Five podcast advertising measurement mistakes and a scoring rubric

What to take away

A media buyer in Manchester reviewing a host-read renewal sees a download figure, a promo code count and a brand lift line. Only one of those usually survives scrutiny, so the checks below separate evidence from decoration.
Podcast advertising measurement mistakes cover five issues: unverified audience data, drifting attribution windows, vanity reach, claims that outrun the rules and reports that cannot be audited later.
Apply these criteria to any UK podcast buy, whether a direct host-read or a network placement. Buyers in Scotland, Wales and Northern Ireland should expect the same standard of evidence.
The scoring rubric below turns the checks into a number you can put to a finance team before committing budget.
If a seller cannot answer a check with a named source and a date, treat the metric as unproven.

Mistake one: unverified audience numbers

The first failure is accepting a reach figure with no stated method. The IAB Tech Lab's Podcast Measurement Guidelines provide common definitions and methods for podcast measurement. For UK audience context, RAJAR's MIDAS research covers listening to podcasts and other audio; it is audience research, not an audit of an individual seller's campaign logs. Audience and media research methodologies are published openly by firms such as Ipsos. A supplier should be able to point to an equivalent method note rather than a slide.

Audience number checks

Ask how the count was produced
Ask who collected the data
Ask when it was collected
Request the method note
Separate downloads from listeners

A download is a delivery event, not a listener. If the seller conflates the two, the plan is built on sand. The IAB Tech Lab's Podcast Measurement Guidelines set out common measurement definitions; put the definitions you will use in the contract before money moves. The wider podcast advertising measurement and reporting guide also sets out which definitions belong in a contract.

Mistake two: attribution windows that drift

Attribution errors appear when the reporting window changes between campaigns. One month the code redemption window is 30 days, the next it is 90, and the comparison is meaningless. Fix the window in writing at booking.

Attribution window drift

Campaign A

Code window: 30 days
Comparison: Meaningless
Fix: Fix in writing

Campaign B

Code window: 90 days
Comparison: Meaningless
Fix: Fix in writing

Promo codes also capture only the listeners who act immediately. For example, a show charging £2,500 per episode may report 40 code redemptions, which says little about listeners who searched the brand later. Pair codes with a second, independent signal.

Mistake three: vanity reach over response

Reach without response is the most common padding in a sponsorship deck. A million impressions across a network tells you nothing about whether the audience matched the brief. Ask for the response measure that sits alongside it.

If the seller cannot supply one, build a dashboard that tracks the measures you agreed. A working example is set out in the podcast advertising reporting dashboard, which shows how to keep delivery and response side by side.

Mistake four: claims that outrun the rules

Advertising claims in audio must still be capable of substantiation. For non-broadcast podcast ads, the ASA administers the CAP Code: Section 3, Misleading advertising, includes rule 3.7, which requires documentary evidence for claims that can be objectively substantiated. For broadcast ads, check the BCAP Code. Published ASA rulings show how complaints about audio and podcast ads have been adjudicated.

The Competition and Markets Authority (CMA) also enforces consumer-protection law. Part 4, Chapter 1 of the Digital Markets, Competition and Consumers Act 2024 sets out the UK consumer-protection regime for unfair commercial practices. This is distinct from the ASA's administration of advertising codes, so check both where applicable.

Mistake five: reports nobody can audit

If the only version of a report lives in a slide deck, it cannot be checked later. Keep raw delivery data, the method note and the agreed definitions together, with dates. Where publishers use Spotify for Podcasters or Apple Podcasts Connect, retain the original platform export and label it as platform-reported data, not independent verification.

A report that cannot be reproduced six months on is marketing, not measurement. That matters most at renewal.

Scoring rubric for a sponsorship report

Use this to score the evidence you are given, not the seller. One point per row, then total.

Check	Pass condition	Score
Audience method	Named method and collection date supplied	0 or 1
Delivery versus listening	Download and listener defined separately	0 or 1
Attribution window	Fixed in the booking terms	0 or 1
Second response signal	Independent of promo codes	0 or 1
Claim substantiation	Script claims match held evidence	0 or 1
Audit trail	Raw data retained and reproducible	0 or 1

Five or six is a report you can renew against. Three or four means fix the gaps first. Two or below means the numbers are not yet evidence.

Common questions

Does this list apply outside England?

Yes. The checks are commercial, not legal, so buyers in Scotland, Wales and Northern Ireland can apply the same rubric. Where a rule differs by nation, check the specific regulator guidance.

How many episodes should I measure before judging?

There is no fixed number. Judge on whether the report meets the rubric, and treat a single episode as too thin a base for a renewal decision.

Can I rely on the platform's own analytics?

Platform analytics are useful delivery data, but they are not independent. Pair them with a second signal and with the method note behind any audience claim.