Boffins have been working on sophisticated algorithms for name and address matching since the days of the mainframe, with each iteration becoming more and more sophisticated, working harder and harder to uncover contact duplicates within and between data sets.
Controlling these matching techniques according to the purpose of the match is something the data owner needs to be not only aware of, but also be in control of.
Matching techniques for Direct Mail deduplication and data suppression are very different to those you’d use if matching your customer data to a validation data set, when the goal is to append individual data or flag within your database.
When it comes to processing data for a Direct Mail campaign, we’re not going to be too concerned about a small amount of overkill in matching. The cautious approach of not wanting to mail someone multiple mail packs or suppressing data from the campaign as the customer may be deceased or moved address is a common sense approach. However, when it comes to updating your customer database with suppression flags, or consolidating customer accounts; it is a very different process. You need to be confident that the information you are applying to your database is accurate. You will probably not mind missing a few matches, but those matches you do accept have to be accurate – unless of course you don’t mind losing contact with active customers?
So how do you control your match levels? In essence, the two different matching types are basically controlled by the Name element matching. We can take it as read, that address level matching is working OK, but the anchor has to be the name.
Here are a few scenarios for you to consider..
- You’re matching your customer database against Royal Mail’s National Change of Address data, in order to identify a customer home mover. Matches to NCOA would retrieve the customer’s new place of residence.
- Matching your customer base against our data pool of Mortality Suppression data. This intention is that you will update your database with the final life-stage information (Deceased).
- Matching your customer base against our data pool of Goneaway Suppression data. This intention is that you will update your database having lost contact with this customer, which will then activate email & telemarketing campaigns to update the customer’s address information.
Take a look at the table below – see what we mean about complexities of name based data matching? It’s not something you can shy away from though. It’s an essential part of data quality update processing. But which of these record groups should in fact match? There is no right or wrong answers to this question. Either way, you now have an enormous challenge in configuring your matching according to the rules you’ve now defined for database update matching.
Data matching and any subsequent updates to your customer database aren’t easy processes. You need to be, not only in control of your matching techniques, but also add as many anchors to your matching process to improve the confidence level of any match.
You can, and should, also use customer activity indicators as part of your matching validation. If a customer has just purchased from you as an example, you certainly won’t accept that the customer is dead – would you?
You’ll also need to consider potential identity fraud…
The use of additional anchors, such as; Mobile Number, Date of Birth would certainly benefit the confidence levels when completing name and address matching, but these do not exist on many 3rd part validation files – perhaps they should?
We’re not going to go into the validity of the data contained on the Industry Suppression files – that’s a topic for another day. However, one thing we’d highly recommend when dealing with Suppression data matches is that you do not accept matches without further validation.
A typical validation workflow (below), would drive potential updates (deceased, goneaway or movers) to a further process, such as; an email contact or phone call.
We hope this is food for thought and sincerely hope this post helps improve the accuracy of customer database matching. If it’s a challenge you need help with, feel free to leave a comment and we’ll be more than happy to chat the options through with you.