Now onboarding AuriaTrack consultants · Become a Consultant →
Guide

Build an Attribution System
That Stays Simple and Still Tells the Truth.

You do not need a data team or a multi-touch model to know which ads pay for themselves. You need a few plain rules, applied in the right order, and kept up every month. Here is the build order we would give any small business starting from scratch.

What “Good Enough” Attribution Actually Answers

A useful attribution system answers three questions for every campaign: how much did it cost, how many real leads did it bring, and how much revenue did those leads become. If yours answers those reliably, it is doing its job. Everything else is refinement.

One list of leads

Every inquiry, from every channel, lands in one place. If leads live in an inbox, a voicemail box, a Facebook page and a notebook, there is nothing to attribute against.

A label on every lead

Each lead carries the source, campaign and click ID that brought it in, captured automatically at the moment it arrives, never typed in afterward from memory.

Revenue tied to the lead

When a job closes, the sale is connected to the original lead record, so the campaign that started the conversation gets the credit for the money.

A monthly habit

Someone reads the numbers once a month and makes a decision. A system nobody looks at decays quietly until the data cannot be trusted.

Six Steps, in the Order That Matters

Each step depends on the one before it. Skipping ahead — buying a dashboard before your leads are in one place, say — is how most attribution projects stall.

1
Pick one source of truth for leads

Choose the CRM or lead list that every inquiry will end up in, and decide that it wins every argument. A lead that is not in it does not exist for reporting purposes. This sounds administrative, but it is the foundation: you cannot measure cost per lead until you agree on what counts as a lead and where they are counted.

2
Tag every ad link consistently, using platform IDs

Put UTM parameters on every ad's destination URL with one naming convention: lowercase, no spaces, the same spelling every time. For the campaign value, use the ad platform's campaign ID (Google Ads, Meta and Microsoft Advertising all offer URL macros or parameters that fill it in) rather than the campaign name. Names get edited; IDs do not. Our UTM guide walks through the details.

3
Capture click IDs and UTMs into first-party cookies

When a visitor lands from an ad, store the UTMs and any click ID (Google's gclid, Meta's fbclid) in a cookie on your own domain, and keep a separate first-touch record that is written once and never overwritten. Most people do not convert on their first visit, so if you only read the URL at the moment of the form fill, the ad that brought them in is already gone.

4
Cover every lead path, not just the web form

Forms are the easy part. Calls need tracking numbers, ideally with dynamic number insertion so the number on the page matches the visitor's source. Chat and booking widgets need to pass the same stored values along. Lead-form ads on Facebook, Instagram and Google collect the inquiry inside the platform, so those leads need to be pulled or synced into your lead list rather than waiting for someone to export them.

5
Match closed revenue back to the lead

When a job is won in your CRM, invoicing or field-service software, connect it to the original lead record — by phone number, email or an ID carried through — so the campaign on the lead is the campaign that earns the revenue. Make sure the same sale cannot be counted twice if it arrives by two routes.

6
Review it monthly and act on it

Once a month, put spend, real leads, customers and revenue side by side for each campaign and platform. Decide what to cut, what to scale and what to test. Then spot-check a handful of leads to confirm their sources look right; that ten-minute check is what keeps the whole system honest.

Steps 1 and 4 are where most of the value is. A business that gets every lead into one list, with a source on each, already knows more than one with a sophisticated model fed by half its leads. Get coverage right before you worry about precision.

Where Simple Systems Go Wrong

None of these are exotic. They are the ordinary ways a reasonable setup drifts until the reports stop matching reality.

Over-engineering the model

Debating linear versus time-decay versus position-based credit is premature when a third of your leads have no source at all. Multi-touch modeling is a refinement for complete data, not a substitute for it.

The spreadsheet nobody updates

A sheet that depends on someone typing in the source and the sale amount works for the first month. By the third, half the rows are blank and the totals are fiction. If it is manual, it will lapse.

Renamed campaigns breaking reports

When reporting keys off campaign names, renaming “Spring Promo” to “Spring Promo – v2” splits its history in two, and older leads stop matching. Platform IDs survive any rename.

