Does the pixel track my listeners?

Not directly. The pixel lives on the advertiser's site and only sees people who visit it. The listener side of the match comes from your ad server's delivery logs, which already record the IP and user agent of every download request. The pixel supplies the other half of the comparison.

Why does pixel attribution report fewer conversions than my promo code?

They fail in opposite directions. Pixels miss conversions when the listener downloads on one network and buys on another, which is common. Promo codes miss conversions when someone hears the ad, forgets the code and buys anyway. Most advertisers run both and treat the pixel number as a floor.

Is IP-based podcast attribution going away?

It is getting harder rather than disappearing. Private relay services, carrier-grade NAT and IPv6 rotation all weaken the match. The direction of travel is that pixel attribution keeps working as a directional signal while becoming less reliable as an exact count.

What is a podcast attribution pixel?

A podcast attribution pixel is a small tracking tag installed on an advertiser's website that records the IP address and user agent of each visitor, so an attribution vendor can match those visitors back to the devices that downloaded episodes carrying the ad. The pixel never touches your podcast. It sits entirely on the advertiser's side.

Which means the advertiser controls the measurement. You supply the delivery logs and see the report at the end.

How the match actually works

Podcast advertising has no click. Nobody taps an audio ad, so the industry built a way to connect two separate events that happen to share a network address.

The download. A listener's app requests the episode. The ad server logs the request, including the IP address, the user agent and which ad was stitched in. Those are the same two fields the IAB Podcast Measurement Technical Guidelines v2.2 name as factors for identifying a unique download.
The visit. Some time later, someone loads the advertiser's site. The pixel fires and records that visitor's IP address and user agent.
The comparison. The vendor looks for IP addresses that appear on both lists, inside a configured lookback window.
The attribution. A match is reported as a podcast-driven visit, and a match that reaches a purchase page is reported as a conversion.

The lookback window is a campaign setting rather than a standard. A short window attributes only immediate response. A long one catches slow consideration but raises the odds of crediting a visit the ad had nothing to do with.

This is household-level probabilistic matching, not identity. It says a device on this network heard the ad and a device on this network visited the site. It does not prove they were the same person, and the vendors are generally straight about that.

What it catches and what it misses

Scenario	Does the pixel catch it?
Listens at home on wifi, buys on the same laptop	Yes
Listens on wifi, buys on a phone on the same wifi	Usually, same household IP
Downloads on mobile data, buys at home	No, different networks
Downloads at home, buys at the office	No
Uses a private relay or VPN	No, the address is obscured
Behind carrier-grade NAT with thousands of others	Unreliable, the address is shared
Hears the ad, buys in a physical store	No

The pattern is consistent: the pixel undercounts. Every failure mode listed above removes a real conversion from the report, and almost none of them invent one. That is why experienced buyers read a pixel number as a floor rather than a total, and why a campaign that looks marginal on pixel data is not automatically a campaign that failed.

The carrier-grade NAT case deserves a flag. When an internet provider puts many subscribers behind one shared address, a match on that address means very little. The IETF's requirements for carrier-grade NATs treat that address sharing as the normal way these networks operate, not an edge case. Good vendors suppress or discount those matches instead of claiming them, so a campaign that skews toward mobile-network listeners loses credit it earned.

Why it matters for your show

If you sell sponsorship, the pixel report is increasingly the thing your renewal depends on. Three practical implications.

Ask which window was used before the campaign starts, not when the report arrives. A seven-day window and a thirty-day window on the same campaign produce different stories, and the choice was the buyer's.

Expect mobile-heavy shows to underperform on pixel attribution. A commuting audience downloads on cellular and converts somewhere else entirely, so the mechanism itself works against you. That is worth saying out loud in a renewal conversation rather than letting the number speak unchallenged.

Bring a second signal. Promo codes and vanity URLs fail differently from pixels, so agreement between two weak methods is a stronger argument than one method alone. Our ad attribution examples walk through what that looks like on a real campaign.

Setting expectations with an advertiser

The conversation that goes badly is the one where nobody defined success first.

Agree the attribution method, the window and the conversion event before the insertion order is signed. Those terms are negotiable in advance and effectively fixed once delivery begins.

Agree what a good result looks like in the advertiser's own terms. A vendor's benchmark for a different category on different inventory is not a target your show agreed to.

Keep your own delivery record. When an attribution report comes back lower than expected, the first question is whether the ad actually delivered to the audience you sold, and that is a downloads question before it is an attribution question. How OP3 and Podsights differ covers why those two jobs need separate tools.

Related terms

Worth reading next to this: ad attribution for the full set of methods, and measuring podcast ROI for what to do with the numbers once you have them.

For the delivery side of that comparison, Podder's prefix analytics counts downloads independently of whichever attribution vendor your advertiser chose.