Marketing attribution for small service companies

What attribution means for a ten-truck company

Attribution is working out which marketing produced each job. Large companies use multi-touch models and data teams. A small service company needs something simpler: for every booked job, a reliable answer to where it came from, so next month's money goes to what works.

Most of your leads arrive by phone, which is why phone tracking sits at the center. Pew reports that 98% of US adults own a cellphone, 91% a smartphone, and 16% are smartphone-only internet users with no home broadband^[3]. The customer searches, reads and calls on the same device, often without visiting your site at all.

Perfect attribution is not the goal. The goal is data good enough to make one better decision each month: which source to give more, which to fix and which to drop. A system that answers those questions for most jobs, kept up consistently, beats an elaborate model nobody maintains. The four parts below take a few hours to set up and a few minutes a day to keep running.

Part 1: tracking numbers for measured sources

Assign a tracking number to each source you pay for and want to measure: search ads, Local Services Ads, each mailer drop, door hangers, yard signs, truck wraps and any directory you pay for. Keep your main number on your website footer, Google profile and free citations, so your listings stay consistent. Our guide to how call tracking numbers work covers static and dynamic numbers in detail.

• One number per source, not per campaign, unless you will genuinely compare campaigns.
• Separate numbers for offline sources that run for years, such as wraps and signs, so you can still see their calls long after the launch.
• Name each number in the tracking platform exactly as it appears in your lead-source list.

Part 2: UTM tags on every link you control

UTM tags are short labels added to a link so your analytics can tell where a visit came from. Google says you should always use utm_source, utm_medium and utm_campaign, warns that missing tags show as (not set), and notes that values are case sensitive, so google and Google count as different sources^[1]. It recommends a strict naming convention, usually lowercase^[1].

Keep the sheet in a shared document and build every new link from it. Shorten tagged links for print with a QR code generator that you control, so a code on a truck keeps working even if you change tools.

Part 3: the CSR lead-source question

Tracking cannot see a neighbor's referral, a truck parked on the street or a sponsorship banner. People can tell you. Make lead source a required field on every new customer, with a fixed list, and give CSRs a natural way to ask.

• Use a dropdown with fixed options: Google search, Google map, Google ad, Local Services Ad, other website, referral (name), past customer, truck, yard sign, mailer, door hanger, social, event or sponsor, AI answer, other.
• Put the question near the end of the call, after the job is booked, not before.
• Never let CSRs skip it; an empty field is worse than “other”.
• Review a sample of recorded calls monthly to check the question is asked naturally.

If you record calls for review, follow consent law. Federal law allows recording when one party consents^[4], but some states, such as Florida, require all parties to consent^[5]. A short announcement at the start of each call covers both.

When the tracking and the customer disagree

A caller dials the number from your Google ad but tells the CSR she saw your truck in her neighbor's driveway. Both are true: the truck made her remember you, the ad made it easy to call. Record both.

• Tracked source: what the number, tag or form says. This is the last step before contact.
• Stated source: what the customer says. This is often the first step, the reason they chose you.
• When reviewing, credit demand creation channels such as wraps, signs and sponsorships from stated sources, and demand capture channels such as search ads from tracked sources.

This matters more as search changes. Pew's 2025 study found that when Google showed an AI summary, users clicked a traditional result in 8% of visits, against 15% without one^[2]. More customers decide on the results page and call, so your website analytics see less of the journey, and the stated source fills the gap.

Part 4: the monthly review

1. 01

   Export the month's jobs

   Pull every new-customer job with tracked source, stated source, booked or lost status and invoice value.

2. 02

   Add spend by source

   Ad invoices, print and postage, sponsorships, agency fees and a monthly share of one-off costs like wraps.

3. 03

   Build one table

   Rows are sources; columns are spend, leads, booked jobs, revenue and cost per booked job.

4. 04

   Look for three things

   The source with the lowest cost per booked job, the one whose cost rose most, and any source with an unusual number of “other” answers.

5. 05

   Make one change

   Move a small amount of budget, fix a CSR habit or retire a tracking number. One change a month is enough to learn from.

Measuring what you cannot tag

Some of the most valuable marketing for a service company never produces a click: a clean truck parked in a driveway, a crew that leaves a yard tidy, a banner at the Little League field. You will never track these perfectly, but you can measure them well enough to decide.

• Use stated sources as the main measure and watch the trend, not single months.
• Look for geographic lift: count new customers on streets near a burst of jobs, signs or door hangers compared with similar streets nearby.
• Run on-and-off tests where you can, for example door hangers around jobs in alternate weeks, and compare new-customer counts.
• Watch branded searches and direct calls to your main number, which tend to rise when awareness work is working.
• Give each offline asset a unique QR code and tagged link, accepting that only a minority of people will use it.

Example: what a pest control company learned in three months

Example (fictional): Greenline Pest added a required lead-source dropdown and tracking numbers on its door hangers and mailers in March. By June the review showed that door hangers placed around active jobs produced few tracked calls but were the stated source for a steady share of new quarterly-plan customers, many of whom called through the Google profile. Mailers to whole ZIP codes produced calls, but mostly one-time treatments.

The owner kept the door hangers, narrowed the mailers to streets near existing customers and stopped paying for a directory listing that produced no booked jobs in three months. None of this needed software beyond the tracking numbers and a spreadsheet.

Common mistakes

• Letting “Google” stand for search, maps, ads and Local Services Ads at once, which hides which one is working.
• Allowing free-text lead sources, which produce dozens of spellings of the same thing.
• Counting tracking-number calls from existing customers as new leads.
• Changing the UTM naming sheet halfway through a season.
• Reviewing attribution only when cash is tight, instead of every month.