Does SEO keep working after you stop paying for it?

In December 2025 we told a client to stop spending money with us on SEO. The foundation was built, the numbers looked like they'd keep moving without us, and their budget was going to do more somewhere else. So we said so, and it went into social instead.

Eight months later their search performance was higher than it had been on the day we stopped.

All the numbers below come from the client's own Search Console exports, twelve months to 6 August 2026.

The short version: we recommended Norfolk Boards move their budget off SEO in December 2025. One of their two sites then improved every month from December to July. A search foundation can keep paying after the invoices stop.

What we told Norfolk Boards in December 2025

Norfolk Boards makes garden and indoor games. They came to us in December 2024 on a recommendation from a previous agency, and we built three new websites for them, including Cornhole.co.uk and Shuffleboard.co.uk. The SEO ran through to December 2025, deliberately aimed at the following year instead of the one it was happening in.

In October 2025 they went through Think First, our strategy workshop, which found that the real constraint on the business was the calendar and not the marketing. A British outdoor-games company has a few good months and then a long wait.

So in December we sat down and looked at where the next pound should go. Our read, and theirs, was that the search foundation was in good shape and projected to grow on its own. Social was doing nothing and had more headroom. We recommended moving the budget off SEO, which meant recommending they pay us less for the thing we'd been paid to do.

Most agencies never have to find out whether that kind of call was right, because they never make it. We wrote down what we expected to happen, and then we watched the account to see whether it did.

What Shuffleboard did next

Shuffleboard.co.uk is the smaller of the two measurable sites, and it's the one that carries the argument.

Client Search Console export, twelve months to 6 August 2026, so the last full month in it is July 2026.

The part that matters is not the two ends of that table, it's the shape of the line between them. In the client's export, no month went backwards after December. The climb runs straight through the months we were no longer being paid for, and July finished above where December sat.

Over the full twelve months the site took 1,861 organic clicks.

Was it just the season?

Both of these brands sell seasonal products. Google's own list of reasons search traffic moves puts "seasonality and changing interests" alongside algorithm updates and technical faults (Google Search Central), so it's the first thing to rule out and the easiest thing to get wrong.

For Shuffleboard, the calendar can't have done it. Shuffleboard is an indoor game and UK demand for it doesn't peak in summer. Semrush puts "shuffleboard table", which is unambiguously the buying query, at roughly 1,600 UK searches a month, with its trend index hitting an annual high in December and falling to its low in June. The broader "shuffleboard" term, at roughly 27,100 a month, has the same shape. So the site was climbing hardest through the months when fewest people were looking, which is the reverse of what a seasonal explanation needs.

We checked the tool before we trusted it. Semrush doesn't label its trend index with months, so we ran the same report on four keywords whose season nobody argues about, "bonfire night", "hay fever", "mulled wine" and "pancake recipe", and lined the shapes up against the calendar. All four landed exactly where they should, so the shuffleboard curve can be read the same way. Semrush UK database, 13 August 2026.

Did the second site show the same thing?

Cornhole.co.uk is the bigger of the two, and it moved the same way. Comparing its 2026 selling season against the same months of 2025, its average position went from 10.6 to 6.9, impressions from 40,543 to 71,259, and monthly clicks from 1,041 to 1,156. Across both sites the group took 9,804 organic clicks from 768,096 impressions over the twelve months.

There's something in those middle figures most agencies would report and not read. Impressions rose about 76% while clicks rose about 11%, so a much smaller share of the people who saw the site went on to visit it. Visibility grew far faster than traffic did, which is the same gap we've measured on pages that rank without being clicked. If your reporting shows impressions climbing and your enquiries flat, that gap is what you're looking at, and nobody is going to point it out for you.

Do rankings switch off when the invoice does?

There's a mechanical reason none of this is surprising, and it's worth being clear about because a lot of people assume the opposite.

You've never been renting the position. Google's own documentation puts it plainly: "Google never accepts money to include or rank sites in our search results, and it costs nothing to appear in our organic search results" (Google Search Central). What an agency fee buys is the work, and the work leaves things behind. Pages, structure, the internal links between them, whatever other sites have said about you along the way. When the fee stops, none of that gets taken away.

What does change is that you stop moving while other people carry on. Our own read, from the accounts we watch rather than from any published study, is that decline after an engagement tends to be relative rather than absolute: the pages don't get worse, the results page around them gets better. So the question worth putting to your own account is whether your pages would survive the competitors who are still publishing, which depends on your market far more than it depends on your invoice.

How do you measure SEO ROI once the work has stopped?

You can't measure it against the year that didn't happen. There's no version of 2026 where Norfolk Boards did none of this and we can hold the two up against each other. What you can do is pick a comparison you're able to defend and then stick to it: the same months a year apart, on the same set of pages, for a business selling the same things.

Then count the right thing at the end of it. Search Console will give you clicks, impressions and average position, and it will never tell you what any of that was worth. Positions improving is not a return. An enquiry is, and the only place that number lives is the client's own sales record.

That's why you won't find a revenue figure on this page. We have the search data, we don't have the client's books, and we don't put a pound sign on somebody else's year. Plenty of agencies will. If you're being sold SEO ROI as a multiple, ask which of those two sources it came out of, and watch what happens.

What would have to be true for yours to behave the same way

"You have to keep paying or it all disappears" is the line the industry runs on, and this account is one place it wasn't true. Whether it's true for you depends on four things, and they're worth going through before your next renewal:

1. The work was foundational rather than a publishing habit. Site structure, technical state and product pages, not a content calendar that stops producing the day it's switched off.
2. Demand for the product held up. Nothing here helps if people stop wanting what you sell.
3. Nobody was out-publishing them in that window. In a market where three competitors are shipping weekly, standing still costs you more.
4. The rankings weren't propped up by anything that ends with the engagement. Anything you were renting, from paid placements to bought links, disappears the moment the renting does.

If you want to know which of those describes you, the quickest version is in Search Console. Compare the last three months against the same three months a year ago, then look at which pages are producing the clicks and how old they are. Traffic carried by pages you published this quarter needs next quarter. Traffic carried by pages that have been earning quietly for a year is a different kind of asset.

Would yours hold?

We told a client to stop paying us for SEO, wrote down what we expected to happen, and it happened. That's the sort of call you can only make if you're reading the account rather than protecting the retainer.

The question worth asking about your own site is whether the traffic you're paying for is being carried by pages that have been earning quietly for a year, or by pages you published last quarter. One of those survives you stopping. The other doesn't.

If you're weighing up whether search deserves your next budget, how much SEO costs in the UK sets out what the money buys, and why strategy should come before SEO covers the maths worth doing before you commit to any channel at all. And if you want somebody to look at your account and tell you honestly whether the foundation would hold without you, that's worth doing before the renewal date rather than after it.