Deduplication may not be the stuff of everyone’s dreams, but it could turn out to be more interesting than you expect.
How many duplicates can you expect to find?
Our rule of thumb is that within any single customer system there will normally be between 5% and 25% duplicates.
However the more ways you have of identifying an individual, the higher the level of duplicates normally uncovered. For instance, if names and addresses can be combined with email or mobile number, many records can be brought together that otherwise would have been kept separate.
There are no rules of thumb however about the level of duplication between different customer systems held by the same organisation; but as an example our recent work with a media company selling a range of direct to consumer services revealed that for every 100 customer records held across their systems, there were in fact a net 75 individual people.
So why does this matter?
Perhaps the most obvious reason is that deduplication will stop you sending two communications to a proportion of your customer base.
Just stop to think just how irritating it is to have to open or delete two emails from the same source with the same content.
And then, if you are using paper and post, there is a big cost implication of not getting the deduplication right.
The second reason is GDPR. How will you handle individuals’ requests to be forgotten when there are two versions of these peoples’ records? And how foolish would you feel when sending customers copies of the data you hold on them when clearly it came from two separate sources?
But probably the most interesting aspect of deduplication comes when you pull data together from across different systems and sources.
Mr Smith who buys holidays, and is on a dating sight you run, has a distinct profile; so does Mrs Smith who buys wine and cooking equipment, and so probably likes entertaining at home.
So whether you have a single customer file, or customers spread across several systems, the case for deduplication is clear, but just how clear it is can only be quantified when you have matched all those customer records together.
To find out more about Unifida please click here