Which channels produced signed remodels?

Remodeling attracts an enormous amount of early-stage research traffic. Most of it is years from buying, and a channel optimized for volume will produce almost entirely that.

Ranking on signed contract value shows which channel produces people who actually build. It frequently turns out that a smaller, more expensive channel produces most of the revenue.

A remodeling consultation involves a designer, measurements and a proposal. Channels producing consultations that never sign are not free leads — they consume the scarcest resource in the business.

Attributing on signed work makes that cost visible in the channel ranking, which is where a budget decision can act on it.

A kitchen or bathroom remodel is one of the largest discretionary purchases a household makes, and it behaves like one. The inquiry comes months before the contract; the decision involves a couple who do not always agree; and the homeowner will almost always get two or three competing quotes before signing anything. Six to nine months from first inquiry to signed contract is ordinary rather than slow.

That horizon defeats every window-based measurement on the market. The campaign that produced the inquiry has long since been paused or restructured by the time the contract is signed, and last-click reporting hands the credit to whichever branded search happened in the final two weeks before the deposit cleared.

Matching a signed contract back to the household who inquired removes the horizon entirely. The person who filled in the form in February and the contract signed in October are one match, and the February spend is credited with October's revenue rather than written off as a cost with no return.

The most valuable thing a remodeler spends is designer and estimator time. An in-home consultation, a measure, a design concept and a detailed quote can consume many hours before anybody has committed to anything, and a channel that fills the diary with homeowners who were never going to proceed is costing far more than its media spend suggests.

Cost per lead cannot see that, because it treats every consultation as identical. Ranking on signed contract value shows which channels produce households that actually build, and by implication which ones are consuming the design team on work that will never be invoiced.

Uploading both the consultation list and the signed contracts lets the report show the conversion rate between them per channel, which for most remodelers is the more actionable of the two numbers.

A remodel is bought by a couple and inquired about by one of them. The wife submits the form with her own mobile and personal email; nine months later the contract is signed and the accounts system records the husband, because he handled the deposit. Neither file is wrong and neither system is at fault, but the inquiry and the revenue now sit under two different people. A match that reads a person rather than a browser cookie will not join those two rows without being given something they have in common.

This is the most frequent reason a remodeler's match rate disappoints, and it is fixed in the export rather than in the software. A sales row has room for one phone number and one email, and the two do not have to belong to the same person: the homeowner's mobile and the partner's email give the engine two keys to test instead of one. A secondary phone or a site contact beyond those two is not used. Where nothing on the contract row appears in the inquiry file, the job is not linked at all — there is no queue of plausible-but-uncertain pairs — so filling both slots is what joins February's inquiry to October's contract.

Two small habits remove much of what remains. Ask for a mobile number as well as an email at the point of inquiry, because households that research by email frequently book by phone. Record the inquiry against the property as well as the person, so somebody checking a match can recognize the job. Anything unmatched stays in Direct and Unknown rather than being assigned on a shared surname, which keeps the attributed figure conservative.