More work from the clients you already have

You delivered a project in March. It went well, the client said so in writing, the final invoice was paid without chasing. It's now September and neither of you has said anything since.

That client is the through-line of this post. Over those six months you probably spent a working week on prospecting: research, messages, calls with people who had no budget. The buyer who already knows your work, already has your bank details on file and already got a result got none of that attention.

This post is not about finding new clients. Every other post in this series covers that. This one is about the revenue sitting inside relationships you already own, which is the cheapest revenue available to you and the one almost nobody has a process for.

Both big marketplaces now grow by spending more per client, not by finding more

The clearest evidence for the argument comes from the platforms whose whole business is new-client matching.

Fiverr's annual active buyers fell to 2.7 million in Q2 2026, down 21.9% year over year, while annual spend per buyer rose 15.6% to $368 (Fiverr Q2 2026 results). Upwork reported 784,000 active clients in Q1 2026, down 3% year over year, with gross services volume per active client up 5% to $5,138 (Upwork Q1 2026 results).

Two different companies, same shape: fewer buyers, more spend each. The businesses built on acquiring clients are increasingly living on depth inside the clients they already have.

Your economics point the same way, more sharply, because you have no marketing budget and your acquisition cost is paid in unbillable hours.

Expansion doesn't happen because nothing triggers it

Prospecting has a trigger, and it's a bad one: an empty calendar four weeks out. Panic is at least reliable.

Expansion has no trigger at all. Nothing tells you that the client from March is planning next year's budget in October. Nobody sends a calendar invite for "the moment your buyer's problem recurs." So the work goes to whoever happens to be in front of them, and by the time you think about it, it's been six months and getting back in touch feels like it needs a reason.

The fix is unromantic. You put the trigger in your own calendar and it fires whether or not you feel like it.

Four touches a year per past client. For twelve past clients that's forty-eight short messages annually, or roughly one a week. Set against a week of prospecting per quarter, it is not a large investment.

1. The follow-on: name the next project while you still have their attention

Every delivery creates the next problem. You rebuilt the checkout, so now the analytics don't match the new funnel. You wrote the pricing page, so now the sales deck contradicts it. You know this at handover, when you have more context about their business than you will ever have again.

Most freelancers say nothing at that moment, because naming more work at the point of delivery feels like upselling. The alternative is that the observation goes unsaid and the work goes to someone with less context in four months.

Put it in the handover, not in a separate pitch:

Subject: [Project] — handover, plus two things I noticed

Everything's in [location]. Short summary of what changed and why is at
the top of that doc.

Two things I ran into that are outside what you asked me for:

1. [Specific observation]. Left alone, [concrete consequence]. Roughly
[duration] of work to fix properly.
2. [Second observation]. Less urgent — worth looking at before [event or
season].

No need to decide anything now. If either becomes a priority, tell me and
I'll scope it properly.

What each part is doing:

"Plus two things I noticed" in the subject. The handover gets opened anyway. This attaches the next conversation to a message that was already going to be read.

Two items, not five. Two reads as observation. Five reads as a sales document, and the buyer starts discounting all of them.

The consequence, stated concretely. "The analytics will keep misattributing paid traffic" is a problem. "This could be optimised further" is not. Without a consequence there's nothing to prioritise against.

The rough duration. A range is enough. It lets the buyer decide whether this is a next-quarter conversation or a next-year one, without needing a proposal from you first.

"No need to decide anything now." Removes the pressure that makes the message feel like a pitch. It also means the follow-up is legitimate later, because you explicitly left it open.

Send this on handover day. The hours you spent noticing those two things are real hours, and if you tracked the project they're already logged. Worklyn's Money on the Table flags hours worked but never billed, which is usually where the evidence for a follow-on scope lives.

2. The next budget is one floor away

Your buyer controls one budget. The company has several. Marketing, product, sales enablement, internal comms and HR all commission the same kinds of work from different pots, and they rarely compare notes about suppliers.

An internal introduction costs your buyer almost nothing. They are not vouching for you to an outsider, they're recommending a supplier who already made them look competent. That is a comfortable thing to do.

The ask is one line, and it belongs in the 30-day result check rather than in the handover:

"The [deliverable] seems to be doing what we wanted. Out of interest, does anyone in [adjacent team] deal with the same problem? Happy for you to pass my name along if it's ever useful."

Two things make this work. It arrives after a result, not after a delivery, so there's evidence attached. And it names a specific team, because "anyone else in the company" is a question nobody can answer.

