What is a conversion value?

Simply put, in the world of iOS and SKAdNetwork (SKAN) measurement, a conversion value represents an action or a series of actions taken by a user in an app. It’s how app developers and marketers gain critical insights into campaign performance for non-opted-in users on iOS. Each app tracks and measures different in-app events according to their app’s vertical/category, business needs, monetization model, and KPIs. To do this, Apple’s privacy-centric SKAN lets app developers configure up to 64 conversion values (0-63), which can be mapped directly to 64 individual conversion values, or via six bits that map to six individual user actions.

These actions can represent in-app events, chains of events, or characteristics that an app developer determines as an indication of a valuable user. It’s up to the developer to determine how they map their conversion values, which combination of events makes sense, and what they want to measure.

Apple’s SKAN 4 introduced coarse conversion values, which can be mapped to three values—low, medium, and high.

Let’s start with the 0-63 conversion values, or what Apple now calls fine conversion values. In the binary numeral system, there are six bits of downstream metrics between 000000 and 111111. These 6-bit characters can be converted to the numbers 0 through 63, which is what the 64 conversion values are based on. Here’s what binary numerals look like compared to decimal:

SKAN was created to increase user privacy while still giving advertisers a way to attribute installs and events. To further protect and obfuscate the individual device ID, Apple introduced a random timer that is triggered following a registered conversion event. This timer dictates the length of time before a device sends a postback, containing the conversion data, to:

• The ad network
• The advertiser
• The mobile measurement partner (MMP) – if they are set up as an additional endpoint.

Since the postback is not received immediately, a single user cannot be linked to the data, guaranteeing anonymity.

With SKAN 3, events or conversions are measured for 24 hours post-install. If a mapped event is triggered before the 24 hours is complete, the conversion value it corresponds to is logged and the 24-hour window resets. If another one of the events is triggered, the conversion value gets updated, meaning the previous one is no longer recorded. Due to the industry standard largely set by Facebook, many partners and most SDKs will limit the window to 24 hours total, meaning the potential to extend the measurement or conversion window is essentially mitigated.

Once the measurement window is complete and the conversion value is finalized, the random postback timer is triggered. SKAN then randomly returns the postback, including the conversion value, to the network or advertiser at any point between 0-24 hours after the finalization. With the introduction of SKAN 4, there are now two additional measurement windows with two corresponding, additional postbacks.

While SKAN 3 has one measurement window and one postback, SKAN 4 has three measurement windows and three postbacks. Together, they make up what Apple refers to as multiple conversions. Only the first postback can contain the 64 conversion values as explored above—these are referred to as fine values in the SKAN 4 framework. Postback 1 can also contain coarse values, and these will be received instead of fine values if a lower privacy threshold is met. In Postback 2 and Postback 3, coarse values are the best case scenario for the highest crowd anonymity. These can be mapped to the three values of low, medium, and high.

The three SKAN 4 postbacks at a glance:

First postback (Postback 1)

Measurement window: Now 0-2 days instead of (generally) 24 hours

Postback timer: Now up to 48 hours instead of up to 24 hours

Potential conversion values: Fine values or coarse values

Second postback (Postback 2)

Measurement window: 3-7 days

Postback timer: Up to 144 hours (6 days)

Potential conversion values: Coarse values

Third postback (Postback 3)

Measurement window: 8-35 days

Postback timer: Up to 144 hours (6 days)

Potential conversion values: Coarse values

Adjust’s Conversion Hub, part of our overall iOS Solutions, offers clients everything they need to confidently measure their iOS campaigns for the entire dataset with SKAN. This can then be complemented by opted-in, device-level data for side-by-side analysis and the most granular possible insights.

Depending on a client’s needs, we can work within one of two modes to get the best possible conversion values for your campaign goals:

63 CVs mode

Smart setup flow: Best for revenue mapping

This method focuses on giving clients granular insight into their monetization efforts for their iOS 14.5+ user base. Make use of all 64 possible conversion values by mapping revenue ranges to each value. Low revenue ranges are assigned to low conversion values from CV0 up, and the highest revenue range is assigned to the highest conversion value, CV63.

Events can also be measured with this mode, but they must be ordered according to the user journey so that visibility is maintained when the conversion value updates.

6-bit mode

Advanced setup flow: Best for behavior mapping

Maximize the transparency of user engagements post-install for complete user journey optimization. Up to six events can be chosen and mapped randomly to the binary characters’ six bits. Unlike with events in the 63 CVs mode, the order of your events won’t impact the conversion value you receive because the bits will directly represent the events.

These solutions allow clients to gather critical data, allowing them to optimize campaign efforts and deliver maximum return on ad spend (ROAS) on iOS.

Regardless of the mode you choose, Adjust will help you and your team gain a comprehensive understanding of SKAN 4, conversion values, postbacks, and everything in between to make sure you are getting the most out of your iOS campaign insights.

SKAN 4 and coarse values

Conversion Hub is now ready and set up for SKAN 4 conversion value mapping, as well as lockWindow functionality. You can map to 64 conversion values or 6 bits for Postback 1 and to the three coarse values for Postbacks 1-3. We currently recommend porting your existing SKAN 3 strategy over for 1-1 use with Postback 1’s fine values, and developing a coarse value strategy based on that configuration.

For example, if you’re using the 63 CVs mode for SKAN 3 currently, you can copy this over for SKAN 4 fine mapping, then split the values into two buckets (1-31 and 32-63). These two buckets can be assigned to the medium and high values, with the low value being reserved for a session.