A 30-60-90 onboarding plan: what goes in each stage
Echipa HR 365 · reviewed 2026-09-02 · 5 min read
A 30-60-90 plan splits the first three months into three different questions: at 30 days the person should know how the place works, at 60 work with supervision, at 90 deliver alone. If all three stages contain the same kind of task, you do not have a plan, you have a long list.
Days 1–30: context, not performance
The first month has one objective: the person understands where they have arrived and who to ask. Anything requiring output in that period actually delays the moment they become autonomous, because it forces them to deliver before they understand the context.
What happens, and who owns it
- IT and facilities — account, access, equipment — ready on day 1, not “in the first week”
- HR — the initial training and a walk through the structure: who does what, where the procedures live
- Manager — explains what “done well” means for this role, with two concrete examples
- Mentor — shadowing on the main workflow, in the first two weeks
- Manager — gives a first small task, with a visible result, in week two
Days 31–60: autonomy with a safety net
The second month shifts the weight from information to practice. The manager no longer explains beforehand, they check afterwards. The difference matters: in month one the question is “did you understand?”, in month two it is “show me what you did”.
- They take on part of the normal workflow, with growing volume — not all of it at once.
- A weekly one-to-one, with the same agenda every time, so the progress is visible.
- The first piece of work they carry end to end alone, checked only at the finish.
- The first direct interaction with an internal or external recipient, if the role requires it.
At the end of month two, the manager should be able to answer one question: do I still need to check everything they do? If the answer is yes, the next question is not about the person, but about what was never explained to them.
Days 61–90: delivery and decision
The last stage has a result, not just activities: an assessment against criteria written in advance, which the person saw on day 1. An assessment whose criteria appear on day 89 does not assess the person, it assesses the surprise.
| Area | What you show | How it is checked |
|---|---|---|
| Autonomy | Carries work through with no intermediate check | How many needed intervention in the last month |
| Quality | The result does not come back for corrections | The return rate, or concrete examples |
| Integration | They know who to ask and get asked in return | Direct observation, not a questionnaire |
| Pace | Volume approaches that of a colleague with tenure | Comparison with the team average, as a reference, not a target |
What a month of delayed autonomy costs
An employee on a gross salary of 1,000 EUR, reaching full productivity a month later than they could have. Manager hourly cost 15 EUR, mentor 11 EUR.
- Salary cost for the month of partial output (at 50%)
- 500 EUR
- Additional manager time (10 h × 15)
- 150 EUR
- Additional mentor time (8 h × 11)
- 88 EUR
- Cost of a single delayed month
- ≈ 740 EUR
Nu intră în calcul:
- the cost of recruitment, if the person leaves because of poor integration
- the effect on the team absorbing the difference
- employer contributions, which vary from case to case
The figure is useful in the conversation about who does the onboarding. A mentor spending eight extra hours in a month costs under a hundred euros; a month of delayed autonomy costs seven times more. It is one of the few calculations in HR where the answer is obvious.
Three mistakes that appear in almost every plan
- The plan is written by HR and the manager does not see it until day 1. Then it stays a document, not a working plan.
- Every task is due “in the first month”. With no specific dates, they all get done in the last week or not at all.
- Nobody checks at day 30 and 60, only at 90. The two intermediate checks are the only ones that can still correct anything.
A new post and an inherited post do not share a plan
The difference is bigger than it looks and is almost always ignored. Somebody taking over an existing post inherits work in progress, relationships already formed and expectations built by their predecessor. Somebody opening a new post has none of that, but has a different problem: nobody in the organisation knows exactly what they should be doing.
| Stage | Inherited post | New post |
|---|---|---|
| Days 1–30 | A structured handover, with work in progress listed | Clarifying the role with everyone who will work with them |
| Days 31–60 | Taking over gradually, with the predecessor or manager as a safety net | A first visible result, so the role takes shape in the team’s eyes |
| Days 61–90 | Comparison with how things ran before | Defining in writing what the role does, based on what proved necessary |
| The main risk | Permanent comparison with the predecessor | The role stays vague and absorbs everything nobody wants |
The risk in the last row, on a new post, is what produces departures at month four: an undefined role attracts, in the first weeks, every orphaned task in the team. By day 90, the person is doing something other than what they accepted, and usually nobody decided that explicitly.
Does it apply to operational roles that are learned quickly?
The structure yes, the timescale no. For a role learned in three weeks, the same three stages compress to 7-14-30 days. What stays is the order: context, then supervised practice, then autonomy.
Who writes the plan?
The manager, with a template from HR. The reverse — HR writes it and the manager receives it — produces plans that are correct on paper and that nobody follows.
Onboarding templates per role, with tasks and owners assigned automatically
The plan activates itself at hiring, with real deadlines at each stage and tasks for IT, the manager and the mentor — so day 1 no longer depends on anybody’s memory.
Free account, every module for 7 days, no card required.
Sources
- Michael D. Watkins — The First 90 Days — revised edition, ch. 1–2: the first months split into stages with different objectives, and the most frequent mistake is asking for results before the person understands the context. · read on 2026-09-03