How small HOAs lose institutional knowledge

Every two years a strata board turns over and the building forgets things it spent ten years learning. Why it happens, and what to do about it without turning into a paperwork factory.

The most expensive thing in a small building isn’t the boiler or the roof. It’s the institutional knowledge of how the building actually works — and nobody puts it on the balance sheet.

A strata council turns over every one to three years. The members who knew that the riser on the north side runs cold in the winter, that the Otis tech named Tony is the one who actually understands the elevator, that the 2017 leak originated three floors up and not where the water finally appeared — those members leave, and almost none of what they knew is written down anywhere a successor can find it.

This isn’t a discipline problem. Boards are volunteers. Volunteers write things down when writing things down is easier than not writing them down. For most strata boards today, “writing something down” means a group email chain that disappears into people’s inboxes, or a one-line entry in a paper logbook in the storage room, or a Dropbox folder named “BUILDING STUFF” that grows by sediment and is searchable only in the lossiest sense.

Here’s what gets lost when a board rolls over.

What gets lost

Who’s good. The plumber you’ve had three times who actually knows where the shutoffs are. The roofer who, unlike the other roofer, showed up on time and didn’t try to upsell you on a re-flash you didn’t need. The elevator tech whose home cell you have because the elevator company’s dispatcher is useless after 6pm.

Why we did it that way. Five years ago the board decided to do the boiler runs in October instead of November because of a specific pressure issue that surfaced once. The current board has no idea, and will quite reasonably move it back to November this year, and then have the same pressure issue, and call the same plumber, and pay him to diagnose a problem someone diagnosed in 2021.

Who’s allergic to what. The unit on the second floor whose owner left town three years ago and rents to a tenant who didn’t know the sub-floor was being replaced when she signed the lease, and now her two cats are upstairs at 2am because the sub-floor work was scheduled during the day. Three boards ago someone had a system for this.

What was tried and rejected. A previous board got a quote for a new front entry system. They rejected it because the vendor wouldn’t sell parts separately and they’d be locked in for a decade. The current board has no memory of this and is going to get the same quote from the same vendor next month.

These aren’t dramatic losses. They’re slow ones, paid for in re-investigation fees and avoidable mistakes, $300 at a time, spread over years.

The temptation: build a paperwork factory

The first instinct is always to formalise. Adopt a procedure manual, require the secretary to take minutes, build a wiki, set up Notion or SharePoint or Confluence or whatever the previous job of one of the board members used. This works for about three months. Then it stops working because:

  • The structure was designed by one volunteer and is mysterious to every successor.
  • Nobody can remember whether a particular bit of knowledge belongs in the wiki, the meeting minutes, or the document store.
  • The volunteer who set it up rolls off, and the new volunteer doesn’t want to maintain someone else’s filing system.

The pattern is reliable: the more elaborate the system, the faster it becomes a tombstone — a structure that records the first board’s diligence and the eventual abandonment.

What actually works

Three things. None of them require a new role or a new procedure.

One stream, not many places. Every observation, decision, vendor visit, and incident lands in the same log, with a date and an author and an optional photo. Not the meeting minutes, not the wiki, not the shared drive — the log. Other surfaces are views into the log, not separate things.

Plain language, in the moment. A 30-second log entry written from a phone in the boiler room (“Tony from Otis serviced the south elevator. Replaced the door rollers. He said it’ll last another two years before the controller needs upgrading.”) is worth more than a beautifully formatted incident report written six weeks later.

Searchable like a text message. If the way you find old context is “hunt through twelve email threads”, you’re not going to do it. If typing “boiler” into a search bar returns the seven things that happened to the boiler in the last three years, you’ll use it.

These aren’t novel ideas. They’re the same ideas that made chat-based work tools (Slack, Teams) successful in companies: a single stream, low-friction capture, real search. The strata equivalent just hasn’t existed.

The five things a board actually needs

After watching enough buildings, here’s the working theory of the minimum a board needs to remember:

  1. What happened. Every observation, in one line, with a date.
  2. The things that did it. Equipment with serials and history.
  3. The people who do it. Vendors with contact info and history.
  4. The docs. Manuals, scoped to the equipment they document.
  5. The procedures. How this building does the recurring stuff.

That’s it. Five surfaces, all built on top of (1). Everything else — checklists, reminders, announcements, the AGM packet — is a view or a helper on top. The substance is the log.

If you’re on a board this year and you’re worried about the knowledge walking out the door next year: start writing things down. In one place. Today, before the boiler service in October. The format barely matters; the streak matters. Boards that have been keeping a log for two years are remarkably hard to surprise.


We built BuildingHQ because we wanted exactly this for a building one of us lived in. The Basic tier is free forever for any building. If you’re considering a switch, say hi.