Ignoring the phone

For many service businesses the phone is a large share of inquiries. A system that only sees web forms will consistently undercredit whichever campaigns make people call.

Double-counting sales

The same job can arrive as a webhook from your CRM and again in a CSV import. Without a rule that one external sale ID is one sale, revenue per campaign inflates without anyone noticing.

Judging too early

Leads take time to close. Read a campaign's revenue from last month before deciding it failed, or you will cut the ones with longer sales cycles just as their jobs start landing.

Why First Touch Plus Revenue Beats a Fancy Model

The appeal of a sophisticated attribution model is that it seems to account for everything: the search ad, the retargeting ad, the email, the review site. In practice, every model is only as good as the touches it can see, and a small business's data has gaps — calls, cleared cookies, people who switch from phone to laptop. A model that divides credit precisely among the touches it saw can still be confidently wrong.

A simpler approach holds up better. Record where each lead first came from and where it came from most recently, keep both, and tie the closed revenue to the lead. That tells you which campaigns introduce new customers and which ones bring them back to convert. For most budget decisions — is this campaign worth what it costs? — that is enough, and it is verifiable: you can open any lead and see why it was credited where it was.

The other thing a simple system gets right is that it measures in your numbers, not the ad platforms’. Platform dashboards each count conversions under their own rules, and they routinely disagree with your lead list for reasons covered in Google Ads Says 30 Conversions. Your CRM Says 9. and Facebook Says 40 Leads. You Got 12. Those dashboards are still useful for optimizing inside each platform. Your own lead list and revenue are what you should use to decide where the money goes — see Platform ROI Versus Click Metrics for how to read the two together.

Once the basics are in place, there is one more step worth taking: send the outcomes back. Google and Meta both accept server-side conversion data, so you can report which leads became customers, with their value, against the original click. Their bidding then learns from closed jobs rather than form fills. That is a refinement on top of a working system, not a starting point — and if you want to see where you stand now, the free 5 Numbers guide is a good place to begin.

The Same Build Order, Already Built

Everything above can be assembled by hand. AuriaTrack is what it looks like when it is set up and run for you.

Codes from platform IDs

Campaign codes are derived from platform campaign IDs, not names, so renaming a campaign never breaks the reporting.

First-party click capture

The tracking script stores UTMs and click IDs in first-party cookies, with a write-once first-touch record.

Revenue synced in

Won deals sync in from GoHighLevel, Square, ServiceTitan, Housecall Pro, Jobber or a CSV and are matched to the lead that started them.

Monthly reporting

Every package includes revenue attribution reporting and monthly reporting & strategy, with ROAS shown at the campaign and platform level.

Call tracking with dynamic number insertion is part of the Ascent and Summit packages, and every package includes a GoHighLevel account. See how attribution works → or compare the packages on the pricing page.

What People Ask About Getting Started

What is the minimum I need for attribution to work?
One list where every lead ends up, a source recorded automatically on each lead, and closed revenue tied back to the lead. With those three in place you can calculate cost per lead, cost per customer and return on ad spend per campaign.
Do I need a multi-touch attribution model?
Not to start. Recording first touch and most recent touch on each lead, and tying revenue to the lead, answers most budget questions and can be checked lead by lead. Multi-touch models only help once your tracking is close to complete.
Why use campaign IDs instead of campaign names in UTMs?
Names change when someone edits a campaign, and every change splits the reporting. Platform campaign IDs stay fixed for the life of the campaign, so leads from before and after a rename still group together.
How do I attribute phone calls?
Use tracking numbers, one per source or with dynamic number insertion so each website visitor sees a number tied to how they arrived. The call is then credited like a form fill. Our guide on tracking phone calls from ads covers the options.
How often should I review attribution data?
Monthly is right for most small businesses: often enough to catch a campaign that is wasting money, and long enough for leads to turn into closed jobs. Spot-check a few individual leads each time to confirm the sources look right.

More on calls in How to Track Phone Calls From Your Ads.

Skip the Build. Keep the Answers.

Book a demo and we'll show you every lead, its source and the revenue it became — in one place.