The short version
- The common complaint in growing companies is not that the document was missing. It is not knowing who to ask.
- Org charts answer who reports to whom, which is almost never what you needed.
- Build a who knows what list from questions people actually asked, not from a skills survey. Surveys produce aspirations.
- Once the list exists, the next problem appears: the right person is now identified and unavailable.
Ask people in a company of forty what slows them down and you rarely hear about missing documents. You hear a version of this:
"I did not know who to ask, so I asked in the general channel, and then I asked three people, and the third one knew."
The complaint, stated properly
There are two different failures hiding under "I could not find it", and they need different fixes.
The first is that the information exists in writing and is badly organised. Search and structure help.
The second is that the information exists only in a person, and you do not know which person. No search index helps, because there is nothing indexed. What you need is a map from topics to names.
In companies under a hundred people, the second failure is much more common, and it appears suddenly. Below about twenty five, everyone knows who does what by osmosis. Above it, new joiners arrive faster than the osmosis works, and the informal map stops being shared.
Why expertise directories fail
The standard response is a directory: a page per person with their skills and areas. Three reasons it does not last.
- It is built from self description. People list what they want to be known for, not what they are actually asked about. Those differ, and the second is what you need.
- It uses the wrong vocabulary. Someone lists "data infrastructure". The person searching types "why is the revenue number different in two places". No match.
- It goes stale silently. People change areas and nobody updates a page, so the directory sends you to someone who last touched that system two years ago.
The org chart has the same problem in a purer form. It answers reporting lines, and almost nobody's question is about reporting lines.
The low tech version that works
Build it backwards, from real questions. One shared page, three columns, maintained by the questions themselves.
| If you need to know about | Ask | Second name |
|---|---|---|
| Why the revenue number differs between the dashboard and the invoice | Marta | Kofi |
| Anything about the German contracts | Ana | Ben |
| Getting a deploy out on a Friday | Lena | Kofi |
Two rules make it survive. Phrase the left column as the question someone would actually type, in their words, including the confused phrasing. And add a row every time someone had to ask around, which makes the list grow from real friction rather than from an exercise.
The second name column does two jobs: it gives people an alternative when the first is away, and it makes your key person dependencies visible the moment a column is empty. See how to find key person dependency.
The second problem you have just created
Solve the routing and you immediately meet the next wall. Now everyone knows Marta is the person for the revenue question. Marta gets asked constantly, and Marta is in a different time zone, or in a meeting, or on leave.
The directory has converted "I do not know who to ask" into "I know exactly who to ask and they are not there", which is progress and is not an answer.
This is what StandIn is built for. Each day Marta spends about ninety seconds on a brief: what moved, what is open, what is blocked, what is next, mostly drafted from the work that already happened. When someone needs the revenue answer and Marta is not available, they ask and get an answer from that record, in her words, with a source under it. It never guesses, and when the answer is not in what she wrote it says so and names who to ask, which is when the second name column earns its place. The routing problem and the availability problem get solved by the same page. See how it works and answering the same questions at work.
Common Questions
At what size do we need this?
Around twenty five to thirty people, or earlier if the team is distributed, because osmosis needs shared rooms and shared hours to work at all.
Who maintains the list?
Everyone adds rows, one person prunes monthly. Pruning matters more than adding. A list with a wrong name in it will be abandoned after two bad experiences.
Is a general help channel a reasonable substitute?
It is better than nothing and it costs everyone's attention, because a question broadcast to forty people interrupts forty people to reach one. Use it to find the name, then add the name to the list so nobody has to broadcast twice.
What about people who do not want to be listed?
Usually they are already the most asked person and are protecting themselves, which is reasonable. Listing a second name alongside them helps more than leaving them off, which just moves the asking into private messages.
When you're off, your StandIn is on.
It answers your teammates' questions from work you've already done, in your words, with a source under every answer.