Data Quality and the Golden Record
Welcome to my new series: “Questions that I got asked while presenting at a Conference that I have been ruminating on since”.
Working title. We’ll come back to that later.
Long and short, thinking on my feet is something I do well, but thinking out an answer over time and doing some research to reinforce my stance is something I do better.
Todays Question –>“What is your stance on Data Quality in the universe of CRM’s and CDP’s?”
My answer there can be summed as, “It depends, but chasing the golden record is probably foolish.” This is something I whole heartedly agree with, but doesn’t really tell the whole story. Data quality is absolutely pivotal and essential to getting what you want out of your process. That is something I did not emphasize enough.
What I meant was, data quality is going to mean something different to each business and business process. Pushing all of your data systems into a single source and single record is likely a lesson in futility and may actually hinder your sales process. Just take a step back for a second and ask yourself what you’re actually doing with that record. What kinds of questions are you trying to answer about your customers? What kinds of processes are you trying to power with that data? Even answers to those basic questions will get you closer to right sizing your “Golden Record” definition.
High Level Strategy
Generally speaking, think in tiers of quality. I prioritize getting the information about paying customers AIR TIGHT. You don’t want to accidentally send a paying customer wrong information or try to upsell them something they already purchased. This should also be the easiest layer to get correct. This layer is about getting your CRM and POS or Sales department processes in alignment and letting them run.
That said, the less information I have about a customer, the less I value attempting to get that information “right”, and more about respecting how they gave me information and for what purpose. Fill out a form asking about X? Save the record and give them X. Unless that person already shares an email address with another record, trying to do the dance around “uniqueness” just takes a lot of processing with little pay out.
Final thought here is STOP HOARDING DATA. If you’re saving a whole contact record and their whole history in your primary database with hot storage that has not interacted with your brand in 10 years, you’re just setting money on fire. Do some testing, pull out your pareto chart, and find the reasonable baseline for max tenure for win back campaigns and then let the rest go. Or at least push their data into cold storage to reactivate should they return.
