Paying more for less: the quiet squeeze on estate agents

For more than two decades, the relationship between estate agents and the major property portals has run on a simple premise: agents pay, portals deliver eyeballs, listings convert. It was never a friction free arrangement but the maths broadly worked. A line item in every branch’s P&L bought the reach that no agent could realistically replicate on their own.

That premise is now under pressure from two directions at once. Portal fees have continued their long, upward march with annual increases that have outpaced inflation, transaction volumes and in many cases the revenue growth of the agencies paying them. At the same time, a more fundamental shift is underway in how buyers and sellers actually find property in the first place. The default starting point for a search is no longer guaranteed to be a portal homepage or a Google query that lands on one. Increasingly, it is an AI-powered assistant that answers the question directly, summarises what’s available and only sometimes surfaces the underlying source.

Taken together, these two trends point to an awkward conclusion that deserves a clear-eyed look: estate agents may soon be paying more money for less visibility than ever before.

Two trends, one collision

The first trend is well understood inside the industry, even if it is rarely discussed openly with consumers. The dominant players have used their pricing power to push fees higher year after year. The justification typically offered — more features, more reach, more data — has become harder to defend as the rate of price increases has decoupled from the rate at which any of those things are actually improving for the average agent.

The second trend is newer and until recently, easy to dismiss as a future problem. Generative AI tools have moved very quickly from novelty to default. A material and growing share of consumer search now takes place inside conversational interfaces that aim to answer the user, not point them somewhere else. Even traditional search engines have re-architected their results pages around AI-generated summaries that reduce the need to click through. The clicks that do still happen are increasingly going to whichever source the AI chose to surface, on its own terms.

Either trend in isolation would be a manageable headwind. Together though they create a structural squeeze. The cost of a listing on a major portal continues to climb, while the underlying value of that listing — measured in actual buyer eyeballs that originate from a portal-led journey — is starting to erode.

How portal economics quietly unravelled

The original portal value proposition was reach plus efficiency. Sellers were on portals because buyers were on portals; buyers were on portals because that’s where the stock was. Agents were the connective tissue, paying for the privilege of putting their inventory in front of the demand the portals had aggregated. For a long time, that was a fair trade, even when the price went up.

What has shifted is not the logic of the model but the leverage within it. As portal pricing has scaled, agents have looked for alternatives and found that there genuinely aren’t many. Listing on more platforms means writing more cheques. And in markets where one or two portals carry the bulk of demand, the negotiating position of any individual agent — particularly an independent — is close to nil.

The result is a recurring cost line that grows reliably each year, regardless of whether transaction volumes, average fees or branch profitability grow with it. In good markets, the increase is absorbed. In quieter markets, it bites hard. And in either case, it sets the bar that any alternative channel has to clear.

What changes when search becomes AI-led

For agents, the practical question is not whether AI search is “real” — it plainly is — but how it changes the shape of buyer journeys.

Portal economics depend on the portal being the destination. The buyer arrives, browses, filters and lands on a listing. Every part of that funnel is monetisable and the agent’s listing is part of the inventory that makes the destination valuable. AI-led search does something subtly but profoundly different: it treats the entire web, including the portals, as source material. The buyer asks a question, the AI synthesises an answer and the listings or market context that come back are filtered through whatever the model judges to be relevant.

Three consequences follow.

First, the brand of the portal becomes less load-bearing in the buyer’s experience. A buyer who used to “go to the portal” may now “ask the assistant.” The portal logo is no longer the front door. It might appear as a citation, a link or — increasingly — not at all.

Second, the data and content that gets surfaced is whatever the AI can reach, understand and trust. That is not always the portal listing. It can be the agent’s own website, a property’s individual landing page, a local market commentary or an authoritative third-party source. Agents who have invested in their own discoverable, structured content have a meaningful advantage in this environment. Agents who haven’t are increasingly invisible outside the portals.

Third, the click-through dynamics flatten. If the AI answers the buyer’s question directly, the buyer may never click through to the portal at all. The visibility the agent paid for, on the assumption that it would translate into traffic, simply doesn’t generate the journey it used to.

The maths agents should be running

For most agencies, the portal line is one of the largest controllable marketing costs on the P&L. It deserves the same scrutiny as any other significant spend.

The honest analysis has three parts. The first is straightforward: what has the portal cost grown by over the last three to five years, in absolute terms and as a percentage of branch revenue? The second is harder but more important: what proportion of new instructions, applicant registrations, and viewings can actually be attributed to portal-originated leads, rather than being claimed by them through last-click attribution? The third is forward-looking: as a meaningful share of consumer search migrates to AI-led journeys, what is the realistic trajectory of portal-originated leads over the next two to three years?

Few agencies will like the picture that emerges. The cost line is reliably going up. The attributable benefit is, in many cases, already softer than the headline numbers suggest. And the structural tailwind that made portal economics work — the portal being the buyer’s default destination — is no longer as reliable as it was.

What a measured response looks like

None of this is an argument for unilateral disarmament. The portals still aggregate enormous demand, and for the next several years they will continue to be a meaningful part of the discovery layer. Pulling listings purely on principle is not a strategy.

The measured response is to rebalance. That means investing in the assets that AI-led search rewards — fast, structured and verifiable. It also means treating portal spend as a negotiable, reviewable cost rather than a fixed tax, and being prepared to have harder conversations with portal account managers as the value equation shifts.

It means, more broadly, accepting that the next decade of property marketing will not look like the last one. The agents who quietly do the work now to be discoverable in their own right — outside the portal walls, and inside the AI tools their buyers are already using — will find themselves in a stronger position when the squeeze gets tighter, as it almost certainly will.

The portals built a very effective tollbooth on a road that, until recently, every buyer had to travel. The road is starting to fork. Paying more, every year, to maintain a presence on a stretch of it that fewer buyers will use is a position no agency should drift into without noticing.

— Matty