How to compare Audience Studio segment populations in Activation Partner systems

Background

Audience Studio allows you to send audience data to many external systems for targeting, suppression, personalization, insights and more. This is what we call Activation. In order to understand how data is handled, we’ll need to provide you with a little background.

When an Activation Partner is integrated within the Audience Studio platform, they either utilize our standard server-to-server integration or a custom integration. With standard integration, there are many supported configuration options that determine how the data is delivered. Even when a partner uses the standard integration, they can apply additional rules to determine how the data is handled once they retrieve it from the location where Audience Studio delivers it. When the partner uses a custom integration, much of the logic the partner applies to the data is specific to the partner’s requirements.

User Matching

In order to integrate with partner’s systems, we map the partner’s device identifier to our device identifier. This is known as User Matching. A partner may choose to receive the Audience Studio ID or their PartnerID when receiving data.

Example

Audience Studio ID PartnerID Partner
123 543 Google
123 7645 Twitter
123 65311 TheTradeDesk
234 334 Google

 

What Data Gets Sent

When sending audience data to a partner, there are two processes:

  1. The initial segment gets created in the partner system. For standard integrations, this is what we call Segment Sync. Each customer may choose to send only segment ids or also send the segment name when configuring a connection between Audience Studio and an Activation Partner. For custom integrations, we often use the partner’s Taxonomy API.
  2. The user segment membership. This is the list of segment ids along with the device ids that belong in the segment. The partner may choose to receive their segment id or our segment id and they may also choose to receive their device id (from user matching) or our device id.

When a partner receives THEIR device ID, the integration becomes dependent on user matching. However, user matching is not required for anything other than cookies. A mobile ID is consistent across systems as is an OTT device, hashed email address, first party ID, etc.

If a partner supports mobile devices, Audience Studio will not limit them to only the mobile devices that were included in user matching.

Using the user match example above, if we have three mobile devices to send to Google in SegmentA, the we would send all three instead of sending only the two that appear in the user match table.

Partners may choose to receive a complete data replacement each day or to receive incremental data each day.

If a complete data replacement is sent to the partner each day, you’ll be able to reference the total population sent in Activation > Partner Management.

blobid0.png

If incremental data is sent to the partner each day, you may reference a cumulative population in Activation > Partner Management. With the incremental option, you may also see added (new devices that qualified for the segment) and deleted (devices that expired from the segment) numbers for each day.

blobid1.png

Comparing Apples to Apples

Match rates can be used to identify expected reach, but match rates can be very misleading and are becoming less relevant when dealing with device IDs other than cookies.

Additionally, each partner supports a different user match recency and / or an expiry timeframe, often referred to as ttl. Audience Studio sends device data for the last 90 days of user matching as well as the more persistent ids, such as OTT, hashed email addresses (HEMs), etc. However, partners may only accept device data matched within the last 30 or 60 days. The partner will expire devices out of the segment based on their expiry timeframe (e.g. 30 days). There are also some partners that accept all devices we deliver. This is rare, but it does occur.

Some Activation Partner systems also only include audience counts for devices that are in user match. In the example above, we stated that we’d send all three mobile devices to Google, but Google may only accept one of those mobile devices if the recency was not within the last 30 days for the other two devices.

In order to determine which audience population you will see in a partner’s system, you will need to compare the population sent or cumulative population in Activation> Partner Management to the audience population in the partner’s system.

Additional information you’ll need from the partner to compare the numbers:

  • What is the partner’s supported user match recency? If it’s less than 90 days, then you’ll expect to see a smaller population in the partner’s system
  • What is the partner’s supported ttl? You should expect to see some of the population drop off over time based on the expiry timeframe. If Audience Studio is sending incremental data and the partner expires data after 30 days, then devices will start expiring after 30 days (unless a periodic full refresh is sent)
  • Does the partner include all device types received, only specific device types (e.g. cookies) or are they limiting the data to devices included in user match even though consistent device ids, such as mobile, shouldn’t require user matching?
  • Do they apply any other rules to how the data is handled or counted?

Considering All The Data

There are some competitors that may claim 100% match rate, especially for publishers. You can also achieve 100% match rate with Activation Partners through Audience Studio if you limit the data sent to only the first party data where partner user match is available 100% of the time for only the recency supported by the partner. However, since you are able to use both online and offline data, media data, third party data, etc., you will be able to reach more customers and prospects than limiting to only 1st party contextual use cases, but the match rate will never be 100%. It will be the subset of devices where the Activation Partner has a match and / or is able to use a more persistent device id, such as mobile, OTT, hashed email, etc.

Again, match rate becomes less relevant and total addressable population is what we should be focused on optimizing.

Have more questions? Submit a request

0 Comments

Please sign in to leave a comment.