How an engagement runs.

Measurement audit — weeks 1 and 2

The audit starts with access, not a meeting: read access to the ad accounts and analytics currently in use, and either an export or a read-only connection to whatever system records a booked outcome — a CRM, an EMR, a property management system. From that, we build a channel-by-location map showing what is tracked today against what the revenue system actually recorded, and name the specific gap between the two rather than describing it in general terms.

The output is a written document: which channels and locations have a real, auditable connection to a booked outcome, which are inferred or estimated, and which have none at all. It is scoped to two weeks regardless of how many locations or channels are involved, and it stands on its own — a business that stops here walks away with an accurate account of where its reporting can be trusted and where it cannot.

Build the layer — weeks 3 through 6

This stage requires a different kind of access: administrative rights to the ad platforms and call tracking system, since offline conversion import and call scoring both make changes inside those tools directly. Call scoring goes live first, against a fixed rubric agreed with the client’s team, followed by the CRM or EMR matching that connects a scored call or form to what actually got booked.

The output is attribution built to your revenue system: a cost-per-booked-outcome model, broken out by channel and location, feeding back into each ad platform’s own bidding algorithm through offline conversion import. Four weeks is typical; a stack with more than one CRM or a manual, non-digital intake process usually needs the full six.

Run the programs — week 7 onward

Campaigns that were previously judged on platform-reported clicks and conversions start reporting against the new standard instead, and budget moves accordingly — sometimes toward a channel that had looked mediocre under the old measurement, and sometimes away from one that had looked strong. The first full reporting cycle against the new layer typically produces at least one reallocation nobody would have made under the previous numbers.

That shift is not confined to one channel. It reaches every stage a program runs in — top-of-funnel programs, middle-of-funnel programs, and bottom-of-funnel programs alike — since all three report against the identical cost-per-booked-outcome standard from week seven onward.

This is also where the account team we recommend hiring or reallocating in-house starts working from the same standard the account itself is judged on, rather than a second, informal read built from whatever dashboard happened to be open. Creative tests, launch sequencing, and channel-mix decisions all draw on the same cost-per-booked-outcome figures the standing review reports each month, so a decision made in week nine and a decision made in month six are answerable to the identical number.

Standing review — ongoing

Reporting runs on a monthly cadence: cost per booked outcome by channel and location, checked against the previous month, with a decision log recording what changed and the specific data that prompted it. The cadence is fixed from the start; what changes over the first two cycles is how much of the reporting still needs caveats attached, as more of a location’s intake gets fully instrumented.

Access is the largest requirement, and it grows across the four stages: read-only for the audit, administrative for the build stage, since call scoring and offline conversion import both make changes inside the ad platforms directly. Beyond access, one person who can answer how intake actually works day to day — which forms route where, what a call sounds like before anyone scores it, where a lead can quietly get lost. That person does not need to be senior. They need to be the one who would actually notice if a step silently stopped happening.

Response time on access requests matters more than most clients expect going in. A build stage waiting three weeks on an administrative login runs three weeks longer, in full, since call scoring and CRM matching cannot be tested against real traffic until the access behind them actually exists. Setting a single point of contact for access requests before the audit even finishes is the fastest way to keep the six-week build window realistic rather than aspirational.

Sometimes the audit itself is the honest end point: a business with too little tracking infrastructure in place to build a real measurement layer on top of it, or one not ready to grant the administrative access the build stage requires. When that happens, we say so directly at the end of the audit rather than starting a retainer we do not believe will produce a trustworthy number, and point toward what would need to change first for the next stage to be worth doing.

A single-location business with a short, simple sales cycle is another common case: the audit still produces a useful account of what is tracked, but the layer built in stage two earns its cost fastest at multi-location or multi-channel scale, where a blended report otherwise hides which site or program is actually producing customers. Telling a single-location business that plainly, before spend, costs us a retainer and saves them one they would not have gotten full value from.