3 Replies Latest reply on Jun 22, 2017 8:36 AM by 831fa11f7d15c46f1bdc2d96bb1f35ecb08a1782

    De-duping

      Does anyone have any strategies for how to accomplish the below projects without 'adding' up leads scores of prospects involved? For context, I plan on using Cloudingo (a deduping tool, not marketo) to execute the below. Would appreciate any insight or approaches.  The below is not difficult, but figuring out a kosher way to preserve one of the merged prospect's 'lead score' or to maybe even accomplish a 'blended' score has been hard to come by.

       

      Projects:

      • Perform a massive deduping project of leads->leads, contacts>contacts, leads->contacts.
      • Perform an automated 'ongoing' de-duping project that will run nightly.
        • Re: De-duping
          Josh Hill

          Does Cloudingo allow you to design workflows that would do what you want?

           

          When you say "kosher" what would that mean to you or your firm? Clearly, you want to avoid false MQLs on additive Lead Scores when merged. So you could average them, you could keep the newer record's score, you could attempt to re-score them based on recent activity on the merged record. (a littler harder).

           

          The reality is some data is lost and you have to make a call.

          1 of 1 people found this helpful
            • Re: De-duping
              Grégoire Michel

              +1 on Josh.

               

              Vote here: Define how Marketo should deal with score fields on merge

               

              There is always the solution to export the data before merge, then reimport it fter. Be aware that they are some restrictions on whether or not you can update score fields with Import, though (If I remeber well, you cannot update standard score fields, but I might be wrong here).

               

              -Greg

              1 of 1 people found this helpful
                • Re: De-duping

                  "There is always the solution to export the data before merge, then reimport it fter." -GM

                   

                   

                  I'm having a hard time understanding how in the long term that piece of advice would work. I agree a sort of 'imperfect' judgment call would have to be made at one point or another - let's say that the imperfect judgment call I make is that I always want the *higher* score to be preserved in a field I call 'Lead Score - Post Merge'.  These judgement calls mean that I would:

                   

                  1) Export the data I've listed below in the  'Part 1: Pre-Merge' section

                  2) Merge the records

                  3) Upload the 'higher' lead score to the field called 'Lead Score - Post Merge'

                  4) Be left with the data listed below in the 'Part 2: Post-Merge' section

                   

                  Part 1: Pre-Merge

                  -Jane Doe - Lead Score: 15

                  -Jane Doe -  Lead Score: 30

                  -Michael Baker -  Lead score:  11

                  -Michael Baker - Lead Score: 4

                   

                  Part 2: Post-Merge

                  -Jane Doe: 'Lead Score: 45', 'Lead Score - Post Merge: 30'

                  -Michael Baker 'Lead score: 15','Lead Score - Post Merge: 11'

                   

                  *****

                   

                  For that 'moment in time' the data is ok.  I'm comfortable with the level of 'imperfection' that using the ''Lead Score - Post Merge' field as the lead score of record offers.

                   

                  But what happens down the road? When Jane Doe downloads an eBook (let say +5)  and her scores changes to:

                  -Jane Doe: 'Lead Score: 50', 'Lead Score - Post Merge: 30'

                   

                  How am I supposed to incrementally capture that +5 in "Lead Score - Post Merge"? If I set up a workflow in process builder  that says 'every time a lead is created or edited' then update 'Lead Score - Post Merge' with the 'Lead Score' value then I will wind up with the below which leads me to the sort of horrible-quality data I'm looking to avoid.

                  -Jane Doe: 'Lead Score: 50', 'Lead Score - Post Merge: 50'

                   

                   

                  In light of this more flushed out description of the problem does anyone have any advice?