The Trial Period, and How to Actually End One

Almost every group that runs a trial period ends almost every trial period with a yes. Not because the fit was always right, but because by the time the review arrives the person has a desk, friendships, work in flight, and a reasonable expectation of staying. Saying no has become socially expensive, so the group finds reasons not to.

A trial period that cannot end in no is not a trial period. It is a delayed start date, and it costs you the one mechanism that protects a small group from the wrong person.

Design it so no is possible

Say the word "no" out loud on day one. Not "we'll see how it goes". Say: "at the end of these eight weeks we will decide yes or no, and no is a real outcome that happens here." If it has never actually happened, say that too, honestly, and say it is still real.

Put the review date in the calendar before they start, with the decision-makers named. A date that gets scheduled later gets scheduled late, and late means never.

Agree what you are assessing, in writing, before day one. Three to five specific things. Not "fit". Things like: can they take a decision without asking; do they raise a tension rather than sit on it; is their work at the standard we need; how did they handle being disagreed with.

Give them the list. Someone who knows what they are being assessed on can actually address it, and the ones who do are usually the ones you want.

Keep the first weeks light on entanglement. Do not give them the only copy of something critical in week two. Not out of distrust, but because a trial where saying no would break three projects is a trial that will end in yes.

Name the cost of a wrong yes. Once, to the group, before the review. In a small self-managing organization, removing someone later takes months, damages relationships, and lands on people who did not sign up for it. Everyone should be holding that when they weigh a marginal yes.

Run real check-ins, not encouragement

Weekly, twenty minutes, both directions.

Ask them:

• What has been harder than expected?
• What have you not been able to get an answer to?
• What would you change about how we work?

That third one has a short shelf life. A new person can see your dysfunction clearly for about six weeks, then they normalize and lose the ability. Ask it every week and write down the answers, because they are the most valuable outside perspective you will get all year, regardless of whether the person stays.

Tell them:

• One specific thing that is going well.
• One specific thing that is not, as it happens, not saved for the review.

That second one is the whole discipline. A trial review that contains the first mention of a problem is a failure of the process, not of the person. If they never heard it, they never had the chance to change it, and a no at that point is unjust.

The decision

Ask everyone who worked with them, individually and in writing, three questions:

1. Would you want to work with this person on something difficult?
2. What is your hesitation? (Assume there is one. Asking "any concerns?" gets "no". Asking "what is your hesitation" gets an answer.)
3. If you had to say no, what would your reason be?

That third question is the useful one. It gives permission to voice a doubt hypothetically, which is much easier than asserting it, and it surfaces the quiet unease that otherwise stays quiet until it is far too late.

Then decide by consent, not by vote. The question is: does anyone have an objection to this person joining? An objection has to be a reason, not a feeling. But treat a feeling as a signal to go looking: "I can't say why but something is off" is not an objection, and it is worth an hour of finding out what is underneath it.

Saying no well

If it is a no, the person should not be surprised. If they are, you skipped the weekly conversations.

Do it in person, early in the week, early in the day. Not Friday afternoon. They will need to talk to someone afterwards.

Be specific and brief. "The work was good. The thing we could not get to work was that decisions kept coming back to us for approval, and we talked about it three times. In a group without managers, that costs more than we can carry." Then stop.

Do not soften it into ambiguity. "It's not the right time" invites a follow-up and leaves them without anything they can use.

Do not litigate it. You are informing, not building a case. If they disagree, let them disagree.

Pay them properly through the notice you promised, and say so in the same conversation.

Offer something real if you can. An introduction, a reference for what they were genuinely good at. Most people who are wrong for a self-managing group of eleven are right for something else, and saying so is both true and kind.

Where it goes wrong

Extending the trial. Almost always avoidance. If you genuinely need more information, name exactly what information and the date you will have it, and extend once. Twice is a no you are not saying.

One person carries the decision. In a group with no manager, the whole group's judgment is the mechanism. Using it means everyone is asked and everyone owns the answer.

Nobody raised anything until the review. Then the review becomes a pile-on, the person hears a list of things nobody mentioned, and everyone leaves feeling grubby. Ongoing feedback is what makes the ending fair.

Treating a no as a failure. A trial period that ends in no worked. It found something before it became expensive. Say that out loud to the group afterwards, because the next trial is watching.

What good looks like

The person is not surprised. The group is not divided. Somebody says out loud, a month later, "that was the right call and it was hard", and the next trial period is run just as seriously.

Try this week

If you have someone in a trial period, ask them the three questions today, especially "what would you change about how we work".

If you do not have anyone on trial, write down the three to five things you would assess, before you need them. Writing that list under time pressure with a candidate waiting is how it becomes "fit".