Client Offboarding: The Process That Wins the Next Project

Agencies invest heavily in the first two weeks of a client relationship and almost nothing in the last two.

That is backwards, commercially. A client at the end of a successful engagement is at the peak of their goodwill toward you. They have just received something valuable, the frustrations of delivery have faded, and they are more willing to say something generous about you than they will ever be again.

Nine months later, when you finally get round to asking for a testimonial, the details have gone, the enthusiasm has cooled, and the person who championed you may have moved on.

This guide covers what offboarding actually is, the five stages, exactly what belongs in a handover pack, when and how to ask for testimonials and case studies, how to handle an ending that is not amicable, and what to do in the 90 days afterwards.

Why it gets skipped

Three reasons, all understandable.

It is unbillable. Offboarding happens after the last invoice, when attention has moved to the next project. There is no line item for it and no client chasing it, so it loses to whatever is urgent.

It feels like an ending. Nobody enjoys formalising the end of a good working relationship, and there is a superstition that acknowledging it makes it final. In practice the opposite happens: a vague ending is what makes it final, because nothing is set up to continue.

Nobody owns it. Onboarding usually has a named owner. Offboarding is everyone's job at the point where everyone is busy, which means it is nobody's.

The result is a predictable pattern. The last deliverable is sent, a few loose emails follow, access lingers for months, the invoice eventually gets paid, and the relationship simply stops. No testimonial, no case study, no referral, no re-engagement.

What offboarding actually is

The framing that makes this click: offboarding is not an event, it is an arc.

ManyRequests' offboarding guide and ClientManager's walkthrough both describe it as a process rather than a handover, and the timeline that works in practice runs roughly:

• T-2 weeks: signal the ending, book the wrap-up.
• T-0: final delivery, handover session, handover pack.
• T+1 week: access closed, billing closed, feedback requested.
• T+2 to 4 weeks: testimonial requested.
• T+6 to 12 weeks: case study, results check-in, re-engagement conversation.

Three months, mostly in small increments. The total effort is a few hours; the value is a testimonial, a case study, a referral path, and a meaningful chance of the client coming back.

The five stages

1. Signal the ending early (T-2 weeks)

Two weeks before the final deliverable, say explicitly that the engagement is approaching its end and book a wrap-up session.

This sounds trivial. It is the highest-leverage step, because it converts a fade into a scheduled event - and scheduled events get prepared for.

What to send:

We're two weeks from wrapping up. I've put 45 minutes in the diary on the 18th to walk through everything, hand over access and documentation, and talk about what makes sense next. Anything you'd like to make sure we cover?

That message does four things at once: sets the end date, books the session, promises a handover, and opens the "what next" conversation without pitching. The last one is important - the re-engagement conversation is far easier when it is a scheduled agenda item than when it arrives as a sales email in six weeks.

2. The handover pack (T-0)

The single artefact that separates professional offboarding from a set of emails.

What goes in it depends on the work, but the structure holds:

A summary of what was delivered. Against the original objectives, with results where they exist. Two or three paragraphs. This is not for them to read once - it is the document that gets forwarded internally when someone asks what the agency actually did, including to people who join later.

Every asset, organised. Source files, editable formats, exports. Not a link to a folder that will lose permissions in six months - a structured, downloadable set they own.

Credentials and access. Everything you hold, listed, with what it is for. Domains, hosting, analytics, third-party accounts.

Documentation. How to update the thing you built. Written, not just a recorded call - though a recorded walkthrough alongside it is genuinely useful and takes twenty minutes.

What to watch. Known limitations, things that will need attention in six months, decisions you made that a future supplier should understand. This is the section that feels risky and is the most valuable: it is the clearest possible signal that you are optimising for their outcome rather than for lock-in.

Who to contact. For questions, for future work, and for emergencies during any agreed support window.

Agency Vista's offboarding packet guide and this offboarding checklist from Envato Tuts+ both cover comparable ground; the useful discipline is templating the structure so building one takes an hour rather than a day.

3. Close access and billing cleanly (T+1 week)

Unglamorous and worth doing precisely.

Access. Remove your team from their systems. Tell them you have done it and what was removed. Lingering access to a former client's systems is a real security exposure for both parties, and it is the kind of thing that surfaces badly during their next audit.

Billing. Final invoice, clearly marked as final. If it is a retainer, stop it at the end of the cycle and confirm in writing that it has stopped. An unexpected charge after the engagement ends undoes months of goodwill in one email.

