Pixel attribution

Who provides it

Dedicated vendors (Podscribe is the name you will hear most; its methodology docs are the ones to read) sell cross-platform attribution as a service; per-seat and per-spend pricing is quoted, not published, so no figure ships here. Spotify includes its own pixel free for campaigns on its platform, which is the cheapest way to try the method at all. The platform-owned versions only see their own inventory; the point of a third-party pixel is one methodology across every show and platform you buy.

The hard limits, stated

IP matching resolves to households and offices, not people: the “converter” may be the listener’s roommate. Carrier-grade NAT on mobile networks pools thousands of users behind shared IPs, forcing modeled discounts. Privacy platform changes keep removing signals, so a vendor’s match rate two years ago says little about today’s. And attribution windows are choices: a 30-day window credits the podcast with conversions a 7-day window would not. None of this makes pixels worthless; it makes them modeled evidence, to be reported as such.

When it beats codes, and how to sanity-check lift

The crossover is scale: codes stay readable at any size, pixels need enough impressions for matched-visitor counts to clear noise, which in practice means the $15k tierand above. Sanity checks to run on any vendor lift report: ask for the match methodology and window in writing, compare pixel-attributed conversions against code redemptions for the same flight (they should bracket each other), and check whether “lift” is against a true holdout or a modeled baseline. A vendor who resists those questions is answering them.