Broken attribution: tying your Meta ads to real sales when conversion happens offline

We've got a client who doesn't sell in one click. The prospect fills out a form or books a Calendly, and then there are human steps before they become a paying customer. Form filled, then appointment, then visit, then signature. A four-stage funnel with humans in it.

The problem is that Meta only sees what goes through Meta, and the client's business CRM doesn't track where sales come from. So we're flying blind on the campaigns. We put budget in, we see leads coming in, but we have no idea which ones turn into customers. And for arbitrating ad spend, that's the worst possible situation.

What marketing attribution actually is

Attribution is simply how you decide who gets credit for a sale, across the different touchpoints a customer went through before buying. A Meta ad, a Google search, an email, a direct visit. They all played a part in the journey, and you have to decide who deserves credit for the sale.

The thing is, every platform pulls the blanket towards itself. Meta claims the sale, Google claims the sale, your CRM claims the sale. End result, if you add up what each one tells you, you blow way past the actual number of sales. That's exactly what happens on a funnel that ends offline.

Why Meta is feeding you nonsense

When a customer clicks an ad and doesn't convert within 7 days, ad platforms don't report the ad's performance. That's the attribution window. On a business where the sale happens in one click, fine, it works. But here, the buying process is much longer. Between the moment the guy fills out the form and the moment he signs, a week can go by, two weeks, sometimes more.

So Meta tells you "you got 120 leads on this campaign." But it has no clue whether those leads bought. It optimises on the form filled, not on the paying customer that should actually count.

I've seen accounts where the lead count was exploding, everyone was happy, and when you looked at who actually signed at the end of the funnel, it didn't hold up. Unqualified leads from forms, that's super common on Meta. The volume climbs, the profitability doesn't follow.

Last-click attribution versus everything else

The most widespread attribution model is last click. All the credit goes to the last touchpoint before the purchase. It's simple to set up, but it completely distorts how you read your channels. It means the channel that warmed up the prospect for two weeks gets its sale stolen by the one the guy happened to click on right before signing.

There are other models. Last click, first click, linear distribution, and data-driven models that look at converted and non-converted journeys to distribute credit. I'll skip the model details, because in a business where the sale ends offline, none of these platform models really work. They all stop at the click or the form. They don't see the signature.

That's why running on any platform model whatsoever leads you straight into a wall here. Your only reliable reference stays your actual sale. The one that hits the till.

The gap in the net, cookies and cross-device

There's a second layer of problems that makes all this even murkier. Between the platforms each claiming the conversion and people refusing cookies, your reference point goes haywire.

On Meta, you can have 1,000 clicks and only report 300 page views. At first you think "the tracking is broken." Actually no. It's just that 30% of people accept the cookie banner. The rest, you don't see. The pixel only fires after the click on "accept," so Meta optimises on barely a third of your traffic. Before you even talk about the final sale, you've already lost two thirds of your signal at the top of the funnel.

Add cross-device on top. Your prospect sees your ad on their phone, and they sign from their computer two weeks later. To the platform, that's two different people. The link is broken. Between cookies disappearing and the jump from one screen to another, classic attribution models lose a good chunk of reality.

That's why it's better to steer at the global level, on your actual revenue, rather than on the platform view. I go deeper into this logic in the article on why an eCommerce audit should start with your P&L, because it's the same reasoning, you start from the real result, not from what the platform is willing to tell you.

The clean solution, dynamic UTMs

The solution is dynamic UTMs.

In practice, you insert dynamic variables into the destination URL behind each ad. The campaign name, the campaign ID, the ad ID. And you have them take the value lead, lead gen, traffic, or awareness depending on what the campaign is doing.

The upside, when a lead enters your CRM, you capture the name of the original Meta campaign. So you know whether it's lead gen or lead, and you can calculate the conversion rate and cost per channel all the way to the signed sale, not just to the form filled.

