When the same question gets asked for the third time, most owners conclude that members do not read. The more useful conclusion is that the answer is not where people look for it.
"It's in the pinned post" is the community version of an excuse every software team has made: the feature exists, you just had to know the magic word. It is true, and it is not a defence.
The repeat question is data
A question asked three times is not three lazy members. It is one broken path, found independently by three people.
That makes it the cheapest research you will ever get. Nobody has to fill in a survey, nobody has to be interviewed — the community is telling you, unprompted and for free, exactly which thing it cannot locate. The owners who treat this as a nuisance are throwing away the only unsolicited usability testing they will ever receive.
Keep a list for two weeks. Every question you answer that you have answered before, write down in one line. At the end you will have a ranked list of everything wrong with your community's findability, in priority order, written by the people affected.
Where things go is not where people look
These are two different problems and almost every owner solves only the first.
Structure is a filing decision: which space does this belong in, should it be public, when do we archive it. It is the owner's mental model, and a good one genuinely helps. But a member arriving with a question in their head is not browsing your filing system. They are doing one specific thing, and the question is whether that thing terminates in an answer.
A perfectly structured community can be completely unnavigable, because structure is optimised for putting things away and findability is optimised for taking things out. The second does not follow from the first.
What members are actually doing
There are only about four, and they want different things from you.
| What they are doing | What they need | What usually happens |
|---|---|---|
| Looking for a specific answer they know exists | Search that matches their words | They type their words; your content uses yours |
| Wondering what this place is for | One obvious starting point | They land in a feed mid-conversation |
| Coming back to a thing they saw last week | A way back that is not scrolling | It is four hundred posts down |
| Checking whether it is worth asking | Evidence somebody answers | They see an unanswered question and leave |
Only the first of these is a search problem. The other three are answered by what is visible without searching, which is why "just use search" is never the whole fix.
Pinning is not a filing system
Pinning is the tool every owner reaches for, and it stops working at about item four.
A pin says "this matters more than the feed." Five pins say nothing at all, because the member now has to read five things to find out which one is theirs, which is the job they were trying to avoid. Worse, pins never expire on their own, so unless you unpin things when their deadline passes, the pinned set slowly becomes an archaeology of what mattered in March.
Treat pins as a small fixed budget — two or three, permanently. Adding a fourth means deciding which existing one has stopped earning its place. If that decision feels impossible, the thing you wanted to pin probably belongs somewhere permanent instead.
Answer where the question gets asked
The instinct is to send people to the right place. The better move is usually to bring the answer to the wrong one.
Someone asks in chat a thing that is covered in the resources space. "It's in resources" costs them a search and teaches them that asking has a small penalty attached. Answering in two lines and then adding "I'll put this in the space so it's easier to find next time" costs you thirty seconds and produces a better indexed answer than the original.
If you find yourself saying "as I mentioned", stop and move the thing. That phrase is a reliable signal that you are defending your filing system instead of fixing it, and members hear it as mild criticism even when none is meant.
Name things the way people ask for them
Most findability failures are vocabulary failures, and they are invisible to the person who chose the words.
You named the space Resources; they are looking for "templates". You wrote a post called "Q3 Programme Update"; they are searching for "when is the next call". Owners name things from the inside, using the word that made sense while creating it, and then wonder why nobody finds it. The member's word is the right word, always, because they are the one typing.
The fix is to use both. Put their phrasing in the title and yours in the body, or simply say the thing twice in different words — search does not mind repetition, and the member scanning a list of results does not either.
Everything ages out, so decide what comes back
A feed is a conveyor belt. Anything you do not deliberately re-surface is gone, and "gone" arrives faster than owners expect — usually within a week.
This is the part that quietly breaks as a community grows: the answers that made the place valuable in month three are still there, still correct, and functionally invisible by month nine. Nobody deleted them. The feed simply moved on, and new members have no idea they exist — which is the real argument for making the archive answerable rather than merely stored.
Pick a small number of evergreen things and re-surface them on purpose, on a schedule, without apologising for the repetition. Newcomers have never seen them, and the people who have will scroll past in half a second. The embarrassment you feel about repeating yourself is a cost only you are paying.
What the software can and cannot tell you
This is the one area where you do not have to guess, and almost nobody looks.
Mateflow has search for members — across the community and within a single space — and, on the admin side, an Analytics → Search screen whose own subtitle is "What members search for, and what they can't find". It shows total searches, unique searchers, an average per searcher and an overall search-to-click rate, plus a Top Queries table with a click rate against each term, and a Zero-Result Queries list labelled, plainly, "Searches that returned nothing — content gaps to fill", with a zero-result rate alongside it. When that list is empty it says "No zero-result searches — nice coverage."
Two of those numbers are worth more than the rest. Zero-result queries are your members writing your content backlog for you, in their own vocabulary. And a top query with a low click rate is worse than a zero-result one — it means they searched, got results, and none of them looked like the answer, which is a naming problem rather than a missing-content problem. The honest limit is that it only records searches. The member who never thought to search, or who asked in chat instead, or who quietly left, is not in this data at all — so treat it as the floor of your findability problem rather than the measure of it.
The five-minute test
Ask somebody who joined last week to find three specific things while you watch and say nothing.
Pick things you are confident are findable: the schedule for the next call, the answer to your single most repeated question, and one older post you are proud of. Then stay quiet. The urge to help will be physical, and giving in destroys the only data you were going to get.
Watch where they go first. Almost always it is not where you assumed, and that single observation tends to be worth more than the rest of this post — because it is the difference between the map in your head and the one they are actually using.
Three things not to do
Do not answer a findability problem with more content. Writing a second FAQ because nobody read the first one gives members two places to not find the answer, and gives you two things to keep current.
Do not build a "start here" post and consider it solved. It helps, and it decays — it was written for the community you had, and nobody re-reads it. Onboarding is a flow, not a document.
Do not reorganise everything at once. A big restructure resets everybody's mental map, including the members who had finally learned where things were. Fix the top three repeated questions and stop.
When it is genuinely a content gap
Sometimes the thing really is missing, and it is worth being able to tell the difference.
The signal is consistency: if people search for it, ask for it, and nobody can point them anywhere, you have a content gap rather than a discoverability problem, and the fix is to write the thing once, properly, somewhere permanent. The two failures feel identical from the owner's chair — both arrive as repeat questions — but they need opposite responses, and guessing wrong wastes a month.
Check before you write. If an answer already exists and is not being found, writing a better one does not help; it adds a second unfindable answer, and now they disagree with each other.
The bottom line
Findability is not a feature of your platform. It is the gap between the words you used and the words your members would have used, measured in questions you have to answer twice.
So keep the list of repeat questions, cap your pins, answer where people ask and then move the answer, name things in your members' vocabulary rather than your own, and re-surface the evergreen things without embarrassment. Then go and read what people searched for and did not find — it is the shortest honest list of what your community is missing, and it costs nothing to look.