A buyer looking for a three-bed in your patch does not start on your website. They start on Rightmove, Zoopla or OnTheMarket, because that is where the stock is aggregated and where the property alerts they subscribed to months ago come from. Trying to outrank the portals for 'houses for sale in' plus your town is a fight you will not win, and it is the wrong fight anyway. Buyer traffic to an agent's own site is largely a service cost. It keeps registered applicants informed and gives your listings somewhere to point. It does not grow the business. Instructions grow the business. Every property you take on generates a fee once, brings buyer registrations with it, feeds board presence across the area, and produces the referrals that create the next valuation request. So the honest starting position for an estate agency website is that its job is to convert vendors and landlords, not to compete on search volume the portals were purpose-built to own.
That reframes the keyword set entirely. The searches that matter are 'estate agents in' your area, 'how much is my house worth', 'free property valuation' and the long tail of street-level and development-level queries where somebody is quietly checking what a neighbour's flat sold for. These are low volume compared with buyer terms and worth far more. They are also contested by the portals' own instant valuation tools, by the national online agents, and by every independent within three miles, which is why the ground that decides them is local rather than clever. A Google Business Profile tied to the real branch address, review volume that has kept moving in the last twelve months, and pages built around specific roads, postcode sectors and new developments will do more than a redesign. Two branches four miles apart are working different markets, and a single town-level page treats them as one.
The economics reward patience and punish impatience. A vendor who books a valuation today might instruct in a fortnight or in eight months, and plenty request one with no intention of moving for a year. Between instruction and completion there is usually another few months, so the fee that a January campaign produced can land in the accounts in June. Judge marketing on this month's cost per lead and you will switch off the thing that was working, roughly one month before it pays. Measurement that actually helps tracks valuation requests through to instructions won and fee income, and accepts the lag. It also separates the two ends of the market. The fee on a one-bed flat and the fee on a large family house come out of the same ad account, but the second will tolerate a cost per valuation that would look alarming against the first.