How do I tell which parts of my service area are actually worth advertising in?

Revenue by zip code is the wrong picture

Two zip codes can produce identical revenue while one consumes far more of the day. Travel time is real capacity, and capacity is the scarce resource in a service business. A job forty minutes out can cost you the third job of the day.

The metric that captures this is contribution margin per technician hour including travel. Compute it per zip or per routing zone, and the map redraws itself. Areas that looked fine on a revenue map turn negative, and dense nearby areas you were underweighting turn out to carry the business.

Density beats distance

A distant zip with enough volume to fill a technician's day is efficient. The same zip with one job a week is not, because that job carries the entire round trip. The variable that decides it is job density, and density responds to marketing spend.

This creates a real strategic choice. You can retreat from a thin area, or you can spend deliberately to build density there until it routes efficiently. The second option requires accepting poor unit economics for a defined period, which is a decision an owner should make explicitly rather than discover in a report.

How to build the view

• Spend geography. Ad platforms report by targeted location, which is not the same as where the customer lives. Use the service address on the job as the authoritative geography and treat platform location data as approximate.
• Job geography. Service address from the field service system, normalized. Address quality is usually the weak link; standardize before aggregating or you will split one neighborhood across three spellings.
• Time. Dispatch timestamps give you on-site time. Travel time either comes from the same timestamps or from a routing estimate. Either is better than ignoring it.
• Margin. Modeled margin by job type is sufficient. You are ranking areas, not closing the books.

What to do with the map

Trim the targeting radius where margin per hour is negative and density is not improving. Raise bids where it is strong. And check the boundary cases before cutting: a zip that looks unprofitable may be feeding referrals or repeat work that shows up under a different source entirely.

For multi-location operators this analysis also answers territory questions that get argued about endlessly, including where the real boundary between two locations should sit. Building it requires job-level data joined to spend and to time, which is standard integration work and produces a view no ad platform can generate on its own. It belongs in your marketing reporting, not in a one-time spreadsheet.