Measuring TV and podcast ads without pixels

Digital marketers inherited a strange belief: that a channel is only measurable if it leaves a click trail. So offline channels get measured with duct tape - promo codes, vanity URLs, "how did you hear about us?" surveys. All of them undercount, some dramatically.

Why the duct tape undercounts

Promo codes capture only the listeners who remember the code at checkout - and coupon sites leak codes to people who never heard the ad, polluting the signal in both directions. Post-purchase surveys get single-digit response rates skewed toward your chattiest customers. Vanity URLs measure the tiny minority who type URLs at all.

Meanwhile the actual effect of a good TV campaign shows up everywhere except those trackers: branded search volume, direct traffic, retail velocity, and - awkwardly for attribution - inside the "performance" of your paid search account.

Aggregate measurement is the native fit

MMM works from weekly aggregates: spend (or GRPs, or impressions) per channel against total sales. That makes offline channels first-class citizens:

• TV: air weight varies week to week, which gives the model exactly the variation it needs to isolate the effect, including the 2-6 week adstock tail.
• Podcasts: host-read spots cluster by show and flight dates. Feed the schedule in and the model reads the response, listeners who never touched a code included.
• Out-of-home and radio: regional flighting is a natural geo experiment, and geo-split models exploit it directly.

No cookies is a feature, not a workaround

None of this touches user-level data. No pixels, no fingerprinting, no consent-banner dependencies for measurement. The same property that makes MMM privacy-proof also makes it immune to iOS updates and browser tracking prevention - measurement that does not degrade every time a platform changes its privacy policy.

If offline channels are in your plan but not in your measurement, the model is the missing piece - not another promo code.