Most marketing agencies turn down website work. Not because they don’t want it — because they can’t staff it. A client asks for a new site, the agency says “we don’t really do build work”, and the project goes to someone else. That someone else then has a relationship with the client, and the next brief goes to them too.
This is the story of one agency that stopped saying no. Names and identifying details have been changed at the client’s request, but the project mix, volumes and timelines are taken from our live delivery records.
The agency
A twelve-person marketing agency in the South East. Strong on brand, content and events. Two account directors, a designer, a content team, and a founder who still sold most of the work.
They had a healthy retained client base — Those clients were buying campaign work, brand refreshes and content. Almost every one of them also had a website that was three to six years old.
The founder’s summary when we first spoke: “We know the sites need doing. We just don’t have anyone to do them, and I’m not hiring a developer for work that arrives in lumps.”
The problem with lumpy work
That last point is the honest constraint most agencies hit. Website work does not arrive evenly. You get three briefs in a month and then nothing for six weeks. A full-time developer costs £45–60k plus employer’s NI, pension, equipment, holiday and management time — call it £70k all in. To justify that you need consistent billable work, which is exactly what you don’t have when you’re starting out in web.
So agencies do one of three things:
Turn the work away. Safe, and quietly expensive. You lose the revenue and, over time, some of the relationship.
Refer it out. You keep the client relationship for now, but you make nothing on it, and you have no control over quality or timelines.
Find a freelancer. Works until it doesn’t. Availability, communication and the bus factor
of one all become your problem.
They had tried option three. It worked for two projects and fell over on the third when the freelancer took on a full-time role mid-build.
What we set up
We started as their white-label build team on a single project — a microsite for one of their security clients. Small, defined, low risk. Fifteen hours of build against a design their team supplied.
Once that landed, we put a proper working pattern in place:
A single point of contact. One project manager at our end, one account director at theirs. No group chats with four developers in them.
Fixed quoting off a brief template. They send a scoped brief, we come back with a fixed hours figure and a delivery window, usually within a working day.
Their brand on everything. All documentation, staging links and handover notes carry their logo. We never contact the end client.
Design stays with them. They kept the part their clients hired them for. We took the build.
That last point mattered more than anything commercially. They weren’t outsourcing their value — they were outsourcing the bit that was blocking them.
What actually got built
Over the following eighteen months, the mix looked like this:
| Project type | Number delivered | Typical build hours |
| Brochure and corporate WordPress sites | 9 | 30–55 |
| Microsites and campaign landing pages | 4 | 8–20 |
| Large multi-phase builds | 2 | 85–95 |
| Small updates, fixes and page builds | Ongoing | 1–8 per job |
The two big builds were the ones that changed the numbers. One was a trade-show and exhibitor site rebuilt across three phases, running to just over ninety hours of development with a directory, filtered listings and a content migration. The other was a manufacturer’s group site with four product ranges, gated technical downloads and a dealer locator.
Around those sat a steady flow of smaller jobs: a podcast landing page, a set of microsites for a security brand, seasonal campaign pages, and a rolling stream of two- and three-hour fixes.
Total development delivered in the period: a little over 350 hours.
The first ninety days
For any agency thinking about this, the ramp matters more than the theory. Theirs looked like this.
Weeks 1–3. One microsite, fifteen hours. Deliberately small. The point was not the revenue — it was finding out whether we answered emails, hit dates, and produced something their designer was happy to put their name on.
Weeks 4–8. Two brochure sites quoted and started, both for existing retained clients whose sites were overdue a rebuild. The founder later said the hardest part was not the delivery, it was making the first phone call to a client he had never sold a website to.
Weeks 9–12. The first care plans went live, the brief template was written, and the account directors started raising websites in quarterly reviews as a matter of course rather than waiting to be asked.
By the end of the first quarter the offering was no longer an experiment. It was a line on the rate card.
The commercial model
Here is where agencies usually get stuck, so it’s worth being specific.
They quoted their clients at their own day rate — the same rate they charge for strategy and campaign work. They bought our time at a white-label rate. The margin between the two is the business.
The important thing is that the margin was not thin. Website builds are one of the few agency services where a partner can deliver at a genuine cost advantage — because we are doing WordPress builds all day, every day, and have a component library, a hosting stack and a QA process already paid for. A build that takes an occasional developer sixty hours takes us thirty-five.
On a monthly average across the eighteen months, the agency was invoicing roughly £8,000 a month in website revenue that had not existed before. In their strongest quarter — driven by the two large builds overlapping — it was closer to £13,000 a month. In the quiet stretch after those completed, it was £3,500. That variability is real and worth planning for.
*(Revenue figures here are the agency’s own reported billings, rounded. Build hours and project counts are from our delivery records.)*
What made the difference
Three things, in order of impact.
They stopped treating web as a favour. Early on, small site fixes were being absorbed into retainers as goodwill. Once those were priced and passed to us properly, a meaningful chunk of the monthly figure appeared out of work they were already doing for free.
They sold maintenance from day one. Every build went out with a care plan — updates, backups, security, a small monthly allowance of changes. Individually small, collectively the most predictable line on the sheet, and it renews.
Care plans deserve a note of their own. The temptation is to price them at £25 a month to make them easy to sell, which turns a recurring revenue line into an admin cost. They pitched theirs as a proper managed service — updates, off-site backups, uptime monitoring, security patching, and an hour of content changes a month — at a price that carried real margin over what the hosting and maintenance actually cost them. Clients took it because the alternative was managing a WordPress site themselves, and nobody in a marketing department wants that job.
They quoted quickly. Because they could get a fixed figure from us inside a day, they were coming back to clients with a number while the conversation was still warm. Speed of quote won them at least two projects they would otherwise have lost to a faster competitor.
What didn’t go smoothly
Two things worth flagging, because no honest case study is all upside.
The first was briefing. The first three projects each needed a round of clarification that cost a few days, because the briefs described the look of the site but not the functionality behind it. We built a one-page brief template together — content structure, integrations, forms, migration scope, hosting — and clarification rounds effectively stopped.
The second was expectation-setting on timelines. Their account team was quoting “two weeks” out of habit. A forty-hour build with client feedback rounds is four to six weeks, realistically. Once they started quoting our actual delivery windows, client satisfaction went up even though the dates went out.
What this means for your agency
You do not need a developer on payroll to have a web offering. You need three things: a partner who can quote fast and deliver on time, a brief template that captures functionality rather than just design, and the discipline to price the work properly instead of absorbing it.
The first project should be small. A landing page, a microsite, a set of fixes. Prove the process on something low-risk, then take on the build your best client has been asking about for a year.
If you want to talk through what that would look like for your client base — including what we’d charge you and what you could reasonably charge them — book a call. Bring a live brief if you have one and we’ll price it on the call.