Team & capacity
Onboarding someone onto a live client project in three days
A three-day plan to get a new person contributing to an active client project, with an hour-by-hour schedule, the six documents they need, and the four things you must not let them do in week one.
·5 min read
The default onboarding at a small agency is: here's the Slack, here's the drive, sit with Maria this week, shout if you're stuck. Three weeks later the person is still asking where things are, and the senior they've been shadowing has lost about a day a week of billable time.
The alternative isn't a handbook. It's three deliberate days with a defined output at the end of each, after which the person is contributing real work with review rather than absorbing context by osmosis.
This is written for someone joining an active client project, which is the hard case and the common one.
Prepare before day one
Half a day of your time, done the week before. If these aren't ready, the three days don't work.
- Accounts and access. Every tool, on day one, working. Nothing kills momentum like spending Tuesday morning waiting for a licence.
- One named buddy, who is not you and not the busiest person on the account. Their job is questions, not teaching.
- A first real task, already chosen. Small, genuinely useful, reviewable, and safely reversible. Not a fake exercise — people can tell.
- A one-page project brief written for someone with zero context. See below.
- Two calendar holds: end of day one and end of day three, 30 minutes each, with you.
The six documents they need
Not a wiki. Six things, ideally one page each. If they don't exist, write them once and reuse them for every future hire on that account.
| Document | What's in it |
|---|---|
| Client one-pager | Who the client is, what they sell, who our contacts are, what they're like to work with, what they've rejected before |
| Project one-pager | What we're building/making, the deadline, what's in scope, what's explicitly out |
| How work flows | The path from brief to delivery, who reviews what, who talks to the client |
| Where things live | Files, assets, tickets, credentials, brand or style rules |
| Time and money rules | How to log time, against what, what "billable" means here |
| Glossary | Twenty internal and client-specific terms. Underrated; every agency has fifty |
The three-day plan
Day one — context
| Time | What |
|---|---|
| 9:30 | Kickoff with you, 45 min. Why this role exists, what good looks like at 30 days, who's who |
| 10:30 | Setup and access checks with the buddy. Everything opened once |
| 11:30 | Read the six documents. Alone, uninterrupted. Take notes on anything unclear |
| 13:30 | Walk the project. The buddy demos the actual work — the files, the site, the campaign, the edit — as it stands today |
| 15:00 | Sit in on a real client interaction. A call if one exists, otherwise read the last month of client email or comments |
| 16:00 | Write it back: half a page describing the project in their own words, as if explaining it to a friend |
| 16:45 | End-of-day check-in with you, 30 min. Read their half-page together |
That half-page is the whole point of day one. It surfaces misunderstandings immediately, and it tells you exactly which of your six documents is unclear.
Day two — supervised contribution
| Time | What |
|---|---|
| 9:30 | Answer yesterday's open questions, 20 min with the buddy |
| 10:00 | Start the first real task. Brief it properly, in writing, with a deadline of end of day |
| 12:00 | Working, with the buddy available but not hovering |
| 14:00 | Mid-point check: 15 minutes, look at what exists so far. Correct direction now, not tomorrow |
| 15:00 | Finish and submit for review |
| 16:00 | Review together, live. Explain not just what to change but why this client/project works that way |
| 16:45 | They log their time for the day, with the buddy, against the right project |
Day two exists to establish the loop: brief → work → review → revise. Once someone has been through it, they know what "done" means here, which is the thing that takes new people longest to learn.
Day three — semi-independent work
| Time | What |
|---|---|
| 9:30 | Second task, briefed more lightly. They ask for what they need |
| 12:00 | Introduce them to the client — a short message, or two minutes on a call. Do this in week one, not month two |
| 13:30 | Continue working. Buddy checks in once, unprompted |
| 15:30 | Submit for review |
| 16:00 | The 30-minute session with you. Three questions: What still doesn't make sense? What surprised you? What would you change about the last three days? |
| 16:45 | Agree the first month: what they own, what they don't touch yet, when you next talk |
Write down the answer to "what they own" and send it to them. Ambiguity about ownership is the most common reason new people underperform in month one, and it's entirely the employer's fault.
Four things not to let them do in week one
Reply to the client unsupervised. Introductions yes, decisions no. One confident wrong answer to a client costs more than the whole onboarding.
Touch anything irreversible. Live sites, published campaigns, master files, anything with a delete button and no undo. Not about trust — about giving them a week where mistakes are cheap.
Sit in meetings all day. Observation without doing is the least efficient way to learn a job, and it's how three-day onboarding becomes three weeks.
Log 100% billable time. Their first weeks are not billable and pretending otherwise corrupts your utilisation data at exactly the moment you need it to be honest. Book onboarding to an internal code.
What this costs
Be honest about it up front so you don't resent it on Wednesday. The buddy loses roughly half of days one and two and a couple of hours on day three. You lose about two hours across the three days, plus the half-day of preparation. The new person produces two small pieces of real work.
That's real money against billable time. It is also considerably cheaper than the alternative, where the same senior person answers scattered questions for a month, in fragments, at the worst possible moments, and the new person is still unsure what they own in week five.
Do the reusable part once
The six documents are the leverage. Written once per account, they onboard every future hire, every contractor, and every person who covers that project during someone's holiday.
Pick your most active client and write the client one-pager and the project one-pager this week — two pages, an hour. Then ask the person who joined most recently to read them and mark everything they didn't know on day one. That list is the rest of the work, and it's shorter than you think.
Keep reading
- When to hire your next person, and how to know you can afford themA hire whose fully loaded cost is $118,000 a year needs 66 billable hours a month just to break even. Here's the capacity test, the cash test, and the ramp curve most agencies forget to model.
- Performance reviews for small agency teamsAnnual reviews don't work at eight people. Here's a 45-minute quarterly format with a one-page written template, how to talk about utilisation without weaponising it, and how to keep pay conversations separate.
- Contractor vs employee for agencies: the real mathCompared properly, an $80,000 employee costs $80.44 per billable hour at 70% utilisation and a $65/hour contractor costs $70.65. The comparison flips at about 80% utilisation. Here's the full calculation.