Understand Performance FAQ & Troubleshooting

Answers to the most common questions you may have when trying to understand performance in Spotify Ad Analytics.

Why doesn't the conversion rate (CR) in reports match the CR in the Ad Analytics dashboard? Spotify Ad Analytics calculates the conversion rate (CR) by dividing Visits by Household Reach. Household Reach numbers are not additive, as there will be overlap between line items, therefore, adding up the Household Reach across all line items would cause a discrepancy. Note: The Spotify Ad Analytics dashboard is defaulted to modeled attribution. Therefore, a report may present unmodeled attribution (depending on the report type), which would provide a different CR.
Why do my conversions have decimal places?
Spotify Ad Analytics has two reasons that the data may result in decimal places in your performance metric.
1. First, Ad Analytics uses a modeled approach to attribution. Modeling is a way for us to extrapolate the data of listeners who were exposed to the ad, but we were not able to confidently match the IP address to a household (i.e. noisy IP data). Our reporting is very conservative as we only perform attribution on IP addresses where we know we can get good data.
2. Secondly, if we see the same household exposed to multiple episodes of a podcast and then convert on the website, we will give partial and equal credit to those episodes/shows that drove a conversion.
  1. For example, if a user heard 2 ad impressions from 2 different shows and visited the site once (1 pixel fire) - each show would get 0.5 visits.

What is the "Other" data in the 'Players' and 'Locations' charts? Players and Locations data populate based on the impression data, which can vary per publisher.
- Campaign Attribution: Ad Analytics categorizes the player (software or app) and the location where the impression occurred. If the impression cannot be categorized, it may be counted in multiple categories, or be categorized as "other".
- Campaign Delivery Reporting: In a delivery reporting campaign, it is not unusual to see your chart data look incomplete or inflated under the 'Player' and 'Location' tab. Depending on the ad server, we can receive duplicated data (duplicated IPs with the same timestamp) from the tracking URLs. The attribution job de-dupes the duplicates -- because, this is a reporting campaign, no attribution takes place to dedupe those duplicates, which may result in inflated player and location charts.

Why is Conversion Lift only available in certain markets?
Spotify Ad Analytics uses a device graph to measure confirmed household IP addresses. Due to local regulations, this device graph can only be used in Australia, Canada, Japan, Mexico, and the United States.
I requested a report, but I haven't received it yet. What do I do?
If your attribution campaigns meets Conversion Lift requirements and you have not received your requested report within 5 business days, reach out to support-adanalytics@spotify.com for assistance.
What if my report is unable to detect lift?

Potential reasons can include:
- Your creative didn't effectively communicate the intended message, brand or product.
- Your ads may not have been running for a long enough time, or doesn’t have enough reported conversions to calculate meaningful results.

What is the difference between the 'line_item' field, and 'podcast_title' field in reports? 'Line_item' is the name of the DAI (dynamic) ad placement. It's expected for the 'line_item' field to be blank when the line item is embedded, or when the podcast title was not passed to Ad Analytics. For example, many publishers do not pass podcast-level data on Run of Network or Programmatic line items.
Why are there duplicates in the Campaign Conversions and/or Organization Campaign Conversions report? These reports list all activity related to a user and includes details such as timestamp of the impression and timestamp of the conversion. Sometimes dynamic line items appear to have multiple ad insertions with the exact same timestamp. In those cases, we capture all impressions from dynamically inserted ads, even if they appear to be duplicates. For embedded ads, we count downloads using the IAB v2.0 standards. We assign equal and partial attribution to all of the actions that a user performs, so you would need to use the attribution column to sum the number of conversions.
Which reports are modeled and which ones are unmodeled? All reports reflect modeled data, except the Campaign Conversions and Organization Campaign Conversions reports. The conversion reports will always be unmodeled.

Still have a question? Submit a ticket.