Marketing Attribution

For service businesses the practical stack is small: call tracking numbers that map calls to sources, forms that record landing page and source, and a pipeline that carries that origin through to closed revenue. The output is one honest table — channel, spend, qualified leads, clients, revenue.

Attribution also protects good marketing from bad accounting. Organic work often converts through a branded search or a direct call days later; without source capture, paid channels take credit for it and budgets tilt the wrong way.

The failure it prevents is specific and common: cutting the channel that creates demand because the channel that captures demand gets the credit. Someone sees a truck, hears a recommendation, or reads a guide, then searches the business by name and clicks a brand ad. Last-click reporting records a paid conversion. Cut the paid ads and the conversions reappear as organic; cut the demand creation and everything falls together, several months later, in a way no report connects back.

The AI channel makes this materially harder. An assistant that recommends a business sends no referrer, no UTM, and frequently no click at all — the customer reads a name and searches for it separately. That traffic lands as branded search or direct, and no analytics configuration will label it correctly. The honest instruments are indirect: branded search volume over quarters, and asking new customers how they found you.

Which argues for keeping the standard low and the habit consistent. A single "how did you hear about us" field, asked at intake and recorded in the same place as revenue, beats an elaborate multi-touch model that nobody maintains. Attribution is only useful if it survives contact with a busy front desk, and the perfect version of it does not exist for this mix of channels.