If you work as a small studio rather than alone, this is where capacity becomes the constraint rather than demand. Two departments at one client can be a bigger book than eight one-off buyers, and the operational difference is described on the page for studios.

3. Converting repeat work into a retainer

The client from March who now sends you something every six weeks is a retainer that hasn't been named. Naming it changes three things: your cash becomes predictable, their approvals stop happening one at a time, and the conversation about scope moves to the front instead of arriving as an argument.

The move itself is small. At the third or fourth ad-hoc request, say what you're already seeing: "This is running at roughly two days a month. Want to fix it at that and stop doing a quote each time?"

The structure you pick matters more than the pitch. An hours bank, a capacity reservation, a deliverable subscription and a pure availability fee behave completely differently when the client under-uses the month, and only some of them genuinely stabilise your cash. That's mechanical enough to need its own treatment, and it gets one in retainers: which structures actually smooth your cash flow.

One thing to settle before you propose it: what happens to unused hours. If they roll over indefinitely, you have sold a liability that gets called in during your busiest month.

4. The annual review, which is really the rate conversation

Once a year, in the same month each year, ask the client for 30 minutes. Not a check-in. A review, with an agenda they can see in advance.

Subject: Annual review — 30 minutes in [month]?

We've been working together roughly a year. I'd like to do a short review
rather than just carrying on.

Three things:
- What we shipped this year and what it produced
- What's on your roadmap for next year and where I could be useful
- My rates for [year], which change from [date]

30 minutes, [two or three time options]. I'll send a one-page summary
beforehand so we're not reconstructing it live.

"A review rather than just carrying on." Frames it as governance, which senior buyers are used to and freelancers rarely offer.

"What it produced," not "what we did." Forces you to prepare the outcome, which is the only version of your work that survives a budget discussion you aren't in the room for.

Rates listed as the third item, with a date. They're on the agenda, so nobody is ambushed, and they come after the value discussion. A rate change announced with a future effective date is an operational fact. The same number raised in the middle of scoping a project reads as a negotiation.

"I'll send a one-page summary." Does most of the work of the meeting and gives the buyer something to forward upward.

On the rate itself: holding a number flat is a cut, and the size of the cut is measurable. Consumer prices rose 3.5% in the twelve months to June 2026 in the US (BLS), 2.6% in the UK (ONS), and euro area annual inflation was estimated at 2.9% in July 2026 (Eurostat flash estimate). A US rate unchanged across that year buys about 3.4% less than it did, arithmetic you can redo with your own figures. Two years flat and the erosion is roughly double.

If you haven't recalculated in a while, the rate calculator works backwards from target income, billable hours, weeks off, monthly costs and your tax set-aside, which produces a defensible number rather than a hopeful one.

Why this work is worth more per hour than it looks

Two clients, same fee. Illustrative numbers, but the ratio is the point.

The new client took eleven unbilled hours to win: research, two calls, a proposal, a revised proposal, a procurement form. Then the project itself carried the usual first-time overhead, because you were learning who signs off and how many of them there are.

The returning client took one message and a 30-minute call. You already know their approval path, their file formats, their invoicing quirks and which stakeholder appears at week three with opinions. Scoping is faster and more accurate, so overrun is smaller.

Same fee, ten fewer hours, less estimation risk. Run both through the effective hourly rate calculator with every hour the project consumed, including the winning, and the gap is usually larger than people expect.

Which is also why per-client history is worth keeping properly. Knowing what a client paid you across twelve months, how many hours it took and where the margin ended up is what makes the annual review a conversation rather than a guess. A timer alone doesn't produce that, which is the comparison the Toggl alternative page makes: tracked hours are only useful once they turn into invoices and then into a number per client.

For where new clients come from when you do need them, the channel-by-channel ranking is in where freelance work actually comes from in 2026.

Thirty minutes now, and the review you should book this month

In the next thirty minutes: list every client you've invoiced in the last 24 months and the date of the last thing you sent them. Anything over 90 days old gets one message today — an observation about their business, no ask attached. Three messages is plenty.

This month: put the four triggers from the table above into your calendar as recurring events, and book one annual review with the client who has paid you the most over the last year. That review is where the rate conversation happens, and it happens once whether you schedule it or not.

Worklyn's CFO Mode answers questions from your own numbers, including what each client has actually paid you and which hours never reached an invoice, so the annual review starts from a figure instead of an impression.