I need to fix my demographic lead scoring - is this the right approach?

[Anonymous]

There are a couple things I need to fix:
- We had generic email at -5 but it wasn't set to flow once, but every time so some score are -1000 plus bc it ran every day!
- I want to increase scores for certain target accounts and titles. I found that these were buried amid leads with high behavioral scores.

Is this the right approach? Just wipe the slate clean and start over - for demo only? 

- fix the demo scoring, batch for backwards and trigger for future
- create a campaign to set every demo score to 0 
- then activate the scoring campaigns

Below would be the campaign to set everyone to 0.

[Brigid_Greenwa2]

The value of the flames and stars are stored as fields - so it goes a bit beyond the Sales Insight window as it will be sending a (potential) new value - so it will increase the amount of data that they are going to sync at any given time - hence the Friday afternoon suggestion.

[Josh_Hill13]

Brigid, I think you are thinking of Sales Insight, which won't really affect anyone other than the MQL system. Probably a good idea to do this on a Friday afternoon just in case.

[Josh_Hill13]

You can use that flow, but use =0 to set the score to 0.

But shut off the other demo scoring campaigns before you do this. And you might want to shut down the MQL sync temporarily in case the scoring triggers a bunch of MQLs accidentally.

Then go through each of the flows to verify the Schedule is set right, so that you don't have this issue again!

Then restart the demo scoring.
 

[Brigid_Greenwa2 · Accepted solution]

Hi Christine-
Are you connected to your CRM live?  Reason I ask is that radical score changes are going to affect the entire database as MKTO's algorithm is partly derived from what Lead1's score is relative to the rest of the database.

If you have connected to your CRM, I would suggest making that change on a Friday so that it can re-calculate on the weekend.

As for if you should wipe entirely or try to correct, that is going to be mainly dependent on how broad your scoring spectrum is (do you have leads with a score of +5 to +100000000000) and how large a data set you are trying to course-correct.

I would suggest that you split out demographic information (like account name, title, industry etc) away from behavior score changes (how many times you do something).  If you need help, there are a lot of articles and presentations on this - but at days end, you will be able to right-size your ship with a little bit of investigation work with some small batch changes to see how things get affected.

Hope this helps!