ScorePool

ScorePool is a free, optional system that scores each of your subscribed users to predict how likely they are to engage via email in the next 30 days, based on their activity across all ActionKit clients. You can use ScorePool as one tool to target users that are most likely to engage, or exclude users that are unlikely to engage.

You have to opt in to benefit from ScorePool. There is no cost to join ScorePool, but you do need to agree to permit the use of your activity data. Superusers can do this for your organization by clicking on ScorePool Signup in the configuration menu.

https://s3.amazonaws.com/clientcon/images/editor-2022-01-28.png

How It Works

ScorePool aggregates engagement metrics for every end user across all ActionKit clients into a shared data pool. The data used by ScorePool includes opens, clicks, actions, and more, but it does not include personal information: no plain-text emails, no names, or addresses. The detailed activity data in the pool is kept private: only scores and statistics are shared.

Then, using a machine intelligence algorithm, ScorePool will assign a score that represents the probability that the user will engage over email in the next 30 days. The scores are capped at 50, so a score of 50 means that this user has a 50% likelihood or higher of engaging with email in the next 30 days.

In addition, ScorePool calculates how many days have passed since this user has engaged with any ActionKit client program. Days since last engagement is capped at 90 days, so a user with a last engaged score of 90 has engaged with an ActionKit clients' program within the past 90 days.

Using The Data

You can use ScorePool scores in your mailing targeting and re-engagement queries. Scores can also be useful in dealing with deliverability problems, providing the ability to target a better set of users when an ISP has blocked or delayed your messages.

ScorePool will add data to two tables in your database:

scorepool_userscore

Predicted engagement scores for each subscribed user.

Field Type Description Null Key Default Extra
user_id int(11) joins on core_user.id. NO PRI NULL  
score smallint(5) Score from 0-50, indicating likelihood to engage in the next 30 days. Scores over 50 are capped at 50. NO YES NULL  

scorepool_lastengaged

Days since last engagement across all ScorePool participants for subscribed users.

Field Type Description Null Key Default Extra
user_id int(11) joins on core_user.id. NO PRI NULL  
days smallint(5) Days since last engagement, from 90-65535. Values under 90 are capped at 90. NO YES NULL  

NOTE As noted above, values are capped to try to prevent unintended uses. Scores above 50 will be capped at 50. Last-engaged day counts below 90 will be capped at 90. These thresholds may change in the future.

The data in these tables will be refreshed periodically, initially once a week. New users added to your database will initially have no rows in these tables until the next time the tables are rebuilt.

Because ScorePool data isn't allowed to be synced outside ActionKit, ScorePool data is not copied to dedicated client database replicas. ScorePool data is available on the share ("client-db") database replica.

Query Builder Support

To access ScorePool data in the Query Builder, start with a Users query and then you can add filters for scores and last-engaged day counts:

https://s3.amazonaws.com/clientcon/images/Screenshot_2022-03-10_14-21-11.png

Mailer Targeting Support

You can target mailings using ScorePool scores without needing to use a report. The following targeting criteria is available on the mailing targeting screen:

https://s3.amazonaws.com/clientcon/images/Screenshot_2022-03-10_14-24-11.png

Aside from the minimum and maximum options, a third targeting filter "ScorePool Unscored Users" controls whether your set will include users who have not yet been scored. This is useful because new users added to your database won't have scores until the next time the ScorePool tables are rebuilt.