A mobile number the customer typed themselves is the best key you will ever get.

Most lead files hold a number somebody entered into a form, which can be mistyped, deliberately wrong or a work number for a personal purchase.

A texting conversation only exists because the customer entered a number they actually use and then received a message on it.

That makes this one of the highest-yield lead files in the catalogue, and worth exporting even when another system is the primary source.

Match rates against a customer record are correspondingly high, which means less revenue disappears into Direct / Unknown for want of a usable key.

Where the sale system holds an older number, normalisation runs on both sides and the report shows which key joined rather than dropping the row.

When a call goes unanswered and an automatic text follows, the resulting conversation is with somebody another channel already produced.

Counted as a source, it takes credit for revenue that search, the map listing or a referral actually generated.

The honest treatment is to label it as recovery and report it separately, which also gives the real number: how much revenue was recovered from calls nobody answered.

That figure is usually large enough to justify the tool on its own, and it is a staffing argument rather than a marketing one.

Pooling it into a single messaging channel hides both findings at once, and it is the default configuration in most accounts.

A review programme raises a local rating, and the rating changes how many people choose you from a map result weeks later.

That effect is real and untraceable at the individual level, because no customer in your file interacted with somebody else's review in a way anything recorded.

The defensible reading is to measure the map and directory channel accurately on revenue and watch it against review activity over months.

Any tool claiming to attribute revenue to a specific review is guessing, and a guess presented as a measurement is worse than a stated gap.

The report does the measurable half properly and declines the other half explicitly.

Conversation volume and response times are reported per location because that is how the tool is organised; revenue usually is not.

A site with excellent response times and poor conversion looks like a success, and the group average conceals both it and the site subsidising it.

Where the export carries a location, the report splits revenue on it and the sites separate on the only measure that matters.

The usual finding is operational — one location answering slowly while the group average looked fine — which is cheaper to fix than any channel change.

Response time is the variable to look at first, because in a texting channel the customer is holding the phone. A reply within minutes and a reply the next morning are not the same business, and the revenue difference between them is usually larger than the difference between any two sources in the file.