Update to Adobe Marketo's Bot Filtering Algorithm
Currently, Marketo has two email bot identification methods in production: "Match with IAB List" and "Match with Proximity Pattern." In the next few weeks, Marketo is going to update its "Match with Proximity Pattern" identification method's algorithm to improve its precision in identifying bot activities.
To give an idea, the Proximity Pattern algorithm identifies email activities that happen and originate at precisely the same time or are very close to one another (within the time span of a second) as bot activities.
When comparing two activities, Marketo considers the following attributes:
✔Lead ID
✔User-Agent
✔IP Address
✔Email asset
✔Email Open or Click Link Activities
✔Time between 2 activities
After the update, out of the above items, Marketo will not consider the "User-Agent" and "IP Address" attributes for its Proximity Pattern algorithm.
Of course, Marketo is rolling out this update after cautious consideration and analysis of several real-life data points across its customers.
Since a couple of attributes will no longer be considered for the "Proximity Pattern" algorithm, customers who have it turned on may see an uptick in the number of bot activities and a reduction in the number of non-bot email activities in their subscriptions. Marketo recommends turning on the bot activity logging feature for a while and observing the bot activities identified by the algorithm.