The Continental Drift of the Database

What’s the most basic concept of a data set or a database? I think it’s the primary key. Meaning, the primary key is the single element in a data structure that enables you to integrate two pieces of information. It’s the foundational layer in data management and without it, our ability to analyze data across sources would be severely limited.

Well, take this concept to the marketing database, specifically. For the past 20 years, the primary key of merging marketing data sources has been effectively one thing – the name and physical address of the customer (or prospect). Of course, technology has changed from mainframe processing and NCOA to referential customer data integration (CDI) solutions. But, however you thought about it, the only way to merge data was, basically, the name and address.  And yes, I know, the phone number has been around and things like social security number but effectively, those have not been the primary method of database merging. 

However, that concept of a single primary key is changing. A slow but definitive shift is occurring with the marketing database, where in 5 years, there will be multiple primary keys. Think about how email changed matching in the last 10 years. How many email addresses do you have? (I probably have about 5). Also, the mobile phone has multiple ways to be identified – specifically, the actual phone number, but also the IP address of the data connect and the login id of the app you are using. 

Basically, the foundational layer of the marketing database is changing and it’s changing in an extremely complicated manner. The entire primary basis for CDI technology is going to shift from a name/address construct to a multiple primary key technology where a confounding problem will be introduced – many-to-many matching. The database of tomorrow will have to have a new set of primary keys in addition to the name/address, phone number and email address. It will now have your social handle, the IP addresses of your devices (PC, web TV, phone, tablet, perhaps even your car), cookies, and potentially even an algorithm-based identification prediction. 

So, if you’re still worried about getting name/address matching right, I’d say don’t be worried about that, because the industry has already solved that problem. The new problem that you need to get in front of is a future where the marketing database is the same as the enterprise customer warehouse and data integration is solved through many fronts. This is something we’ve been working on for years and I’d argue that even though we’ve made advancements in this area at Merkle, there is still a long way to go before we solve how to do data integration in today’s completely connected world.

Join the Discussion