Seems like most tech guys crying foul at the market response to Apple’s announcement either don’t get or willfully ignore basic finance: the price of Apple’s shares this morning has already priced in the historical and *expected* earnings from the company, based on past guidance and etc. The drop in price is a result of Apple not meeting expectations and the worries about the company’s future growth. The hand wringing and exclamations of Apple’s current profits have little to do with the share price, which already should have factored in historical AND expected performance.
And really, a bit of me is crying inside at my positions, too.
“If Apple can take this sort-of-passable Maps app and make it into a really great one, with functional transit directions, better map tiles, and all that, it can do anything. If it can’t, for whatever reason, that’s something to legitimately worry about.”
Paging map nerds - how hard of a road is this for Apple to get a respectable maps experience for its users?
Since Tumblr’s reply box seems to have a character limit, I’ll put my attempt to answer that here:
Looking at reviews and sites like http://theamazingios6maps.tumblr.com/ there seem to be multiple issues with Apple’s maps, among them:
- The data quality of the actual maps, sometimes data is entirely missing and sometimes it’s wrong and/or old.
- Search and geocoding, problems here can be results of both bad and incomplete map data as well as weak algorithms.
- Satellite/aerial imagery, they simply don’t seem to have enough up-to-date images at high resolutions.
Now how could you tackle those issues:
Let’s start with the imagery, I’m not as familiar with those as with map data, but from what I know you can get pretty far by simply spending lots of money to buy those images from imagery providers around the world. So you basically need money (not an issue for Apple) and people being able to make those deals (might be an issue for Apple right now). Plus making all those deals and stitching the images together takes time.
To get to Google Maps like quality however I think you would need go beyond that and make exclusive deals and maybe even have some planes fly around for you (see for example what MS/Bing has been doing). That takes more expertise and a lot of time.
Now for the map data, what Apple seems to be doing so far is buy data from different data vendors around the world and combine it to a worldwide map.
To improve here they would need to go out and license more data, which costs money (not an issue) and again you need time and people being able to make those deals. There is simply not a single data provider that has good or even great data for the whole world. For the western world TomTom (previously TeleAtlas) is probably fine, but is basically the only data provider since the other big one, Navteq, is now part of Nokia (that also provides Amazon’s maps now). For POIs and search Apple seems to already license data from additional sources like Yelp, they could also do more deals there (Foursquare comes to mind).
For the rest of the world you can talk to local data providers, obviously there are lots of them around the world with varying degrees of data quality plus you need them to agree to a license which allows you to later mash the data together… And in some third-world countries there are simply no official maps you can buy or license, that’s why Google started their MapMaker program and OpenStreetMap often provides the best maps you can find in those places.
After you’ve licensed all that data you need smart engineers that can combine those data sets in a way that only the best data is used for a given place, which doesn’t seem to be the case with the current edition of Apple’s maps (see also). Again you need expertise and time to do that.
If you want to get to Google like quality, well basically read what Google is doing in the recent Atlantic article.
A few years ago Google moved from licensing data to buying data plus using public datasets to create their own dataset in a lot of countries and then improve it constantly with their own data streams (Street View, location data from Android devices, bug reports from people, crawling the web for POI information and getting people to work for free for Google through MapMaker). This again takes time and you need lots of people who do the manual work of making maps out of those “raw” data sources (Google apparently outsourced most of that to India). Apple doesn’t have access to Street View like imagery and their maps team doesn’t seem to be nearly as big and powerful as Google’s yet.
The only other way to get access to a dataset that is as detailed and as up-to-date is to use OpenStreetMap, which is crowd-sourced by people around the world (including volunteers that actually walk around with a GPS device, but also government agencies, non-profit organizations and even some companies contributing data).
The main issue with using OpenStreetMap data right now is that the quality varies around the world and different regions, simply because in some countries (e.g. Germany) a lot of people have been mapping since 2005 while that hasn’t happened in other countries for various reasons. With OpenStreetMap data you can also run into issues of data quality, that is map data might be fine for making maps to look at (like Foursquare’s maps on their website), but not good enough for turn-by-turn directions.
To sum it up, I think it will definitely take some time before Apple fixes all the issues they now have. I won’t try to predict if or when they will catch up with Google, it obviously depends on which of the outlined ways they choose to pursue and how many resources they decide to invest in it.
I hope that’s somewhat understandable and gives a good overview without going into too much detail. I’ve been contributing to OpenStreetMap since 2008, attended some geo conferences and read quite a few articles over the years, but of course I don’t know nearly as much as people who are actually working on that kind of stuff at Nokia/TomTom/Google/Apple etc. Corrections and additional thoughts welcome.
Pro tip: if you want a “conversation” to be displayed at all in a logical way, you’d best reblog the post as a text post (instead of link or quote).