Concretely, you reconcile your data across your tools. Instead of trusting three sources that contradict each other, you rebuild a single one, based on the real source of each signed sale. In the end, only that cost per sale per channel matters. One channel can hand you cheap leads that never convert, and another can hand you pricier leads that sign three times more often. Without the source tracking, you cut the good one and keep the bad one without even knowing it.

The obstacle you're going to hit

Now the real world. The UTM already exists on the ads side, that's not the issue. The issue is that the business CRM doesn't pull it automatically, because the final sign-up is entered by hand by the sales reps.

The customer signs, sends their bank details and documents, and it's a sales rep who types the info into the system that day. Nowhere in that process do you capture the original UTM. The link breaks between the ad click and the final record.

Two details that matter here. First, UTM integration is not retroactive. If you put it in place after launching the campaign, the link doesn't happen on leads already in. You really have to anticipate and wire it up before launch. Second, even when it's wired up, if the final entry stays manual, you have to discipline the sales reps to report the source. Otherwise the pipe is laid but nobody's pouring into it.

When the CRM can't keep up, disciplined Excel or Trello

I've been told twice that an upstream CRM was useless and that an Excel sheet was enough. That's not entirely wrong on the tooling side. If your sales team is small and your tracking fits in a file, no need to pay 400 bucks a month for a CRM just to look nice.

Right now all the tracking for this client happens in a manual Excel sheet. "Text sent," "asked to sign up," follow-up at 3 days, follow-up at 6 days. For lead gen it's the same, automatic text, one call, two calls, and after the second missed call a text saying "couldn't reach you, re-register" with the Calendly link, then a follow-up at 2 days. It runs, the reps find their way around it.

Excel is enough for follow-ups, but not for arbitrating budget. The moment you want to know which channel pays off, you need one more column, the source of each sale.

Trello in Kanban or the Excel you already have, the principle is the same. Each rep notes where every customer who signs came from. Meta lead, Meta lead gen, word of mouth, whatever. The idea is to manually reconstruct who bought what and where they came from, until the automatic tracking is clean.

It's manual reporting, not glamorous, but an Excel column that gives you 80% visibility instead of nothing is worth it. If we can do 80%, why not do it right now while we wait for the 100%.

Another thing on the ads side, stop sending all your traffic to the same page. Create specific pages based on what you sell and on the campaign. It helps tracking, and it helps conversion. That's the logic we apply in the conversion and Shopify side of things.

The contrarian reflex to fix

The "no need for a CRM" can be true on tooling and false on steering. The two providers who said "Excel is enough" were right for following up with prospects. They were wrong the second you talk about budget arbitrage.

Because without source tracking, you can't know which channel actually pays off. You throw budget at Meta, Google, a bit of awareness, with no idea what's driving sales. You steer on gut feel. And on an account burning 8,000 bucks a month, that adds up fast.

And we always land back on the same logic, you build your budget on what actually pays off, not on what the platform displays. I talk about it in build your ad budget from your margin.

Where to start

If you're in this situation, a business where the sale happens offline after the click, here's the order I'd do things in.

1. Put dynamic UTMs on all your campaigns before any new launch (campaign name, campaign id, ad id). It's not retroactive, so do it now.
2. Check whether your CRM can pull the UTM when the lead comes in. If it can, wire it up. If it can't, you move to manual reporting.
3. Add a "source" column in your Excel or your Trello, and require the reps to fill it in on every signature. The data comes from the reps filling in that column. No tool will make up for it if they forget.
4. Calculate your cost per sale and your conversion rate per channel, not per form. That's the only number that tells you where to put more budget.
5. If you want a clean backup measure, run a post-purchase survey asking "how did you hear about us." It sidesteps the whole attribution problem, and people often answer Instagram or "a friend told me about it," things Meta will never show you.

Forget the perfect dashboard. What you want is to know which Meta campaign brings you paying customers, and at what cost. The day you have that, even in Excel form, you stop flying blind.