Redrawing the map

[unixronin]

This article by Bill Gurley talks about how Google has disrupted the turn-by-turn navigation market by offering turn-by-turn map data "less than free" with every Android OS copy, now that Google Maps is based entirely on Google's own mapping data.  Not only do they not charge for the map data, but if you use it and their mapping API, they give you ad revenue splits.  This announcement from Google immediately knocked a big chunk off the values of both NavTeq and Tele Atlas, previously the only two sources of turn-by-turn map data.  TomTom's stock dropped 16% after Google's announcement, and Garmin's dropped 21%; between the two of them, they own the majority of the GPS navigation market in the US, particularly after Magellan got bought out.  (I found it surprisingly difficult to find current hard numbers, but it looks like as of the end of 2Q2008, Garmin owned 55% of the US GPS market, with TomTom in second place with 18%.  In Europe, the situation is reversed, with TomTom at 38% of the Eurpoean market and Garmin in second place with 19% as of 2007.)  When you're selling GPS navigation capability based on map data that you're charged a hefty fee for the use of, it's pretty hard to compete with GPS navigation based on map data that you're paid to use.

There's an interesting side effect to all this, too.

Let's recap a little.  Google used to use both NavTeq and Tele Atlas.  But in late July 2007, TomTom bought TeleAtlas, and less than three months later, Nokia bought NavTeq.  That put the writing on the wall for turn-by-turn navigation.  A year later, Google dropped NavTeq, keeping Tele Atlas after a negotiation for looser license terms; and just over two weeks ago now, Google dropped Tele Atlas as well, cut over entirely to its own mapping databases, and made its announcement.

Now, when we moved here, it didn't take us too long to notice that the map data for our neighboorhood was wrong. Streets shown as connecting on the map don't, and the map shows streets that don't exist.  This probably comes from the long-standing practice of commercial mapmakers of introducing deliberate errors into their maps in order to be able to detect and prosecute unauthorized copying.  If you can show in court that a competitor's map faithfully reproduces intentional errors that you have placed in your maps, it makes a strong argument that they simply copied your maps.  Of course, it's bad news for you if you, as a map user, are relying upon that section of the map.

Well, we tried to report the errors on the map.  It took considerable hunting by several different people, one of them a Google employee at the time, to find a place to submit an error report to NavTeq.  I even submitted, along with it, a digitally corrected version of that section of the map.  Not only did NavTeq not correct the error, they never even responded.  More recently — after, I now know, Google dropped NavTeq and began using only Tele Atlas — we noticed that the map had changed, but was still wrong, although the major error had changed — the non-existent connection between Cheshire Circle and Briarcliff Road was gone, but Briarcliff now trailed off to the south instead of running westward.

So, after finding out the news of Google's switchover, I just looked again at the map of our area.  And it's now almost entirely correct.  There's only one error remaining ... and there's a link to report errors, right there on the map.

I'd say this change at Google Maps is going to end up a clear win for mapping users.

This entry was originally posted at http://alaric.dreamwidth.org/7248.html. That post currently has comments.
You may comment there via OpenID even if you do not have a Dreamwidth account.

Comments

[perspicuity.livejournal.com]

a lot of the USA mapping data was free from our govt, so some errors are common... map companies acquire data, massage it, correct (sometimes), and that's their value add.

most map data in europe is EXPENSIVE and very not free. oh my no. or just restricted entirely. maps? you don't need no stinkin' maps.

google... they have freaking fleets of GPS cars, and they are mapping all KINDS of crap. they buy (and make) sat data, photo data, all kinds of data.

M$ of course, uses helicopter and plane fly bys to get multi depth 4 way ordinal data (see maps.live.com if you haven't, and zillow.com for an interesting use, if they're still using it that way).

it's semi-exciting. any iphone has maps on it now. it's the MAJOR reason i got the stupid thing 2.25 years ago now. COMPELLING doesn't cover it.

i'm on my third garmin (upgrade/use path, not because they broke - great warrant and flat rate coverage if out of warranty), and will not give that up easily (60csx) - i paid $$$ extra for a route-enabled map set. it has a host of functions and a precision that most other devices (including the very latest iphones) cannot touch. more importantly: it's completely passive. i'm not tracked or aggregating my use of the data to anyone (yes overlord Google ;> spider this, you can't seeeeee me, i'm not using your maps for certain things ;>) oh yes, google android devices? "please track me, off, but well, they still can, they just aren't sharing it with your friends, just the spooks)

many of the services now recognize (finally), the need for timely data. they have yearly disc upgrades, quarterly subscription models, etc. neat stuff.

#

[bikergeek.livejournal.com]

also lots of the time that government data is just plain wrong. Developers file plans for new housing developments with town planning commissions. Then the housing market collapses or the builder goes bankrupt or something else happens and the streets never get laid out and the houses never get built. But maps for years afterwards will show the "paper streets" for the development that never got built.

Or the plan as-filed doesn't match what actually gets built, but the copy at the town hall never gets corrected. E.g. a street that's shown as going through to a main road ends up getting built as a dead end. I would suspect that there's a lot of that going on in that part of Gilford.

If you look at an area of Google Maps that shows lot lines it's easy to see where paper streets were laid out once upon a time. It's not uncommon for paper streets to be given as part of metes-and-bounds (http://en.wikipedia.org/wiki/Metes_and_bounds) land descriptions. One house I was looking at semi-seriously had a property description that used streets that were apparently never built, and a different one used as part of its description a road that hasn't been used in probably at least a century, because it wasn't on any maps that I could find.