A woman in Maine, USA receives 500 letters from United Healthcare within five days

Interesting story I came across tonight;

A woman in Maine, USA receives 500 letters from United Healthcare within five days

I work with and process very large, complex databases regularly, and the article had me speculating about how this could have occured. A worker from United Healthcare claimed it was a coding issue. The article did not elaborate if it was either a United Healthcare or the mail house coding error.

There are a couple of scenarios where multiple letters to the same recipient could occur:

1. No deduping, automated or manual checking process used by United Healthcare or the mail house.
2. Problem within United Healthcare system not flagging recipient as already mailed to, therefore generating new data, which is subsequently processed by the mail house.

Typically, a mail house could potentially pick this up but there a few scenarios where it may be impossible for the mail house to do so;

1. Data sent from United Healthcare to the mail house occurs frequently, perhaps daily, so these daily mail outs could be treated as individual jobs and therefore unlikely to be deduped against the previous days mail outs by the mail house.
2. If data received by the mail house contains hundreds of thousands of records, then a block of 500 identical recipient name and address could feasibly slip through, especially if there are no processes to check for duplicates.

In the letters received, instead of the recipients name in the ‘Dear’ field, it was “Maine’s Department of Health and Human Services”, suggesting that there was a major problem with the database or the letter template.

These problems typically arised when dealing with large volumes of complex data and using automated systems that require no user intervention. As the article suggests, it’s very easy to see how just a small coding error could have resulted in a single individual receiving mutliples of the same letter, which is unfortunate for the recipient.

On rare ocaasions, when I am processing customer database, sometimes the data ‘just does not feel right’; it might be a large allocation to a single BSP code, or a high percentage of unbarcoded to barcoded recipients but I always use that as an opportunity to go back through the data, or even start over from scratch to verify the output is indeed correct. Like any experienced home renovator would tell you; measure twice, cut once.

Copyright secured by DigiproveSome Rights Reserved