Strategy Unfiltered.

The Attribution Problem Your Board Doesn’t Understand

You don’t own your attribution model. You rent it. And the landlords keep raising the rent.

---

There’s a moment in every board meeting where the marketing slide comes up and the room gets quiet. Not quiet-good. Quiet like a teacher waiting for a student to explain why the homework isn’t done.

Then someone asks the question: “So how do we actually know this is working?”

And the CMO answers with a number. Cost per lead. ROAS. Pipeline influenced. Whatever the dashboard surfaces that week.

Here’s the part nobody wants to say out loud: that number came from a system you do not control, built by a company whose business model depends on you believing it, and reconciled against a CRM where 30% of the records have a UTM string that died two product launches ago.

You don’t own your attribution model. You rent it — from Google, from Meta, from GA4, from your agency, from your call-tracking vendor.

Every one of those landlords has an incentive that does not perfectly align with yours. Google wants you to spend more on Google. Meta wants you to spend more on Meta. Your agency wants the retainer renewed. Your call-tracking vendor wants the contract renewed. None of them are lying. They are simply optimizing for what they can see.

The board doesn’t understand this. Most of the time, the CMO doesn’t fully understand it either. So the number stays on the slide. The room moves on. And six months later, when growth stalls, everyone is surprised.

In a recent audit of seven PE-backed healthcare platforms, the average attribution variance between platform-reported conversions (Google Ads, Meta) and CRM-confirmed patient acquisitions was 41%.

Not 4.1%. Forty-one. That’s the gap between what the platforms told the operators they were buying and what the platforms actually delivered. And in every case, the platform number was higher.

If you operate a platform and you can’t answer these five questions in under 60 seconds, your board doesn’t actually know what your marketing is doing:

1. 01What percentage of your reported conversions are de-duplicated against the CRM, weekly?
2. 02What is your identity match rate — the share of ad-platform users you can tie to a real patient record?
3. 03What is your consent rate, and what is the modeled-vs-observed split inside GA4?
4. 04Which of your tools sit under a signed BAA, and which are quietly out of scope for PHI?
5. 05If your largest ad platform disappeared on Monday, how would you know your CAC by Friday?

These aren’t analytics questions. They’re governance questions. And they belong at the top of the dashboard, not in a footer nobody reads.

The operators getting this right in 2026 share one trait: they treat attribution like a financial control, not a marketing report. They reconcile weekly. They publish their methodology. They show the board the variance — not the platform headline.

It is not glamorous work. It is the work that survives a downturn.