Who is who
Finding the People You Will Actually Need
Three methods find the people who matter, none of them is the organisation chart, and all three work in the first fortnight.
What the people on this page had to find out, and where from
Who signs off a schema change
you asked4Three people were asked before the right one
That two people on the team do not speak to each other
worked out11Worked out from who was never on the same call
two things you had to know · nobody told you any of them · the median arrived in week 11 · all two you went and got
Two facts about the shape of a team, neither of which appears on any chart and neither of which anybody would have said out loud.
When a new employee is learning an unfamiliar workflow, managers can see the tool in practice to understand where time is going before assuming that a slow start is an effort problem.
For a separate career or employment reference, consult USCIS.
Four occasions of not knowing who to ask, and what each cost
You could not tell who owned the pricing table
6 hanswer: 6 minYou waited because the org chart did not say and you did not want to guess wrong. Two of the three people you asked did not know either
Nobody had given you access to the billing dashboard
5 hanswer: 3 minYou waited because you did not want to look as though you were chasing. The owner turned out to be in another division
You did not know who approves a schema change
4 hanswer: 8 minYou waited because asking felt like admitting you had not read something. Found on the fourth attempt
You could not find out who had written the import pipeline
2 hanswer: 2 minYou waited because it seemed like something you should be able to look up. It was in the commit log and took two minutes once you thought of it
four occasions · 17 hours spent stuck · 19 minutes of answers · the longest single wait was 6 hours
Seventeen hours across four occasions, and the whole of it was one question asked of the wrong people or not at all. Nineteen minutes of answers. One person's own first month.
The people you will need are discoverable in your first fortnight, by three methods, none of which involves asking anybody for an introduction.
One: read the history
Whatever your work touches has a record of who touched it before you.
Commit logs, document revision history, the last twenty messages in a channel, the signatures on previous approvals, the names in old tickets. Thirty minutes of this produces a list of six or eight names that is more accurate than anything anybody will tell you, because it records what people have actually done rather than what they are nominally responsible for.
Do this on your first quiet afternoon. It is the best available use of a week with no access to anything.
Two: watch who answers
In any team channel, over a fortnight, four or five people answer most of the questions.
Write their names down. These are the people who will unblock you, and their position on the chart is nearly irrelevant: the most useful person in a team of thirty is frequently somebody two years in with no title at all.
Notice also what each of them answers about. People's answers cluster, and after two weeks you can usually say which two topics each name belongs to.
Three: ask, in the form that has an answer
"Who owns X?" is a question many people cannot answer honestly, so they guess, and you end up where the panel above ends up: four attempts and seventeen hours.
Three better forms:
Who was the last person to do one of these?
If I get this wrong, who finds out about it?
Who would be annoyed if I changed this without telling them?
All three have findable answers. The second and third are particularly good, because responsibility in practice is defined by consequence rather than by title, and people know the consequences.
The map you are building
| Name | What they actually know | How you found them |
|---|---|---|
| The thing you will ask them about | History, channel, or asked |
Two columns and a note on provenance. The third column matters more than it looks: a name you got from a commit log is reliable, and a name somebody guessed at is a lead.
Asking for twenty minutes
Once you have six or eight names, ask the three or four most relevant for twenty minutes each.
This feels presumptuous and it is not. New people are allowed to do this, it is the one thing a new person can do that nobody questions, and the window for it closes: in month four asking a stranger for twenty minutes requires a reason, and in week three it does not.
I have just joined and I will be working on the pricing side. Could I have twenty minutes some time this week to understand how your bit fits in? No preparation needed.
"No preparation needed" is the clause that gets it accepted.
What to ask them
Three questions, and they are the same three every time.
What do you work on, in your own words? What do you wish people understood about it? And who else should I be talking to?
The third produces the next names, and after four conversations the names start repeating, which is how you know the map is roughly complete.
What to do when the trail goes cold
Occasionally the history names somebody who has left, the channel has nothing, and three people have all said they are not sure.
Two moves remain, and both work more often than they should.
Ask the question in a wider channel, naming what you have tried. "I am trying to find out who owns the pricing table — I have asked in our team channel and looked at the commit history, which points at somebody who has left. Does anybody know where this landed?"
Or ask your manager to ask. This is one of the few things a manager can do instantly that you cannot: one message from them reaches people who would not answer a stranger, and it costs them thirty seconds.
What not to do is keep looking. Past about an hour, the search has stopped being a search and become a way of avoiding asking.
Asking the question of a group
Where you have no candidate at all, a question in a wide channel works better than it looks, provided it is specific.
Does anybody know who looks after the pricing table these days? The last few changes are from somebody who has left.
The second sentence is doing the work. It explains why you are asking a room rather than a person, which otherwise reads as not having tried.
What does not work is the general form: "who should I talk to about pricing?" That is answerable by anybody with an opinion and produces three different names, which is where the four attempts in the panel above came from.