Traffic is not the thing being valued
Almost every guide written about AI traffic loss is about recovering it. This one is about pricing it, which is a different question and the one you need answered if you are selling, buying, or deciding whether to hold. The starting point is that nobody values a website on visitor counts. Buyers value earnings and then adjust for how likely those earnings are to continue. A traffic drop matters exactly as much as it moves those two things, and the answer is frequently less than the chart suggests.
First, separate the traffic loss from the revenue loss
Put the two lines next to each other over the same period. If traffic fell 40% and revenue fell 38%, you have lost 38% of your value and roughly that much of your multiple, because the earnings base has genuinely shrunk. If traffic fell 40% and revenue fell 9%, something quite different happened: you lost visitors who were never going to convert, and the business underneath is close to intact. Those two sites look identical on an analytics screenshot and are worth very different amounts.
Then find out which pages lost it
This is the step that decides the number, and it takes twenty minutes in Search Console. Sort your pages by lost clicks and look at what they are. Informational pages — definitions, how-to guides, what-is explainers — are what AI Overviews absorb, because a summary can answer them completely. Transactional and comparison pages are far less exposed, because somebody choosing what to buy still wants to see options, prices and opinions. A site that lost its glossary and kept its reviews is in much better shape than the headline number implies.
Re-baseline the earnings before you apply any multiple
Do not average across the drop. Take the most recent stable period — three months if the decline has flattened, six if you have it — and treat that as the run rate. Then apply the multiple to that figure rather than to a trailing twelve months that includes pre-decline months you cannot repeat. Sellers resist this because the older number is larger, but a buyer will do it anyway, and presenting the pre-drop average as the earnings base is the fastest way to lose credibility in due diligence.
What buyers actually discount, and roughly how much
A sustained downward trend costs 8 to 12 points of multiple on its own, before any argument about cause, because a buyer is pricing the risk that the line keeps going. Whether the cause was an AI Overview, a core update or a lost affiliate programme matters less to the discount than whether the decline has stopped. A site that fell 40% and has been flat for five months is a fundamentally easier sale than one that fell 20% and is still falling, even though the second looks better on paper.
The argument that genuinely reduces the discount
It is evidence that the remaining traffic converts better than the traffic you lost. This is often true after an AI-driven decline, because the visitors who still click are the ones a summary could not satisfy, and they are further down the buying process. If your conversion rate rose while traffic fell, show that explicitly, month by month, alongside revenue per session. It reframes the story from decline to concentration, and it is the single strongest thing you can bring to the negotiation.
Value informational and transactional pages differently
Treat the page portfolio as two assets. Pages whose only monetisation is display advertising on informational traffic are the exposed part, and a buyer will assign them close to no forward value if the trend is still down. Pages that sit near a transaction, hold a community, run a tool, or carry original data are the defensible part, and they are what the multiple should really be applied to. Presenting the split yourself is more persuasive than letting a buyer discover it.
Whether being cited in AI answers is worth anything
Some buyers now check whether a site is named as a source inside AI answers for its core topics, on the theory that a cited site keeps a form of visibility even without the click. It is a reasonable thing to look at and worth testing on ten of your own core queries before you list. Be careful how much weight you put on it, though: citation does not currently monetise, and nobody can price it reliably yet. Treat it as a supporting argument rather than a line item.
Sell now or wait
Wait if the decline has not flattened, because you will be selling into an open-ended risk and pricing it accordingly. Wait if you can plausibly rebuild through channels that are not search — a list, a community, direct traffic — since diversification is worth several points of multiple in its own right. Sell now if the exposure is structural rather than fixable, meaning the site is mostly informational content answering questions a summary answers better. In that case time is working against you, and the best price available is probably the current one.
What to disclose
All of it, early, with the chart. A buyer running due diligence will find a traffic decline in the first hour, and a seller who did not mention it has told them something about everything else in the listing. Presenting the drop yourself, with the page-level breakdown and the revenue line beside it, is the difference between a discount and a dead deal. It also lets you frame the analysis rather than have it framed for you.
- Value earnings, not traffic — put the revenue line next to the traffic line first.
- Which pages lost the clicks decides the number: informational pages are exposed, transactional ones are not.
- Re-baseline on the most recent stable period, not a trailing twelve months spanning the drop.
- A sustained decline costs 8 to 12 points of multiple; a decline that has flattened costs far less.
- Rising conversion on falling traffic is the strongest argument available. Show it month by month.
Work it out: Profit multiple calculator
Run a free appraisal on your current profit rather than your old traffic — the tool prices from earnings and shows which factors move the multiple.
Value my site →