A new hire is paid in full from day one. Until they have context, part of that salary buys questions instead of work, and every question also costs the teammate who stops to answer it. The cost is salary for the months before full productivity, times the unproductive share, plus teammates' hours. Your own numbers decide how much.
What does a new hire cost before they have context?
There are two invoices, and nobody sends either one.
One is the new hire's own salary. They're capable. You checked. But for a while they produce little, because what they need to know isn't in front of them: who owns what, why the process looks the way it does, what was already tried and abandoned. It's all written down. Across 300 Slack channels.
The second invoice is everyone else's time. Each question the new hire asks is answered by someone who was doing something else. Every question costs twice: once for the person asking, once for the person who stops to answer.
Neither appears on a budget line. The salary was approved months ago. Interruptions aren't tracked anywhere. So the cost is real, it recurs with every hire, and nobody has put a number on it.
This post puts a number on it. Yours, not ours.
The formula
Four steps. A spreadsheet is optional.
- Salary during the wait. Annual salary ÷ 12 × months to full productivity × the share of that time that isn't productive yet.
- Hires. Multiply by the number of people you're hiring this year.
- Teammate hours. Hours per week teammates spend answering the new hire's questions × the number of weeks × what an hour of those teammates costs.
- Add them. That's the cost of a new hire without context.
Two of the inputs need judgment.
Months to full productivity is the point where a manager stops budgeting extra time for this person. Not the end of orientation, and not the end of the 30-60-90 template. The month the questions thin out.
The unproductive share is an average across those months. It's close to all of the time in week one and close to none by the end.
Neither input is precise, and neither needs to be. A rough number you believe beats a precise one you borrowed.
A worked example, with assumptions you should change
Every number below is an assumption, not a measurement.
- Salary: $100,000 (assumption)
- Months to full productivity: 3 (assumption)
- Share of those months that isn't productive yet: 50% (assumption)
$100,000 ÷ 12 × 3 × 50% = $12,500.
That is one hire, salary only, before a single teammate hour is counted.
Now use your own numbers. Your salaries aren't round. Your roles differ: an engineer joining a five-year-old codebase and a CSM inheriting forty accounts are not waiting on the same things. And your managers already know the real months figure. They say it in one-on-ones. They don't say it in budget meetings.
Change one input at a time and watch what moves. Double the months and the figure doubles. Raise the unproductive share and it climbs with it. The arithmetic is not the hard part. Agreeing on honest inputs is.
The onboarding cost calculator runs the same arithmetic with your inputs, and every default in it is labeled as an assumption.
The second invoice: who answers the questions
The salary line is the one people expect. The teammate line is the one they haven't looked at.
Knowledge workers spend, on average, 1.7 hours a week providing duplicate information or repeating the same answers or updates (APQC, 2022). That figure covers knowledge workers in general, in an ordinary week. It is not a measurement of the people sitting nearest a new hire, who would probably like a word.
The people answering are rarely junior. Questions about why a decision was made go to whoever was there when it was made: the senior engineer, the longest-serving account owner, the manager. An hour of their time costs more than an hour of the new hire's, and they pay a second time in whatever they were doing before the interruption.
So count them:
- Name the two or three people a new hire leans on most.
- Estimate the hours a week each of them spends answering.
- Price those hours at what those people cost.
- Multiply by the weeks until the questions thin out.
It is usually the part of the bill nobody had looked at.
How long does it really take?
Longer than the job posting implied.
For technical roles there is a published figure: 4.8 months is the average time it takes to onboard new technical hires to normal productivity, with 58% of organizations requiring more than four months (The Linux Foundation, 2024 State of Tech Talent Report).
That figure is for technical hires only. For every other role, use your own history. Ask three managers how long it took their last hire to get through a week without a daily question, and take the honest answer over the hopeful one.
Whatever the number is, look at what fills those months. Very little of it is learning the job. Most of it is finding out who owns what, why things are the way they are, and what's already in motion. All of which someone, at some point, wrote down.
Run your own numbers
Put your salaries, your hires and your months into the onboarding cost calculator. It won't tell you what Firstweek saves, because there are no measured pilot results yet and we won't invent one.
Your number is the argument. Take it to whoever approves headcount and ask what those months are buying.
The months input is the one worth arguing about. Part of it is skill, and skill takes time. Part of it is context, which is already written down in Slack, Confluence and the CRM, and has never been handed to the person who needs it.
Firstweek reads those tools through read-only connections and writes each new hire one cited onboarding artifact, which the hiring manager approves before day one. You can read a sample artifact, or see how it's put together in the product tour.
FAQ
How do I calculate what a new hire costs before they're productive?
Multiply annual salary ÷ 12 by the months to full productivity and by the share of that time that isn't productive yet. Add the hours teammates spend answering questions, priced at what those teammates cost. Then multiply by the number of hires.
What counts as full productivity?
The point where the hire's manager stops budgeting extra time for them. It varies by role, so use your own history. For technical hires, The Linux Foundation reports an average of 4.8 months to normal productivity.
Why count teammates' time?
Because every question costs twice: once for the person asking, once for the person who stops to answer. The people answering are usually senior, so their hours are the expensive ones.
Does Firstweek reduce this cost?
We don't publish a savings figure, because there are no measured pilot results yet. A pilot runs on one real hire for 30 days, and you judge the artifact yourself.
Sources: APQC, KM Makes Knowledge Workers More Productive and Less Stressed Out (2022). The Linux Foundation, 2024 State of Tech Talent Report, p. 20. Both are listed with links on the onboarding statistics page.
Run it on your next hire. 30 days. Then we talk. Start a pilot or see a sample artifact. How it works for scaling teams.