Removing/Zero out behavior scoring past a certain time period

[dmedua]

Hi everyone!

I'm looking to remove or zero out any behavior scoring older than x number of days. Does anyone know how I may go about this or if it is even possible?

This would simply because I only care about the most recent activity a lead has been doing, as opposed to the form they filled out 11 months ago.

Ex: Let's say our AQL person score threshold is 50. If my lead had a person score of 70, but 50 of their points were from years ago, then those 50 points aren't relevant anymore and that lead will likely not truly be qualified and end up being recycled. So I'd want those 50 old points removed, but I don't want to also lose the 20 points that they did accumulate in the past month or so.

Is there a way to automate having points/activities older than a certain age zeroed out or removed?

Thank you!

[SanfordWhiteman]

Blunt/built-in answer: Using Marketo built-in functionality, there's no way to do this. The date and/or quantity of the last increment/decrement isn't stored within a Score field, it's just an integer representing the current value.

Longer/advanced answer: By accompanying your flows with a webhook-compatible service, this is possible. You store the historical decrements/increments/timestamps in a Textarea field, and also (because this reduces webhook calls) set a Score Last Modified DateTime field to {{system.datetime}} whenever you touch the field.  Then each day, run a recurring batch for people whose Score Last Modified touch was exactly 90 days ago, or whatever you want your decay checkpoint(s) to be.  The batch calls Request Campaign to Call Webhook. The webhook subtracts the offending points and returns the new value.  We have stuff like this running for a couple clients; it takes planning for sure.

[Katja_Keesom · Accepted solution]

As @SanfordWhiteman said, subtracting a specific value that was previously added is not technically possible. However, it is good practice to implement an "inactivity" score in any lead scoring model. This subtracts a specified amount of points from both the behaviour score and person score if no relevant activity was seen in the last x days.

An example for a situation where MQL threshold is 100: If no email click, web visit, form fill or event registration / attendance in the last 30 days, subtract 10 points from the total.

Hope this helps.