Places, Databases and Duplicative Efforts.

Over the past couple months I’ve taken upon myself to get more familiar with the broader world of technological startups, mostly because I’m always on the lookout for new APIs that I can aggregate into Geon. More recently however I’ve started to notice that there’s a big trend towards making any application location aware and there’s an increasing amount of social networking applications that use places and locations as their main selling point. The current hot thing appears to be apps that let you “check-in” at locations, say your local coffee shop or cinema, and give you rewards based on that. It fits in quite well with the formula I came up with for successful social networks (sense of belonging + wanting to share with community + competition element = win) so whilst I can’t see myself using the service I can understand why they’re becoming more popular. Of course with so many of the services starting to come out of the wood works some obvious duplication efforts become apparent, namely that they all roll their own location databases.

Now from both a business and technical point of view this makes quite a lot of sense. Whilst it’s nice to rely on other people’s services to provide you with data it also poses a risk, especially if that service is made available to you free of charge. Usually you’ll be on the rough end of the stick in terms of usage agreements and they’ll absolve themselves of any responsibility should the service go down. With bigger players though you can usually count on them being fairly reliable (I consider most Google services as 6 Sigma, for example) but when your core business relies on services provided by others you have to ensure that you have strict service level agreements with them or you put yourself at quite a large risk. Keeping the service internal, whilst increasing your own risk profile, at least grants you control so that any outages can be dealt with more effectively.

Still any engineer will see duplicative systems as wasteful if they’re not specifically being used for redundancy. The recent explosion in location aware applications hasn’t gone un-noticed and the duplicative efforts managed to catch the eye of one journalist:

Here is the problem: These efforts at creating an underlying database of places are duplicative, and any competitive advantage any single company gets from being more comprehensive than the rest will be short-lived at best. It is time for an open database of places which all companies and developers can both contribute to and borrow from.  But in order for such a database to be useful, the biggest and fastest-growing Geo companies need to contribute to it.

I put this suggestion to Foursquare CEO Dennis Crowley the other night at a party, and he was enthusiastic about the idea. Foursquare is building up its own comprehensive database of places, which it calls “venues,” through its users who add places they want to check into, if they don’t already exist. Foursquare matches their GPS lat/long coordinates to its database of venues (businesses, points of interest, even people’s homes). Later I followed up by email and asked Crowley, “Isn’t the quality of your places directory, built by your users, a competitive advantage?

He makes a good point, it would be quite advantageous for many location aware applications (mine not so much) if there was an open database that contained a list of places worldwide. Whilst I’m not aware of a similar call for geolocation services (translating co-ordinates into names) the service GeoNames seems to embody the exact idea Schonfeld is talking about, albeit for a different kind of service. For anyone looking to use such services GeoNames provides a very quick way of integrating them into your project and if I hadn’t run into them earlier I may well have ended up spending a fair chunk of cash to get the same functionality (or signed my life away to Bing Map Services, which I’ve already done in part). Still the guys over GeoNames have been repeatedly bitten by those looking to make use of the service and I can’t help but feel the same thing would end up happening to those who would build the database of places. They’d provide a service we’d all love and enjoy, but they wouldn’t be getting a lot of tangible benefits as a result.

Schonfeld makes the point that any small advantage of a place database that has an edge over their competition doesn’t really give a company any advantage. To a point that’s true, since most of the legwork has already been done and it wouldn’t take a dedicated programmer more than a few weeks to replicate a similar database. Still anyone who goes ahead and makes this open database wears all the implementation and operational costs as well. They do gain a decent amount of power by being a central authority for something (which screams Google to me) but it will all come down to whether people co-operate or not. The trend towards an open web makes me think that they probably would, but it’s still a risk.

Right now I don’t envisage Geon actually requiring such a database, mostly because it’s focused on information + location and not so much if that happens to be from a bar or convention center. If such a thing would be implemented I’m sure I could augment the data stream with some place information to give the information a bit more context but it’s currently in the same bucket as the weather for a given location. It’s another cool thing to add on but the audience that I’m targeting probably won’t need it (and it will just add to the noise).

It’s really just a modern version of the tragedy of the commons and the solution is not much different than it was back then. I’m sure everyone would cry foul if the first such implementation came with a price tag for access but unless a large company wants to play the good patron to the rest of the world such central resources will be slow to come about, if ever. Any startup making use of such data hasn’t seemed to have any trouble coming up with their own dataset and it appears that will be the situation for a long time to come.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.