One question that comes up often when record matching is, “What rules should I prioritize when matching records?” There’s actually no right or wrong answer to this question—it all depends on your specific business case!
Tim Sidor sat down with SD Times to talk about matching rules and cases in part 5 of this ten minute webinar series about verification! Here are some highlights of his chat that can help you figure out what matching strategies you might want to implement into your processes.
Generally, there are two ways to approach matching: loose or tight. You might want a loose approach if you have too many unwanted records, or are purging your database after an influx of customers who may have added their information multiple times to enter sweepstakes, get discounts, etc.
Some loose matching strategies consist of fuzzy matching algorithms, which flags variations in names or companies, such as Jon vs. Jonathan or Melissa vs. Melissa Data. You can also add multiple OR logic conditions to flag more matches, such as setting a rule to flag Name AND Phone Number OR Name and Email Address. This will find more matches that can then be cleared from your database.
A tight matching approach would be used if you want to prioritize accuracy—if your goal is to create golden records for each customer instead of clearing out an old database, you won’t want to flag false matches. You can customize your rules to reflect what you want to prioritize, whether it’s the most recent, most complete, etc. You will also want more matching fields—for example, First Name AND Last Name AND Address AND Phone Number. You’ll want to avoid the OR logics that a looser strategy might entail. This will lead to less matches, but more overall accuracy.
One last consideration you should have while matching is whether you want to cleanse in real-time or in batch. With batch cleansing, you will want to make sure your matching strategy works the way you need it to, because once your data is merged or deleted, it’s much harder to revert back to where it was before. With real-time matching, you can set a series of logics and evaluate the outcomes. That way, you can see what logic the entry matches, and decide from there whether it’s a true match or not.
To help with your matching needs, Melissa offers MatchUp, where you can select data types, fuzzy algorithms, and other soft matching and settings specific to your business needs and is easy to use and flexible in deployment.
If you’re interested in learning more about matching and other verification processes, check out our 10 minute webinar about "Achieving a 360-degree Customer View with Custom Matching Strategies", hosted by SD Times and presented by Tim Sidor. If you want more information on MatchUp or other address and identity verification solutions, reach out to us at www.melissa.com or call 1-800-MELISSA. Don’t forget to subscribe to our blog for everything related to data quality!