Data. Say what you are keeping, why, and for how long. If they want it deleted, do it and confirm. This matters more every year.

Loose ends. Any outstanding items, explicitly closed or explicitly carried. Ambiguity here is where disputes start.

Telerik's piece on whether offboarding is overkill makes a point worth repeating: the clean exit is not bureaucracy, it is what makes the client comfortable recommending you, because they know working with you has no trailing obligations.

4. Ask for the testimonial and the case study (T+2 to 12 weeks)

Timing here is specific, and getting it wrong is why most agencies have thin proof.

Ask for the testimonial at 2-4 weeks. Late enough that they have lived with the work, early enough that the experience is vivid. Not on delivery day, when they cannot yet speak to the outcome.

Make it easy. The single biggest determinant of whether you get one. "Would you mind writing a testimonial?" is a task on someone's list. Instead:

Would you be happy to be quoted on this? To make it easy, here's a draft based on what you said on the wrap-up call - edit it however you like, or write your own if you'd rather.

Most people edit the draft and send it back the same day. That is not putting words in their mouth - it is removing the blank page, and you are quoting things they actually said.

Ask for the case study separately, and later. A case study needs results, and results need time. Six to twelve weeks is usually right.

The framing that works is mutual benefit: a case study featuring their company, their logo and their results is exposure for them too. ManyRequests' guide makes this point directly - pitching it as joint content, and offering to let them review and approve, converts far better than asking permission to write about them.

Ask for the referral last, and specifically. "Do you know anyone who needs this?" gets nothing, because it asks the person to do a search. "Is there anyone in your network dealing with the same problem you had six months ago?" gets a name, because it is a much smaller question.

5. Stay in contact (T+90 days and beyond)

A check-in at 90 days, then at six and twelve months.

Not a sales email. A short, genuinely useful message: "It's been three months since launch - how are the conversion numbers looking? Anything behaving oddly?"

Two things come out of this. You get the results data that makes your case study real, which you would otherwise never collect. And you are present at the moment their next need emerges, which is the actual mechanism by which former clients become repeat clients.

The agencies with strong repeat business are rarely doing sophisticated account management. They are sending three short emails a year to people they used to work with.

The wrap-up conversation

The 45-minute session at T-0 is the centre of the whole process, and it works best with a fixed agenda.

Ten minutes: what we set out to do, and what happened. Walk through the original objectives and where things landed. Include what did not go to plan - a wrap-up that presents everything as a triumph is less credible than one that names the two things that were harder than expected. It also makes the praise believable.

Ten minutes: the handover. Walk through the pack, screen shared. Access, assets, documentation, what to watch. Record it if they are happy for you to - a twenty-minute recording is worth more to their future team member than any written document.

Ten minutes: what we would do next. Not a pitch. An honest view of what would deliver most value over the next six months, whether or not it involves you. This is the single most effective re-engagement move available, precisely because it is not a pitch - and if you recommend something outside your capability, say so and name someone.

Ten minutes: feedback, both directions. Ask what you could have done better and let the silence sit. Then offer yours, gently, on anything that would help their next supplier - late approvals, unclear decision-making, whatever it actually was. Most clients have never been given this and value it.

Five minutes: the ask. Flag that you will follow up about a testimonial in a couple of weeks. Signalling it here means the later email is expected rather than a cold request.

Offboarding a retainer is different from a project

A project has a natural end. A retainer does not, which changes three things.

There is no natural handover moment, so you have to create one. When notice is given, put a wrap-up session in the diary immediately - otherwise the final month drifts and the relationship ends on the last invoice.

Access and continuity matter more. A retainer client has usually had you inside their systems for months or years, and there is far more institutional knowledge to transfer. The handover pack needs a section on things you have been quietly maintaining that nobody has thought about since month two.

The reason for ending is usually about value, not completion. Which means the feedback conversation is more useful and more uncomfortable. Ask it anyway - a retainer that ends is telling you something about your pricing, your reporting, or your visibility, and it is often the reporting.

The 90-day check-in

The final touchpoint is the one most likely to be skipped and the one with the clearest commercial return.

Keep it genuinely short and genuinely useful:

Hi Sarah - it's been three months since launch. How are the demo numbers looking? Anything behaving oddly that you'd want a second pair of eyes on?

No pitch, no offer, no attachment. The reply does three things for you. It supplies the results data that turns a project write-up into a real case study, which you would otherwise never collect. It puts you in front of them at the moment their next requirement is forming. And it costs two minutes.

