Data cleansing and your name data
 
 

Data cleansing and your name data

by MarkC 20. October 2009 11:38

A large proportion of our data cleansing services revolve around matching the name and address information in your data file to that in our various data sources, so you will get better results from services such as goneaway suppression and deceased suppression if you have more accurate and more complete data already.

The simplest way to ensure you have complete and accurate address information is to integrate our postcode lookup software into every point of contact with your customers, such as your website, call centre and CRM system, and our PAF cleansing service ensures that we can efficiently tidy up any existing addresses in your data.

However, unlike PAF for addresses, there is no single authoritative data source for names for us to use to correct your name data. If you have mis-spelt or incomplete name information in your data, you may be missing out on valuable customer insight.

Our powerful data cleansing systems attempt to intelligently match mispellings of names between your data and ours, coping with names being supplied as separate fields or all together, double-barrelled names and many other combinations, but it cannot fill in any blanks.

If you only store the initials of your customers rather than their full forenames, we have to assume that the "J Smith" in your data is the same person as the "Mr John Smith" that we hold on our goneaway or deceased data source. This can lead to problems where your customer was actually "Jane Smith", giving you a false match.

There are things you can do to avoid this problem. You can get the goneaway and deceased flags anyway, but apply them more selectively based on your trading history with the customer - if they are still buying goods, don't delete them from your database! You can also start a process of enhancing your data with the full forenames during your normal customer interactions and gradually flesh-out your information. Even adding a title of Mr or Ms to customers to indicate if they are male or female can help us eliminate some false matches, as we will not match "Ms J Smith" against "Mr John Smith".

It's worth noting that, if you do have some records with full forenames and some with only initials, you can safely purchase all matches at initial level as we won't match "Jane Smith" against "Mr John Smith" at initial level.

If you have any questions about what matching level you should use for your data, just give your account manager a ring and they'll be able to help you out.

Comments are closed

RecentComments

Comment RSS
Skip Navigation Links
Request handled by server S1 for client 38.107.179.220 at 5/18/2012 7:59:19 AM