A McDonald’s Story
Local is the message at GDC. By sourcing data from the best available sources in any market, we’re quietly building Worldview into a platform that produces the world’s most reliable global address and identity validation results. As a famous politician once said (and we paraphrase), you can trust but you must also verify. To that end, we’re constantly testing the data we get from in-country providers to verify that it actually produces better results.
Every once in a while we run across some data examples that help us tell the story of our mission. Last week we were comparing results for a UK address we received in some sample data. We happened to know it was a McDonald’s Restaurant in London. Presumably a potential customer called to ask for the store address, and someone told him to look up 30-32 Saint John’s Road, London, UK.
Okay, that’s all the input we got. So first we ran that through one of the generic international address engines. The generic parsed the data, re-formatted it, and came back with a postal code. This is a verified address, it told us. Feel free to drive there.
The only problem was…it was all wrong. The generic engine’s rules stripped out the “30-“ and just found an address with “32 Saint John’s Road.” Then it went looking for a postal code that had that same number and street combination, and it found one. But the wrong one. Even though it said confidently it had a reliable result for the address this customer was looking for, it was in the wrong part of London and it certainly wasn’t that McDonald’s you were looking for.
So next we ran it through the GDC Worldview engine. It parsed the address, kept the street number as “30-32” and affixed it with a (different) postal code. The right one, it would turn out. It verified the address as reliable and added one piece of useful information that our best in-country local data provider happens to track…the name of the business establishment at that address. It told us this was a McDonald’s. The very one our hypothetical customer was looking for.
Now data is a tricky thing. No address or identity system is going to be 100 percent accurate. But the data they pull together and the rules they use make a big difference. That’s the story of what we found here. The generic address engine didn’t get the nuances of the street address (that you can have a hyphenated result like 30-32), so its rules stripped out part of the number. That guaranteed it would find the wrong postal code, a full 10 miles from the actual McDonald’s. Moreover, our provider enriched the basic address information, enhancing it with the business name. This extra bit of information – that local touch – makes a big difference when you’re trying to create the right experience for the end user. When you’re helping your customers find your restaurant.
What data examples do you have to share?