You run ads. People buy things. But somewhere between “someone saw your ad” and “someone paid you money” lies one of the hardest questions in marketing: which ad actually caused the sale?
That question is what attribution tries to answer. Get it right and you know where to put your next dollar. Get it wrong and you end up cutting what's working and doubling down on what isn't.
Why attribution is harder than it looks
In a perfect world, every customer would see one ad, click it, and buy immediately. You'd know exactly which campaign drove the revenue. In the real world, a customer might:
- See a Meta video ad on their phone on Monday
- Search Google for your brand name on Wednesday
- Click a retargeting ad on Friday
- Open an email and buy on Sunday, on a different device
That's four touchpoints across three days and two devices. Which one gets the credit? Depending on what you measure, you could make a case for any of them. Each ad platform will happily claim the win for itself.
The overlap problem
When you add up the reported conversions from Meta, Google, and every other platform, the total is almost always higher than the number of actual sales in your payment processor. This isn't anyone cheating. It's the natural result of every platform applying its own attribution rules to the same customer journey. The clicks really happened. The sales didn't multiply.
The delay problem
Not every sale happens the same day as the click. High-ticket products, subscription services, and anything with a sales call in the middle can have a purchase cycle that stretches over weeks or months. A last-click model will miss the awareness campaign that started the whole journey. A 7-day attribution window will miss the sale that closed on day 14.
The main attribution models, in plain terms
Attribution models are just rules for how you divide credit. Here are the four you'll encounter most often:
First-touch attribution
100% of the credit goes to the first thing the customer ever interacted with. Good for measuring which channels bring new people into your funnel. Bad at explaining what actually convinced them to buy.
Last-touch attribution
100% of the credit goes to the final touchpoint before purchase. This is the default on most ad platforms, which is why your retargeting campaigns always look amazing in the platform dashboard. It ignores everything that built the relationship before the final click.
Linear attribution
Credit is divided equally across every touchpoint in the journey. It avoids the all-or-nothing distortion of first- and last-touch, and is a reasonable starting point for businesses with longer sales cycles.
Multi-touch attribution
A family of models (linear, time-decay, position-based) that all distribute credit across multiple touchpoints, each using a different logic. They give the most complete picture of which channels are working together to drive revenue, but they require good data across the whole journey to produce accurate results.
What good attribution actually requires
The model you choose matters less than the quality of the data behind it. Accurate attribution needs three things:
- First-party tracking. You need to capture visitor and click data on your own domain, not depend entirely on the ad platforms to report it back to you. That means a first-party pixel or server-side tracking that isn't blocked by browsers or restricted by privacy changes.
- A complete view of the journey. You need to connect the ad click to the lead, and the lead to the sale, even when those events happen in different systems (your ad platform, your CRM, your payment processor). If those systems don't talk to each other, your attribution model is working with incomplete data.
- A neutral system of record. The ad platforms have an incentive to report high numbers. A neutral first-party system that reconciles click data against actual payment records gives you numbers you can trust.
How to start
You don't have to solve attribution perfectly on day one. Here's a practical sequence:
- Get UTM tags on everything. UTM parameters are the foundation. If every paid link includes a source, medium, and campaign tag, you can attribute sessions to campaigns in even a basic analytics setup.
- Connect your payment data. Compare what your ad platforms report against what actually landed in your payment processor. The gap between those two numbers is your over-reporting problem. Measuring it is the first step to solving it.
- Track leads back to their source. If you have a sales process with a gap between lead and sale, you need a way to carry the original source through to the closed deal. Most CRMs can store this if you pass it in at the point of lead capture.
- Pick a model and stick with it. A consistent, imperfect model you always use is more useful than a perfect model you argue about every week. Linear is a solid default. Adjust as you learn more about your actual customer journey.
The goal isn't perfect attribution. The goal is a consistent, first-party view of your funnel that improves your decisions, even if it isn't complete.
Where Cavor fits in
Cavor connects your ad platforms, CRM, and payment processor into one unified view, runs your first-party pixel to capture click-to-revenue attribution across the full journey, and shows you blended ROAS and per-channel revenue against actual payments, not platform-reported conversions. It's designed to be the neutral system of record that makes attribution trustworthy for the businesses it serves.