When the Golden Record is Tarnished

Golden records, single customer views, call them what you will, are the El Dorado for many organisations struggling with large amounts of data from multiple sources.  They’re a great asset when they’re accurate, but can cause a lot of problems in downstream data quality when they’re not.

Recently, Dutch organisations which should know my new German address, such as the tax service, started sending documents to my old Dutch address, which the new inhabitants were good enough to send on. After some investigation as to why they were doing this when I had informed them all of my new address directly, it became apparent that either I, or somebody in the Amsterdam municipality, had added my new address in the population registry as being at number 17 instead of 14.  This population registry data is used as a golden record by many institutions, and this error had been iterated downstream, overriding any correct information that was already in their systems.

A taxing data quality problem

Much as I would sometimes prefer the tax services not to know how to get hold of me, they have a habit of making wildly inaccurate assessments and taking the money anyway, so I decided it was best to let them know where I really am.

There is a number 17 in my street, but it’s a long way away from number 14, and its inhabitants don’t know me from Adam, so they had been sending back the mail from these Dutch institutions, which in turn led those institutions to question their data and to start sending information instead to the old Dutch address in the hope that it would be sent on.

Once an error like this is in the system, correcting it is far more costly and time-consuming than getting it right in the first place, for both parties.  It quickly became clear that the municipality required proof that there was an error, whilst proof had not been required to create a change of address record in the first instance, and that the correction would not wash down the system with the same efficiency as the error did.

Address validation systems would not have spotted the initial error. Numbers 17 and 14 (in fact, every address in this municipality) share the same place name and postal code. There were checks and balances missing which could have caught this issue, though.  As my better half and I had to fill in the forms separately, and we each indicated that we would both be moving to the same new address, the discrepancy between our records should have been flagged up at that point.

Regardless of whose fat finger caused this problem in the first place, it’s a clear indication on how essential correct data is, especially when it is being used as a trusted and much accessed golden record.