Find Quality E-data at a Fair Price
By John Ripa
In any good e-mail customer-retention campaign, the first step is to secure high-quality, reliable, properly obtained e-mail addresses for your existing customers from a dependable data provider.
This task, however, can be much more difficult than it sounds. There are many choices out there. Which is right for you?
To help you sift the wheat from the chaff, here are two questions to ask each e-mail data provider you interview.
"What's your match routine?"
Ask if they provide "string-level matching" or a more sophisticated system. Traditional string-level matching takes components of the name and address, scores them and creates acceptable levels of difference between the person on the provider's file and the one on the customer's file.
The problem is that there's no knowledge needed in the string-level matching system. It tells you if things are identical or not, and if they aren't, it tells you how close they appear to be. But that's often not good enough. You need a conservative, confident match routine that's based on a comprehensive store of real knowledge about consumers.
Using a knowledge-based match routine, you can ensure that the e-mail you're sending goes to the right person. Why pay for guesswork?
"How many of these records are deliverable?"
Remember, it's not the number of e-mail records you get up front that counts; it's how many of them actually are deliverable. Of course, you can't expect 100-percent deliverability. On any given campaign, odds are that 70 percent to 85 percent of the addresses matched to your customer file actually will be deliverable.
Your provider should send the initial message on your behalf, handle the opt-out issue, check deliverability, then append only the deliverable addresses to your customer file. Avoid simply asking "What's your match rate?" and then letting the higher rate win. Make sure you run a test on the provider's file. It's the only way to get an idea of how deliverable it is.
How do you test? Consider taking a portion of your housefile and breaking it into three segments. Take 10 percent from each segment, and send it to three data providers. Let them do the match and deploy the message.
Then, either apply the data you get to your customer file with source flags to indicate where you bought the address, or simply market to each provider's list separately.
Track results in time to show how each data source performed. And visually scan the e-mail addresses that are returned. Sometimes you'll find nonsensical addresses lumped in with the data for which you're paying.
John Ripa is product leader, InfoBase eProducts, at Acxiom Corp. He can be reached at john.ripa