Career · 7 min read

The first 90 days in a new job: what actually matters

The first three months set an impression that is disproportionately hard to change later. People form a view of a new colleague quickly and revise it slowly, which means early behaviour carries more weight than it deserves. The good news is that the behaviour that works is fairly consistent across roles and levels, and most of it is unglamorous.

The single most common mistake is trying to prove value immediately by changing things. It is understandable — you were hired partly for your experience, and the flaws in a new place are visible precisely because you are new — but it is close to the worst opening move available.

Why proposing changes early goes badly

When you arrive, you can see what is wrong and not why it is that way. Almost every odd process is the scar tissue of something that happened before you: a customer incident, a regulatory requirement, a person who left, a battle someone lost. Proposing to fix it on week two signals that you have not asked, and it quietly insults everyone who lived through the reason.

This is not an argument for silence. It is an argument for sequencing. Observations made in week two are cheap and forgettable; the same observation in week eight, informed by why the thing exists, is credible and usually welcomed. Write them down as you notice them and hold them.

The first week: learn names and the shape of the place

Your only real jobs in week one are to get set up, meet people, and start building a map. Get your access sorted on day one and flag anything missing immediately, because access problems compound and nobody notices you are blocked.

Ask your manager three questions early: what does success look like at ninety days, what is currently on fire, and who should I make sure I meet. That last one produces a list you would not have assembled yourself, and it gives you a reason to approach people that is not awkward.

Write down every name with what they do and something specific about their work. It feels excessive on day three and it is invaluable in week five, when you are trying to recall who owns the thing you need.

The first month: ask the questions while you still can

There is a window in which asking anything is free, and it closes faster than people expect — usually around the six-week mark, when the same question starts to read as not having paid attention. Spend that window aggressively.

Ask why more than what. Why is this deployed on Thursdays, why does this team own that service, why did we choose this vendor. The answers contain the institutional history that nobody writes down and that determines what is actually possible here.

Keep a running list of things that confuse you. Some will resolve themselves as you learn more; the ones still on the list at week eight are frequently real problems that everyone else has stopped seeing, and by then you can raise them with enough context to be useful rather than naive.

Take notes in a form you can search. Onboarding is a firehose and you will be told things once, in passing, that turn out to matter in a month.

Find the first small win

Somewhere in the first month, deliver something real and finish it. Not a strategy, not a proposal — a completed thing that someone else notices. Fix the flaky test everyone works around, write the documentation that does not exist, take the recurring annoyance off someone's plate.

Size matters less than completion. A small finished thing establishes that you deliver, which buys credibility for the larger arguments you will want to make in month three. A large unfinished thing establishes the opposite, and unfinished is the default state of anything ambitious started by someone who does not yet know the system.

The other function of an early win is that it teaches you how work actually ships here: who reviews, what gets blocked, how long things take. That knowledge is worth more than the win itself.

Managing upward without being annoying

Your manager is forming a view of whether you need managing. The way to answer that favourably is a short, regular, unprompted update: what you did, what is next, what is blocked. A few lines weekly is plenty. It is not self-promotion; it removes the need for them to check, which is the thing overloaded managers actually value.

Ask for feedback early and specifically, before the formal review exists. How is this landing, is this the level of detail you want, would you have approached this differently. Generic requests get generic reassurance; specific ones get usable answers.

If your manager is absent or disorganised, which happens, take more responsibility for the rhythm rather than less. The alternative is arriving at month three with no shared understanding of how you are doing.

Relationships beyond your team

Most of what makes a job workable long-term is relationships with people outside your immediate group, and they are far easier to build in the first weeks when being new is a reason to introduce yourself. After that, reaching out requires a pretext.

Talk to the people your work depends on and the people who depend on your work. Ask what they wish your team did differently. It is a disarming question, the answers are usually specific, and you will hear things nobody says in a meeting.

The second month: start contributing your judgement

By now you have earned the right to have opinions, and you should start offering them. The shift is from asking questions to proposing answers — still tentatively, still with the caveat that you may be missing context, but visibly thinking rather than only absorbing.

This is also when to start returning to that list of things that seemed strange. Raise them one at a time, framed as a question rather than a verdict: I noticed we do X, is that deliberate? Sometimes the answer is a good reason you did not know. Sometimes it is nobody has looked at that in three years, which is an invitation.

Month three: the honest checkpoint

Around ninety days, take stock deliberately rather than letting it drift. Are you doing what you were hired to do? Does your manager's picture of your progress match yours? Have you delivered something you would put on a resume?

Ask directly for an assessment if one is not offered. Where am I doing well, where would you like to see something different. It is a normal question at this point and the answers are far more actionable than they will be at a formal review six months in, when opinions have set.

If it is not working

Occasionally the job is not what was described, or the team is in worse shape than anyone admitted. It is worth separating the discomfort of being new — which is universal and passes — from a genuine mismatch, and the reliable test is whether the specifics are getting better. Confusion should be shrinking every week. If it is not, or if what you were hired to do keeps being deprioritised, that is a signal rather than a phase.

Raise it before you conclude anything. A direct conversation about the gap between the role as described and the role as it exists sometimes fixes it outright, because the mismatch is often unintentional. If it does not, you have the information early, which is far better than realising at month nine.

The short version

Do not change anything in the first month; ask why instead, while asking is still free. Write down names and confusions. Finish one small visible thing early. Send your manager a few unprompted lines each week. Meet the people either side of your work while being new is still a reason to. Then, at ninety days, ask plainly how it is going rather than waiting to be told.

Remote and hybrid starts need deliberate effort

Starting remotely removes almost every accidental interaction that normally does the relationship-building for you. Nobody shows you where things are, nobody introduces you in a corridor, and the natural drift into knowing your colleagues does not happen. It has to be replaced on purpose.

Book short introductory calls with everyone you will work with, and say plainly that you are new and would like to understand what they do. Almost nobody refuses this, and it is far easier in week two than week ten. Fifteen minutes each is enough.

Over-communicate in writing, more than feels natural. In an office your presence is evidence that you are working; remotely, your visible output is the only evidence there is. A short daily or weekly note is not performative in that context, it is the substitute for being seen.

If the company is hybrid, find out which days actually matter and be there on those. Turning up on the quiet days and missing the ones where decisions happen is a common and invisible mistake.

A note on the first job, or the first after a long gap

If this is your first role, or your first after a break, expect the adjustment to be about pace and norms rather than skill. Most early-career discomfort comes from not knowing what is normal — how quickly to reply, whether a question is too basic, how much autonomy you have — and the fastest fix is to ask your manager directly rather than inferring it over months.

It is also worth knowing that feeling underwater for the first several weeks is the standard experience and not a verdict on you. Everyone around you has more context, and context is the thing that arrives with time rather than effort.

Related guides

Try JobStraight free →