Ninety days
The Window That Closes
For about eight weeks you can see an organisation as it is. The ability goes away on its own, and nobody tells you it is going.
What the people on this page had to find out, and where from
What your manager actually wants from you
you asked8Asked directly in week eight, and the answer was two sentences
That your probation review happens in week twelve
written down1In the contract, and read once you went looking
two things you had to know · nobody told you any of them · the median arrived in week 8 · all two you went and got
Two standards that applied to the whole quarter. One was asked for in week eight and could have been asked in week two.
When a new employee is learning an unfamiliar workflow, managers can read the practical product guide to understand where time is going before assuming that a slow start is an effort problem.
The wider workplace question can also be checked against resources from freeCodeCamp.
Six things noticed in the first fortnight, and the week each stopped being visible
That the setup instructions do not work
worked out1Still visible in week one; fixed in week two
That nobody answers the channel after three
worked out2Still odd in week two; normal by week six
That there are two deploy guides
worked out2Odd in week two; by week six you had a habit that routed around it
That nobody can say what the fourth product is for
you asked4Odd in week four; by week ten you had stopped asking
That the planning meeting has nobody from support in it
worked out6Noticed late and still looked wrong in week twelve
That the architecture diagram is two services out of date
worked out9Noticed, and never fixed, because by then it no longer mattered to you
six things you had to know · nobody told you any of them · the median arrived in week 4 · all six you went and got
Six observations. Four of them had stopped looking strange within six weeks, one was noticed too late to act on, and one was fixed. Nothing about the organisation changed in the meantime. One person's own twelve weeks.
For roughly the first eight weeks you can see the place you have joined the way an outsider sees it. Then you cannot, and the change happens without being noticed.
This is not a loss of attention. It is what learning a place consists of: the strange becomes the ordinary, and that is how you become able to work in it.
What goes
Three kinds of thing, and they go in this order.
Vocabulary strangeness goes first, within two or three weeks. Words that meant nothing become words you use.
Process oddity goes next, around week six. Two deploy guides stops being a problem and becomes a thing you know the answer to, which is not the same as the problem being solved.
Structural oddity goes last and partially, around week ten. Who is in which meeting, what nobody talks about, which product nobody can explain.
In the panel above, four of six observations had stopped looking strange by week six.
What this costs the organisation
Everybody who could see the problem is now somebody who has learned to work around it.
That is why the same difficulties survive for years in otherwise competent organisations: the only people who can see them are new, the new are not listened to, and within two months they are no longer able to see them either.
What to do about it, for yourself
Write it down. This is the entire remedy and it is in this collection three times because nothing else works.
The document from the first week, kept for a fortnight, is the only instrument that survives the transition. Everything else — memory, good intentions, a plan to raise it later — does not, because the thing being preserved is not a fact but a perception, and the perception expires.
What to do about it, for the organisation
In week eight or nine, read the list back and pick one or two items that still look wrong to you.
| What to raise | What to leave |
|---|---|
| Something concrete, with a cost you can name | A general impression that things are disorganised |
| Something you have already attempted to fix | Something you want somebody else to fix |
| Something that affects the next joiner | Something that affects only you |
| One or two items | A list |
Then offer them as questions, to one person, once.
Something I noticed when I joined and have been meaning to ask about: there are two deploy guides and nothing says which is current. I worked it out but it cost me most of a day. Is there a reason to keep both?
That is answerable, it is not a criticism of anybody, and if there is a reason you will hear it.
Why one or two
Because a list of eight is a verdict on the organisation delivered by somebody who has been there two months, and it will be received as one however carefully it is written.
Two items, raised as questions, are a new person being helpful. The same two items plus six others are a problem.
The last useful act
Around month four or five, somebody else will join.
You are the last person who remembers what was confusing, and your list is still readable. Twenty minutes with them, working through it, is the single most useful thing anybody will do for them — and it is the thing that nobody did for you.
After that you are one of the people who has always been there, and the window is properly shut.
What it feels like from inside
Nothing. That is the difficulty.
There is no moment at which you notice that you have stopped seeing something. The two deploy guides do not become acceptable; they stop being two guides and become "the one in the repository, which is the real one", which is a fact about your working habits rather than an observation about the documentation.
Because there is no sensation attached to it, you cannot use judgement to decide when to write things down or when to raise them. The dates have to do that work instead.
Which is the whole argument for a dated document, and why it is the one recommendation in this collection that has no substitute: it is the only instrument that measures something you cannot feel.
Two months, not three
Worth being specific about the timing, because the common advice to "give feedback at the end of probation" is too late by about a month.
By week twelve you are no longer the right instrument. You have developed working habits around most of what was strange, and what remains is a short list of things that still irritate you personally, which is a different and much less useful thing.
Week eight is about right: late enough to have some standing, early enough that the list still has something in it.
The same applies to fixing documentation. Week two is better than week eight, and week eight is better than never.