Repeat at six and twelve months. Agencies with strong repeat business are rarely running sophisticated account management programmes; they are sending three short emails a year to people they used to work with, and being present when something comes up.

Who should own it

The reason offboarding fails is almost never disagreement about its value. It is that no one is accountable for it at the moment it needs to happen.

Assign it to the person who ran the account, not to an operations function. They have the relationship, they were on the wrap-up call, and a testimonial request from a familiar name converts several times better than one from an unfamiliar one.

Then remove the dependency on their memory. The T-2 week signal, the T+1 week closure, the T+2 week testimonial request and the T+90 day check-in should exist as scheduled tasks with an owner and a date, created automatically when a project moves to its final phase. Anything relying on someone remembering in six weeks will not happen in six weeks - not through negligence, but because by then they are three projects deeper.

This is the whole reason offboarding is worth systematising rather than intending: every individual step is easy, and every one of them falls outside the window in which anyone is paying attention.

What to measure

Four numbers make the case for doing any of this, and they take minutes to track.

Testimonial rate. Testimonials collected ÷ engagements completed. Agencies with no process are usually under 20%. With one, 60-80% is normal.

Referral rate. Referrals received ÷ engagements completed, measured over the following year.

Re-engagement rate. Former clients who return within 18 months. This is the number that most justifies the 90-day check-in, and it is invisible unless you look for it.

Time to final payment. A clean close with a clearly-marked final invoice gets paid faster than a fade. Compare before and after.

When the ending is not amicable

Not every engagement ends well. The process still matters, and arguably more.

Deliver the handover pack anyway. Completely, professionally, without commentary. It is the most effective possible response to a bad ending, and it is the thing the client will remember when the frustration fades.

Do not litigate in writing. The temptation to send an email explaining what actually went wrong is strong and always counterproductive. It will be forwarded, and it will be the artefact that defines you.

Close access promptly and confirm it. Especially here.

Do not ask for a testimonial. Obviously. Do ask for feedback, once, sincerely, and take it seriously - "I'd genuinely value knowing what we could have done differently" is sometimes the message that repairs the relationship.

Write the internal post-mortem. The most valuable output of a bad ending is what you learn. Scope, communication, fit, approvals - most bad endings are traceable to something that was visible early and not acted on. Our guide to difficult client conversations covers the interventions that prevent it reaching this point.

An agency that ends a difficult relationship gracefully quite often gets the referral anyway. The client knows the engagement did not work; what they will describe to others is how you behaved when it did not.

The testimonial email that actually works

Worth spelling out, because the phrasing changes the response rate more than the timing does.

Hi Sarah,

Now the site's been live a few weeks and you've seen the numbers settle, would you be happy for us to quote you on our site?

To save you writing anything, here's a draft based on what you said on our wrap-up call - please edit it however you like, or bin it and write your own if you'd rather:

"We came to SyncHQ with a site that our sales team apologised for. Ten weeks later demo requests were up by a third. What stood out was that they pushed back on ideas that wouldn't work rather than just building what we asked for."

Happy for it to be attributed to you and Northwind, or anonymised to "a B2B software company" - whichever you prefer.

Four things are doing work here. It arrives at the right moment. It removes the blank page. It quotes things they genuinely said, so it is their view rather than yours. And it offers the anonymised option, which converts the small number of people whose company policy prevents named endorsement - a group that otherwise just does not reply.

Send it from the person they worked with, not from a marketing address.

Building the process

Four things to create once:

A handover pack template. Structure, section headings, standard language. Turns a day into an hour.

An offboarding checklist. Access removal, billing closure, data handling, asset delivery. It exists so the person doing it at 5pm on a Friday does not have to remember nine things.

Three email templates. The T-2 week signal, the testimonial request with the draft quote, and the 90-day check-in.

A calendar mechanism. The reason offboarding fails is that the later stages happen after attention has moved on. Whatever system you run projects in, the T+2 week and T+90 day touchpoints need to exist as scheduled tasks with an owner, or they will not happen.

That is the whole thing. Perhaps three hours to build, and it converts the end of every future engagement from a fade into a set of assets: a testimonial, a case study, a referral path, and a former client who is genuinely likely to come back.

Onboarding gets the attention because it is the start of the money. Offboarding is where the next engagement is either set up or quietly lost - and unlike onboarding, almost nobody does it well, which makes it unusually cheap to be good at.