Seven checkpoints in a new hire's first ninety days, and what each one tells you

A probation period is usually described as a trial. In practice most of them are a date on a calendar that arrives without anyone having gathered evidence, which is why so many end in a decision made on a feeling.

The fix is structural. Space the observation points so that each one asks a question the previous one could not answer, and the ninety-day decision makes itself.

Here is why the spacing works. Different qualities become visible at different rates. Willingness shows in days. Comprehension shows in weeks. Judgment only shows once the novelty has worn off and nobody is watching, which takes about two months.

Day one, end of shift

What it tells you: whether your onboarding works.

Ask one question: what was confusing today? A new person on day one has the clearest view of your business that anyone will ever have, and it lasts about a week before they absorb the local assumptions.

Whatever they name is a real defect in your setup, not a defect in them. Write it down. It will save the next hire a week.

End of week one

What it tells you: whether the training is landing, and whether they ask.

The signal here is not competence. Nobody is competent in a week. The signal is the shape of their questions. Specific questions about how something is done mean the training is being absorbed and applied. No questions at all is the warning sign, not the good sign, because it usually means either they are not doing enough to hit an edge or they do not feel able to ask.

Fix the second one immediately by naming a single person whose job it is to answer, so the question has an address.

End of week two

What it tells you: whether the job matches the ad.

By now they have seen the actual work rather than the described work. Ask directly whether anything is different from what they expected.

Two payoffs. If something is meaningfully different, you have a chance to address it before it becomes a quiet resignation in month four. And whatever they name is a line in your job ad that is misleading, which you can fix before you post again.

Day thirty

What it tells you: whether they can do the core task unsupervised.

This is the first real performance checkpoint, and it should be measured against something written on day one. Not a feeling. A short list: the tasks a competent person in this role does without asking, by thirty days.

The mechanism that makes this work is that the standard existed before the assessment. A person who knows the thirty-day list can aim at it. A person being judged against an unstated standard is being asked to read minds, and the ones who are good at reading minds are not necessarily the ones who are good at the job.

Three possible findings, and each has an obvious response:

• Meets it. Move the target. Give them something from the sixty-day list early.
• Close. Name the specific gap and the specific support. This is normal and it is what probation is for.
• Not close, on a task they were trained on. This is the point at which the honest conversation costs the least, for both of you.

Day forty-five

What it tells you: how they behave when something goes wrong.

You cannot schedule this one precisely, so schedule the review and use whatever incident has occurred by then. Every job produces a mistake, a difficult customer, or a day where the plan collapses within six weeks.

What you are watching for is the response pattern, which is stable and predictive in a way that raw skill is not. Did they raise it early or hope it would resolve itself? Did they describe what happened accurately? Did the same category of thing happen twice?

Skill improves with training. The instinct to surface a problem early is much harder to install, which is why it is worth weighting heavily.

Day sixty

What it tells you: whether they have started improving things.

By two months a person who is going to be genuinely valuable has begun to suggest something. A better order for a task, a question about why a form exists, a shortcut they used elsewhere.

This is the checkpoint that separates adequate from good. Someone who has been silently doing exactly what they were shown for two months may still be a solid hire, but they will not be the one who makes the operation better, and it is useful to know which you have hired before you build plans around it.

Ask explicitly if nothing has surfaced. Many people wait to be invited.

Day eighty-five

What it tells you: what the decision is, with five days to act on it.

Not day ninety. Day ninety is too late to prepare anything, and a decision made on the deadline is a decision made under pressure.

Bring the written notes from the six previous checkpoints. By this point the answer is usually obvious and already known to both parties, which is exactly the outcome the structure is designed to produce. Surprises at the end of a probation period are evidence that the middle was not observed.

Whichever way it goes, say it plainly, put it in writing, and if it is a confirmation, say what the next six months look like and when pay gets reviewed.

What the structure actually buys

Two things. A decision supported by dated notes rather than an impression, and a new employee who spent ninety days knowing where they stood.

The second one is the part that shows up in retention. Most people who leave in the first year describe the same experience: they never knew whether they were doing well. Seven short conversations, none of them longer than twenty minutes, remove that entirely.