Deduping accounts when people 'just know' they are duplicates - Open to suggestions!

[GregBKyruus]

Hi all!

I'm still very much a neophyte when it comes to DemandTools. My solution to this issue will make that abundantly clear. Read on, if you dare!

We've done a great deal of data clean up, and now we are getting down to the stuff that is really institutional or even just individual knowledge as well as lack of processes/guardrails. This is very tough to automate, and I have a hokey idea on how to do so. I need help with how to do the matching. I'd also be happy to hear ideas on a better approach to this.

Here's an all too common scenario:

Hospital System A changed its name last year, and used to be named Hospital System B. 

A new rep was told Hospital System A was theirs after the former rep left, and they simply created a new account using it for all their work as they move forward (new opps, contracts, etc).

The account record for Hospital System B (the original record) was never reassigned nor updated, and the Customer Success Team has continued to use that account for existing stuff (cases, customer meeting notes, etc.) 

Hospital System A also acquired Hospital System C six months ago, six months after the name change from B to A. 

Both A and C are customers, and the account record for Hospital System C was never updated to reflect the acquisition.   

All the while, everyone kept on doing their thing with C, and just worked around it.

Often, these accounts will have wrong or incomplete addresses, out of date websites, etc. It's tough to catch these with a DemandTools job and we are left merging these on the fly using built in Salesforce functionality.

The only way I can think of to start moving towards handling this with a DemandTools job is to do the following, using the example above.

Leave Hospital System A as is.

Rename both Hospital System B and C as "Hospital System A-DUPE" respectively.

Figure out how to match on a name before the -DUPE.  I'm not sure how to do that.

Create a Winner Rule that down-scores any account with -DUPE appended. Easy enough.

Merge based on name before or excluding -DUPE, with the account that doesn't have -DUPE appended as the winner.

Again, this is a very janky, hokey approach to cleanup but I can't figure out another way to leverage DemandTools for cleaning up data when there's big gaps in the data and users are basing the claim accounts are duplicates based on their own and/or internal company knowledge. Help!  

(x-posting to my CSM and to Support)

[ajavalidity]

Hi @GregBKyruus Thanks so much for your post. Firstly, I'm so glad that you are thinking outside of the box in how you can utilize DemandTools to accomplish your data challenges. Way to go! Secondly, I believe I have some solutions to provide to you for your questions:

1. Figure out how to match on a name before the -DUPE.  I'm not sure how to do that. You can do this by matching on name and using the "First XX words" comparison type. This will allow DemandTools to only match on the first words so as long as -DUPE is listed after the first {insert word number} words, it would be ignored. Additionally, you can make an edit to your Cleaned Account Name comparison type to have the word "-Dupe" ignored. Just add the text that you want to have ignored within the suffix list. You can edit the Cleaned Account Name suffix list by clicking the paint brush icon from the top of the DemandTools V home screen. click the suffix removal tab and add the word -DUPE.

You can also always try not matching on name at all and instead matching on other identifying fields such as billing address and website.

2. Merge based on name before or excluding -DUPE, with the account that doesn't have -DUPE appended as the winner. Create a winner rule that will always keep the account as the winner that does not contain the word -DUPE. 

I hope this helps answer your questions but if not, please be sure to reply here or you can also email demandtoolscoe@validity.com and a CSM would be happy to join a call with you and walk you through accomplishing your use case.

Thanks again for reaching out! 

Have a great day!

Regards,

Aja