Data Quality Good PracticesSOLVE
Sep 3, 2020 3:29 AM
We're looking to scale and automate big parts of our business and data quality plays a crucial role in that.
Probably the biggest issue is that our Company data is in rather poor state. There are a lot of duplicates, there are no parent/child relationships, there are lots of companies with no contacts and companies with just one deal that was closed as lost, and naming conventions are not aligned - just to name a few issues. I can do de-deduplication through the Merge Duplicates feature but the rest of the work is quite manual. Going through companies manually to create parent/child relationships seems especially tedious.
In essence, I'm looking for good practices to clean our Company data with less manual work as well looking for good practices to prevent bad data from forming in our instance in the future.
Any help is much appreciated!
Solved! Go to Solution.
Sep 3, 2020 4:54 AM
I want to make sure I set the right expectation with all my advice and say right off the bat, cleaning up your data can and most likely will take up a lot of time and will involve manual work. But the good thing is once it's done, it becomes so much easier to keep on top off and to implement strategies to keep your data clean.
When it comes to the parent-child relationship side of things, I would first start with a priority list of companies to work with. If you go into every single company it will, as I mentioned, take a lot of time and you could miss updating important companies or records with the potential to close. Your priority should work from known customers all the way down to companies with whom you have no relationship with. Lists and filters will be your friend in these cases. An ideal priority flow a few of my customers have used from their filtering or lists is:
- Companies/Contacts that are marked as a customer
- Companies/Contacts that have a deal in the pipeline
- Companies/Contacts that are marked as an MQL/SQL
- Companies/Contacts that have a particular lead score benchmark
- Companies/Contacts that have interacted with your emails or website in the past 30 days
- Everyone else
For finding companies with no contacts, this community post should help you out on creating a filter.
With the naming conventions, this is often the biggest struggle because you could be dealing with many different factors such as different names for the same thing, misspelling, etc that make building filters or lists really difficult. Tackling this goes more into how to future proof it from happening again. The number one solution to that is using dropdown properties whenever you can. The beauty of these properties is it gives you control over what names and data should be inputted, no more spelling mistakes, or misalignment on what name should be used.
The next part of future-proofing goes more into how data is inputted into HubSpot. Again, where possible, use the drop-downs to help with data alignment. If you are importing information this will be great as it can also be set-up in Excel or Google Sheets, making for less work on your mapping when you import the sheet. Once this is all done, I'd recommend highlighting what information is needed before someone is contacted by sales and create a snippet for your sales team to follow. You can also create an internal pop-up on the deal pipeline that will ask for the required information before you can move a deal from one stage to the next.
There is so much that I could go into here on how to manage your data post-clean up, especially as you mentioned that you want to implement automation. If you want to share some of those details please do and I'd be happy to give some recommendations. If you have a Customer Success Manager I'd also recommend connecting with them, as this is also something that they would be happy to run through with you.