According to Sirius Decisions, it takes $1 to verify data as it is entered, $10 to cleanse and de-dupe it, and $100 if nothing is done and the mistakes are felt over and over again.  So if you have a database with 500,000 records and even 10% of that data is bad, you are talking about $5M in cost just by doing nothing.  That should get you into the spring cleaning mode!

The question that paralyzes most marketing professionals, however, is where to start.  If you have hundreds of thousands of records, how do you make sure they get and stay clean?  Here are a few suggestions:

  • Look for a reputable data cleansing and data append service.  Many services do one or the other, so you might be looking at a couple of vendors.  But one service I know that does a great job at both and uses multiple sources in the process is our partner ReachForce.  The key is finding someone who doesn’t just use one source of data.  Also make sure the data is verified frequently.
  • For the do-it-yourself types out there, you can start with de-duping your data using some good old Microsoft Excel functions.  On the data tab in Excel 2007, there is actually a remove duplicates button.  You can also do some find and replace functions to make sure you standardize all of your values.  Once you clean the existing data, you can use the internet or your own internal call team to update missing data.
  • Sending your clients a quick form to update their own data always helps.  You are not going to get 100% compliance, but you’ll get some cleaner data.
  • Once you clean your data, go through your forms to make sure you have standardized picklists and switch as many things as you can to picklists vs. text fields to minimize entry errors and duplicate values.  For instance, instead of title, you might want to gather functional level (C-level, VP, director, etc) and functional role (IT, marketing, etc) which are much easier to standardize and segment on than title.
  • Most marketing automation solutions will let you build a program to look for and clean up data.  For example, you can build a program to look for any values of USA, United States, United States of America, etc. and standardize them all to the two character country code US.

Happy cleaning!