Which channels produced signed repair contracts?

Every channel produces inspections. Only some produce homeowners who sign, and the proceed rate can differ threefold between two channels with identical costs per inspection.

Ranking on signed contract value rather than inspections booked is what makes that visible, and it is normally the difference between a channel worth doubling and one worth stopping.

Structural repair is expensive and rarely urgent enough to force an immediate decision. Many homeowners get an inspection, sit on it, and sign the following season.

There is no attribution window, so a contract signed a year after the inspection still credits the campaign that produced it — which is precisely the revenue currently sitting in Direct / Unknown.

Foundation work is almost never bought on the day it is first investigated. A homeowner notices a crack, books a free inspection, receives a quote that is larger than they expected, and then waits — often through a full season — until the problem worsens, a survey demands it, or they decide to sell the house and cannot avoid it any longer.

That produces one of the longest and most irregular gaps between inquiry and revenue in the whole trade. A quote issued in March and accepted in November is entirely ordinary, and no window-based measurement will connect the two, so the March campaign is recorded as spend with nothing to show for it.

Matching the invoiced job back to the household who booked the inspection removes the problem completely. The gap is not modeled or decayed; the same person simply appears in two files, and the revenue is credited to the period the inquiry came from.

The expensive thing in this trade is not the advertising, it is the inspection. A structural technician driving to a property, spending an hour under it, and producing a detailed quote is a significant cost that is incurred before anybody has agreed to anything, and it is incurred on every inquiry regardless of whether the household ever intended to proceed.

Cost per lead treats a serious inquiry and a curious one as identical. A channel that fills the diary with homeowners who wanted a valuation for a sale they are not making is costing far more than its media spend, and on every dashboard available it looks like the cheapest source in the account.

Ranking channels on invoiced repairs makes that visible, and uploading the inspection list alongside the invoices lets the report show inspection-to-job conversion per channel — which is usually the number that changes how the diary is filled.

Only exact email and phone matches count on their own. The one weaker link the engine produces — the same national number under a different country code — is set aside as a decision for a person, and nothing else is: a phone number differing by a digit, an email under a different provider, a shared surname or address produce no link at all. None of it is folded into the revenue quietly. That design is the reason the headline figure can be carried into a budget meeting.

The work that does need a person is reading the matched rows behind a surprising figure, and the person best placed to do it is rarely the person who buys the advertising. Confirming that an inspection booked in March and a pier installation invoiced in November belong to the same property needs somebody who recognizes the address and the technician's note, which means the office manager or whoever runs the schedule. An agency reading the same rows is working from strings; the office is reading its own records.

Anything in the review queue stays out of the revenue until it is confirmed, and confirmations can be undone. An entry left alone stays in Direct and Unknown, so an unreviewed queue understates the attributed revenue rather than inflating it: the failure mode is conservative on purpose. And because the match has no window, a household that decides to proceed a year after the inspection is still linked, which is precisely the pair a window-based tool would have dropped.