accrued to it.

A repository is a filing cabinet. A repository with a conversation happening in front of it is a commons. Same files. Wildly different object.

Nobody shares into a void — or a courtroom

Which leaves the hardest part: getting people to contribute things that are actually worth keeping.

Mettler's study of a professional social network built for hospitals is the most useful thing I've read on this, mostly because it's honest about how badly it can go. Clinicians didn't withhold knowledge because they were selfish. They withheld it because of professional identity, occupational hierarchy, and a well-founded suspicion about who was watching. Privacy mechanisms and clear access rules increased willingness to participate — a finding that should be tattooed somewhere visible. Openness isn't the thing that makes people share. Safety is.

He also found that a small core of consistent contributors could generate enough value to pull in a much larger reluctant group. Not a growth hack. A gravity well.

Muljana and Luo's interviews with 17 instructional designers filled in the other half. Their participants used social media as a just-in-time lifeline — stuck on an Articulate problem, they'd surface it to a professional community, get unstuck, tweak the approach, and then, in at least one case, turn around and make a video walking others through the same problem. Their reported challenges were exactly what you'd guess: information overload, and ethical and privacy concerns.

But look at the shape of that sharing behavior. The contribution wasn't "here is a resource I found." It was "here is the thing that got me unstuck, and here's how." That's a fundamentally different artifact. It arrives pre-vetted by someone who had actual skin in it.

WHAT I'D ACTUALLY BUILD

So if I'm designing this — and I am, for my own field-notes work — I stop asking people to share resources. Asking for resources gets you Pinterest: volume, polish, no teeth.

Instead I'd ask for problems solved. Every contribution has to carry an annotation: what were you trying to do, what did this help with, and where did it fall short? That last clause does more quality control than any voting system I've ever seen, because it forces the contributor to have used the thing.

I'd seed it with a handful of people who'll show up whether or not anyone's watching, and I'd let the lurkers lurk. I'd credit contributors by name in the archive itself, since professional identity is the actual currency here. I'd prune out loud — deleting a stale entry publicly and saying why is the strongest possible signal that this is a maintained space and not a landfill. And I'd resist upvotes, because a metric that rewards popularity in a community of professionals will reliably surface the prettiest worksheet in the room.

I'm still not sure whether that's enough. Every one of those moves adds friction, and friction is the thing that kills participation. It might be that a knowledgebase good enough to trust is, by definition, a knowledgebase most people won't contribute to — and that the two hundred silent scrollers in the #resources channel aren't a failure of design but the expected ratio.

Here's where I'd genuinely like some help: if you've built or belonged to a shared knowledge space that stayed good — where the quality held as it grew — what kept it honest? Was it a person, a rule, or just a small enough room?

---

Resources