Announcing: Custom Attribution Windows
by Dave Zohrob · March 01, 2021

Chartable is pleased to announce configurable attribution windows across all of our attribution products today—SmartLinks, SmartPromos and SmartAds.

An "attribution window" describes the time allowed between an event, like hearing an ad, and a conversion, like visiting the advertiser's website. If the conversion happens within the attribution window, we give "credit" to the event.

Attribution windows are crucial to understanding when a conversion happens. There's often a delay between hearing or seeing an advertisement and taking action on it—for example, hearing an ad for a product on your commute to work and actually remembering to visit the advertiser's website when you get home in the evening.

Changing the window involves tradeoffs. The longer the attribution window, the more likely we'll count conversions which were not directly related to the event. Likewise, as the attribution window is shortened, we link fewer conversions to the original event, even those that may have been influenced by it.

We've been offering a fixed attribution window for each of our products, designed to reduce noise without sacrificing accuracy. However, different ads and advertisers may require fine-tuning, and now that tuning is entirely in your hands.

SmartLinks

For SmartLinks, the event is a click on a podcast link, and the conversion is a download of a podcast episode.

We set the default attribution window to 1 hour. As you can see below, 45% of the downloads occur within the first hour. And since podcast apps do not send cookies, we can only match on IP address when counting a conversion — hence, a short window ensures maximal accuracy, especially for uses where you expect a listener to have high download intent (e.g. on your podcast's website). For other kinds of links, like in ad campaigns, a longer window may more accurately reflect the downloads driven by an ad.

SmartPromos

SmartPromos measures podcast-to-podcast ad campaigns. In this case, the event is an ad for a podcast heard on another show, and the conversion is a download of that podcast. The default attribution window is 7 days.

Again, because we're operating in a cookieless environment, 7 days is a conservative default, which captures 45% of the downloads for the promo. You'll notice peaks every seven days, which happen because most podcasts release weekly and a new episode may cause listeners to listen to previous episodes as well.

A mass market podcast may want to decrease their attribution window when running ad campaigns, to help isolate the effects of their promotional efforts.

SmartAds

For SmartAds, the event is the download of an ad for an advertiser, and the conversion is defined as a particular action on that advertiser's properties — for example, installing an app, visiting their website, or buying a mattress.

The default attribution window is 30 days, which captures 73% of measured conversions. A niche advertiser with a long sales cycle may need to adjust their attribution window upwards to see a full picture, and can likely safely do so without sacrificing accuracy.

You'll notice a different shape for podcast-to-web conversions vs. podcast-to-podcast — it takes longer for listeners to convert. This makes intuitive sense because in the podcast-to-podcast case, a listener can subscribe to a new show after hearing an ad without leaving their podcast app, while visiting an advertiser website may involve switching apps (e.g. Apple Podcasts to Safari) and potentially even devices (e.g. I listen to podcasts on my phone, but do my shopping on my laptop).

We've been wanting to offer configurable windows for a long time now, so we're really excited to finally launch this feature to all Chartable users.

If you have any questions, feel free to reach out to your account manager or email help@chartable.com.

Liked this post? Get our newsletter with podcast insights.
More from the blog

We're launching podcast industry first:...

Podcasts through the pandemic: Trends from 2020...

Chartable, the podcast analytics and...