Client onboarding closes in five working days when each day owns exactly one mandatory box: access, brand materials, a named approver, a tested circuit and the scope in writing. What stretches an onboarding is never the work itself, it is waiting for somebody to reply, so every box gets one name against it and one date.
The five days, and what jams when one slips
An onboarding does not run long because there is too much to do. It runs long because one of the tasks depends on somebody who does not work for you. So the useful checklist is not a list of jobs, it is a list ordered by who has to answer, with the slowest replies asked for first.
| Day | What closes | What jams if it does not |
|---|---|---|
| One | Access requested, client created with their brands | Everything else, because the calendar cannot be tested |
| Two | Brand materials collected and the brief filled in | The first piece is written blind and gets redone |
| Three | A named approver and an agreed review deadline | The first review turns into an email chain |
| Four | The portal opened and the rest of their people inside | You become the client inbox yourself |
| Five | The first month planned and the scope in writing | The month three argument, which you can no longer win |
That table reads downwards and not upwards, because the first two rows carry somebody else's latency. Access is granted by a person inside the client's company who was often not even in the kick off meeting, and the brand materials usually live on the laptop of somebody who left a year ago. Ask those two questions on Monday and Friday still leaves room to chase. Ask them on Thursday and the onboarding rolls into another full week.
The last three rows depend only on you and on one decision taken in a call. They are the ones you can compress when you have to, and knowing which they are is what stops you cutting into the ones you cannot.
The approver is a person, not a department
This is the box that jams most onboardings, and it jams because people fill it with a job title instead of a name. "Marketing approves it" is not an approver: it is a place where a piece sits still while two people each assume the other was looking at it.
Client approval is switched on client by client here, and with it on, one yes is enough for the piece to go out: somebody on your team gives it from the app, or the client gives it from their portal, whichever comes first. There is no order, and it never needs both. And for a client who does not need to review anything, the switch stays off and pieces are scheduled straight away, which is the default.
What you actually agree on day three is not the switch, it is the deadline. How many working days the client has to look at a piece, and what happens when nobody answers inside them. A reminder goes out before the date of a post that is still waiting, and a piece whose hour passes while it waits is not published late: it is parked for rescheduling. That protects the account, but the slot in the calendar does not come back, and that is exactly the conversation to have in week one rather than next month. The whole mechanism is in content approvals.
The next day follows the same logic and almost nobody does it early enough: open the portal and invite not just your main contact but everybody on their side who will ever need to look at something. The person in accounts who asks about an invoice and whoever uploads the product photos should not have to go through your inbox to reach their own things. Each one signs in as themselves, reaches their own company and nothing else, and every decision carries a name. What they see is described on the client portal page.
Access goes out on Monday morning
Requesting access is the task with the most dead days in the whole process, and the one most people leave until everything else is done. Between asking who administers the Business Manager, the client finding out, tracking down that person and that person sitting down to grant it, three good days pass without anybody doing anything wrong.
There is a route that saves half of that wait and it is worth proposing in the kick off meeting itself: let the client connect their own accounts from the connections row of their portal. It inverts the usual flow, because instead of asking them for a password over email, they authorise it from their own session, and on the day the work ends access is withdrawn from the same place. The full script, platform by platform, is in how to request access without holding their passwords.
On the advertising side the wait is longer still, because there is usually one more third party involved: the previous agency, still listed as an administrator, who has to be asked to step back. That request goes out on day one without exception, even when you are not touching ads until month two.
Brand materials get collected before anything is written
Day two is collection, and it goes better as a form than as a conversation. A brief built from the fields you genuinely need for that kind of work saves the client writing three vague paragraphs and saves you three follow up emails. When a brand document already exists, the PDF goes up and the brief is built around it rather than making anybody retype what was already written down.
What you need to walk away with on day two is short and almost always incomplete on the first pass: the logo in vector, the typefaces, the colours with their codes, two or three pieces they like and two they do not, which accounts exist and which do not, and one person to ask when something does not fit. Files go up to the shared Drive from the portal, in both directions, with no loose links that nobody can later account for.
One structural point saves a scare in week two. A client's brands hang off that client, and every brand uses at least one of the connected accounts you have contracted. There is no brand limit in the plan, there is a connected account limit, so the time to count them is before you promise four profiles and not on the day you try to connect the fifth. How workspace, client and brand fit together is in workspaces, brands and clients.
Everything in this article happens in one place in GoFeed.
Try it freeThe first post is a circuit test
The piece that opens the account does not have to be the best work of the quarter. It has to travel the whole route and prove no section of it is broken: produced, uploaded, approved by somebody, published on time, and its numbers coming back to the same place. That return leg is the part almost nobody checks and the most expensive one when it fails, because it fails quietly.
It also helps if it is deliberately boring. A simple piece with no campaign behind it and no critical date lets the client press the wrong button without consequences, and lets you find out whether the person you invited is the one who actually decides. Very often they are not, and finding that out with a filler photo costs nothing.
If the first piece takes more than ten days to go out after signing, the problem is not production. It is one of the five boxes in the table, almost always the third one, and saying so out loud is more useful than speeding up the part that was already working.
What gets written down on day five
Onboarding ends with a document, not with a feeling. It does not need to be long, it needs numbers wherever an argument is going to happen later: how many pieces a month and in what format, how many rounds of revision are included, what happens with urgent requests, who owns the accounts and the material, and how the relationship ends if either side wants it to.
Those five lines are what turns month three into a conversation instead of a negotiation. The long version, with the clauses that go missing most often and what each one costs when it does, is in contracts and deliverables.
There is one last habit that appears on no checklist and decides more than it looks like it should: book the first monthly meeting on the same day the onboarding closes, with the calendar open and the invitation sent. A client who has the next appointment in their diary from that first Friday starts the relationship knowing when they will see results. One who does not starts by wondering.