For service businesses where a job takes 5 to 30 days to close, a position-based (U-shaped) marketing attribution model usually works best. It credits the channel that produced the first contact and the touch that turned the lead into a booked job. Pair it with an attribution window longer than your slowest close.
Most attribution reports were built for a shopper who clicks an ad and checks out four minutes later. Your customer doesn't buy that way. A homeowner finds you through a search on Monday, reads your reviews Wednesday, clicks a retargeting ad the next week, calls for an estimate, and signs nine days after that. If your report credits only the last click, you are moving budget based on the final step and ignoring how the lead started. This guide explains which models hold up across a multi-day close, which ones quietly mislead you, and how to set up tracking you can trust.
TruLata is an AI marketing platform for service businesses, including contractors, home services, clinics and other local operators. It counts leads from the business's own forms rather than from ad platform conversion counts and connects each lead to its source and outcome on one live screen.
Why do most attribution models fail service businesses?
Most attribution models were designed around ecommerce, where the gap between first click and purchase is short and the purchase happens on the website. Service businesses break both of those assumptions.
The conversion happens offline, days later
For a roofer, a plumber or a med spa, the website form or phone call is not the sale. It is the start of a sales process that includes an estimate, a site visit, a quote, follow-up and finally a signed job. That process commonly runs anywhere from a few days to a month. An ad platform that sees a form fill and calls it a conversion has measured a lead. It has not measured revenue.
Ad platforms grade their own homework
Every ad platform reports conversions using its own rules, its own lookback windows and its own definition of what counts. Add up the conversions claimed by Google, Meta and your directory listings and you will often get more leads than your inbox actually received. That is why an honest attribution model service business owners can rely on has to start from a lead count you control. Use the lead count from your own forms and phone system, not totals each platform reports for itself.
Short attribution windows erase the start of the journey
Many reporting setups default to a 30 day lookback, or shorter. If a commercial HVAC job or a full remodel takes six weeks from first visit to signature, the search or referral that started the conversation can fall outside the window. That channel then looks like it produced nothing, even though it brought in the customer.
What marketing attribution models are there, and how do they handle a long close?
Here is how each common marketing attribution model behaves when a job takes 5 to 30 days to close.
Last-touch attribution
All credit goes to the final interaction before the conversion. It is simple and it is the default in many tools. For a service business it tends to overcredit branded search and direct traffic, because people who already know your name search for it right before they call. The channel that introduced them gets nothing. We break this down in more detail in our piece on why the final click doesn't tell you what actually closed the job.
- Works when: most jobs close on the same day as the first contact, such as emergency calls.
- Fails when: customers research, compare quotes and come back later.
First-touch attribution
All credit goes to the first recorded interaction. This is useful for seeing which channels create new demand, but it ignores everything that kept the lead warm through the estimate and follow-up stage.
- Works when: your main question is "where are new customers discovering us?"
- Fails when: you need to know what actually moves quotes to signed jobs.
Linear attribution
Credit is split evenly across every touchpoint. It is fair in a blunt way and works reasonably well when many interactions each nudge the customer forward. Its weakness is that it gives a casual social impression the same weight as the estimate request that started the sales conversation.
Time-decay attribution
Touches closer to the conversion get more credit. This suits journeys where later steps such as a quote follow-up or a second call do the heavy lifting. The tradeoff is that the channel that first brought the lead in gets shortchanged, which can push budget away from the sources that fill your pipeline.
Position-based (U-shaped) attribution
The first touch and the lead-creating touch each receive a large share of credit, often 40 percent apiece. The remaining credit is spread across the middle interactions. For most service businesses this matches reality: one channel introduced you, one interaction turned interest into a lead, and the rest helped.
Weighted or data-driven multi-touch attribution
Credit is distributed across touchpoints based on their measured influence. According to Luth Research's overview of attribution for long sales cycles, multi-touch attribution is generally the best fit for extended journeys because it gives a comprehensive view of how various interactions contribute to conversions. Weighted versions assign more credit to the touchpoints with more influence on the decision. The catch for local service businesses is volume. Algorithmic models need a steady flow of conversions to learn from, and a company closing a few dozen jobs a month often doesn't have enough data for the model to be reliable.
Which marketing attribution model should a service business actually use?
Match the model to how your jobs actually close. Use this as a starting point:
- Same-day close (emergency plumbing, lockouts, towing): last-touch is acceptable, because the first and last touch are usually the same.
- 5 to 14 day close (HVAC replacement, roofing repair, cleaning contracts): position-based. Credit the channel that found the customer and the touch that created the lead.
- 15 to 30+ day close (remodels, new roofs, commercial service, elective medical): position-based or linear, with an attribution window set to at least your slowest typical close plus a buffer.
- High volume with clean data: consider a weighted multi-touch model once you have enough closed jobs each month for the numbers to be stable.
