NAVTEQ VS Tele Atlas
One of the big debates in the USA GPS market is the source of map databases. Most GPS enthusiasts have maintained that NAVTEQ data is more complete in the USA. I’ve always maintained there are errors found in both systems and the differences between the two are negligible. With the debate as hot as ever, I decided to put the two map sources to the test.
My own anecdotal experience is that I’ve found errors in the maps from both vendors. I never saw any large gaps in accuracy between the two. Therefore I’ve always recommended to people that the map vendor should not be a large factor in choosing between two GPS devices. Now I will find out if my advice had been any good.
To perform the test, I made sure I had the most recent databases available from each vendor in a popular GPS device. On the NAVTEQ side I used the most recent maps available from Garmin. For the test I used a Garmin Nuvi loaded with Garmin’s “v8″ maps which shipped in July of last year. On the Tele Atlas side I used a TomTom GPS loaded with their newest map release which shipped in December of last year. So yes, there is a difference in ship dates, however these are still the most recent maps available on those devices today.
I wanted to get a variety of addresses from a variety of locations throughout the USA. I decided to test 250 business addresses and 250 residential addresses. For the business addresses I picked a large group of retail stores which had all opened within the past five years. While the business is new, the address probably existed beforehand. For the residential addresses I took a random sample from a membership database of a USA organization. Hoping to get at least some addresses from new housing developments, I only took members who had changed their address within the past two years. (Hopefully some of those were moves to new homes rather than moves to existing homes to make the test harder.)
Now that my list of 500 total addresses was complete I set out on the painstaking task (literally painstaking) of trying to find each of those 500 addresses on both GPS devices. If I couldn’t find the address right away I didn’t stop there since I didn’t want to test how easy it was to enter an address on each device (although I did learn quite a bit!). Rather I wanted test to see if the address existed in the GPS.
I double checked spelling, tried different ordering of words in the street names, tried different prefixes, suffixes, etc. And here is one important point… A match was only counted if there was an exact location match to the street number in addition to the street. Just finding the street wasn’t enough, I wanted to find an exact location match for the street number as well.
As for the 250 business addresses, I was able to find 83% of them within the NAVTEQ maps. I found a few more, 89%, with the Tele Atlas maps. A difference between the two of only 6%.
The residential results surprised me. I expected to have overall lower matches since I would assume new housing developments (streets) pop-up more often than new business addresses. I was able to find 86% of the addresses from the NAVTEQ maps and 92% of addresses from the Tele Atlas based maps. The difference between map vendors here was again 6%.
A few footnotes pertaining to this study. While searching for 1,000 matches may seem like a large sample, it only represents a tiny fraction of each address number on each street in each city in each state in the USA. While in this test the Tele Atlas data was more complete and resulted in a higher percentage of matches, there wasn’t a staggering difference between the two. Also just because a street address was found, doesn’t necessarily mean that it the address is drawn exactly where it should be.
TeleAtlas VS NAVTEQ
So would I now recommend Tele Atlas maps over NAVTEQ in the USA? Not necessarily. I’m keeping my prior position that the map vendor isn’t currently something that should be a big consideration when looking to purchase a GPS device in the USA. If these results are representative of the entire database the difference between them would only amount to finding less than a dozen more addresses in a year, navigating to a new address every day.