If you are unsure, start with position-based. It is easy to explain to everyone on your team, it doesn't need large data volumes, and it keeps you from defunding the channels that start conversations. For a contractor-specific walkthrough, see our guide to multi touch marketing attribution for contractors.
How do you set up lead source tracking that survives a long sales cycle?
No model can fix bad inputs. Good lead source tracking is what makes any attribution model worth reading. Here is the setup that holds up when a job takes weeks to close.
1. Count leads from your own forms and phone lines
Your source of truth should be the forms on your site and your tracked phone numbers, not the conversion column inside each ad account. When every lead is captured where it actually arrives, you get one number that doesn't double count. TruLata's lead tracking counted from your own forms works this way, recording each lead with its source and letting your team set the outcome.
2. Capture the source at the moment the lead is created
Store UTM parameters, referrer and landing page in hidden form fields so the source travels with the lead record. For calls, use dynamic number insertion or source-specific numbers. If you only record the source when the job closes, you'll be relying on memory, and "they found us on Google" can mean a paid ad, a map listing or an organic result.
3. Ask the customer, then compare
Add a "How did you hear about us?" question to your intake. Self-reported answers catch what tracking can't, such as a neighbor's recommendation or a truck wrap. When the tracked source says "paid search" and the customer says "my brother-in-law," you've learned that the ad captured demand a referral created.
4. Record the outcome, not just the lead
Mark every lead as booked, quoted, won, lost or junk. Attribution measured on leads alone rewards channels that produce volume. Attribution measured on won jobs rewards channels that produce revenue. Those are often different channels.
5. Set your attribution window to match your close
Pull your last few months of closed jobs and measure the days from first contact to signature. Set your window to cover the slow end of that range, not the average. If your remodel jobs routinely take five weeks, a 30 day window will cut off the start of the journey for many of them.
What does multi-channel attribution tracking look like in practice?
Good multi-channel attribution tracking puts ads, organic search, email, referrals and phone calls side by side, measured against the same lead count. The practical test is simple: can you answer "where did last month's won jobs come from?" in under a minute without opening five tools?
Here's a typical 12 day path for a water heater replacement:
- Day 1: Customer finds you in a local search result and reads reviews.
- Day 3: Sees your search ad and visits the pricing page.
- Day 4: Submits an estimate request form (lead created).
- Day 6: Technician visits and leaves a quote.
- Day 9: Opens a follow-up email.
- Day 12: Calls back and books the job.
Last-touch credits the phone call, which tells you nothing about marketing. First-touch credits organic search alone. Position-based credits organic search for introducing you and the search ad that drove the form submission, with the follow-up email getting a smaller share. That's the version that helps you decide where the next marketing hour goes.
To read this kind of path, you need your lead data, ad data and job data in one place. Your TruLata dashboard puts those workstreams on one screen and refreshes when you open it, so the numbers reflect the data available at that moment rather than a weekly export. If your jobs live in a field service tool, connecting it means the outcome sits next to the source. Our post on how every tracked call lands on your leads tab shows how phone leads join the same view.
What mistakes make attribution data unreliable for service businesses?
Trusting platform-reported conversions as lead counts
Each platform counts in its own favor. Reconcile against your own form and call records every month.
Counting junk leads as wins
Spam submissions, wrong-service inquiries and out-of-area calls inflate the channels that attract them. Mark them so they drop out of your attribution.
Changing models every month
Switching models makes trend lines meaningless. Pick one, run it for at least a full quarter, then evaluate.
Ignoring referrals and repeat customers
For many local service businesses, word of mouth is a major source of work. If your tracking only sees digital touchpoints, ads will get credit for jobs that a happy customer actually sent you.
Judging channels on cost per lead alone
A channel with cheap leads and a low close rate can cost more per job than a channel with expensive leads that sign. Always measure through to the won job.
How do you turn attribution data into budget decisions?
Attribution matters only if it changes what you do next. Review your numbers on a fixed schedule, monthly for most service businesses, and work through these questions:
- Which channels appear most often as the first touch on won jobs? Protect those, even if they look weak in last-click reports.
- Which channels appear most often as the lead-creating touch? These are your conversion engines. Watch their cost per won job closely.
- Which channels produce leads that rarely close? Investigate targeting, landing pages or follow-up speed before cutting them.
- Are any jobs closing outside your attribution window? If so, extend the window.
Make one budget change at a time and give it a full close cycle to show up in results. With a 20 day average close, a change you make today won't fully appear in won jobs for three to four weeks. Judging it after one week will mislead you.
See every lead, its source and its outcome on one screen
The right marketing attribution model is only as good as the lead count beneath it. TruLata counts leads from your own forms, ties each one to its source and outcome, and puts ads, search, email and jobs on one dashboard that refreshes when you open it. If you want attribution you can defend in a budget meeting, explore how marketing attribution in TruLata works, or sign up at login.trulata.com/signup and connect your accounts.
