PlaceHoldr : On what we read where… / Mar 12th 2010

So the thing I should probably have focussed on a little bit more at Bonnier Hackday was this. It could have been so much more and hopefully, with a little bit of thinking and a bit of effort, it can be. By half-way through the hackday everything was working and instead of polishing and honing an idea, I made other things. Mea culpa, and I won’t be doing that again (although I think they’re all interesting too). 

I’m calling it as often at a Hackday you can’t think of a name and end up with a placeholder one. So this is my placeholder one and since it is about holding place…

OK, I’ll back up. I think we’re entering a phase where the information we get from our current analytics isn’t really going to help us understand audiences so well. There’s so much more we can learn. For some background, Pretty, Pertinent and Polite is the canon. We’ll have people reading on devices that are wirelessly connected and know their location, but we won’t know where they are or what they’re reading where if we rely on our paltry arsenal of server logs which use IP addresses to guestimate locations. We’ll have a lot of sensors trying to tell us things, but it’ll be like our analytics are wearing earplugs.

We all read at a point in space and time. And we all read something on something, be it a magazine or an iPhone, a web browser or an iPad. Some of these devices know where they are and their orientation and possibly even their angle of tilt and could tell us so we get a better picture of our audience and can think more about them as we design services and try to understand them more. 

So this was the thinking behing placeholdr, some simple service that could tell you something about what people are reading where and on what. The prototype is very simple. I’ll try and work out how to geo restrict one component of it to the Stockholm area and then set it into the wild so that people can play and not totally muddy the data pool. I’ll also work a bit more on making a less localised service.

So how does it work? Well for a hackday I wasn’t going to manage to get Dagens Nyheter to use some calls to an experimental API, that’s all it is; it’s an API for you to post location and what is being read to, well maybe there’s a bit more hidden in there too. What I did was write a simple proxy which in a horribly hacky way fixed all of the image links to be absolute ones with the domain fully qualified and made all of the content and navigation links go through the proxy. Then I added a little status bar at the top and some JavaScript which detects whether the browser is location aware and then posts to the API. There’s also a little status call using the lovely YQL Geo Javascript library from Christian Heilmann.

On the server side, there’s some asynchronous processing of the raw “marks” to yield the likely platform, location details and some information about the content. This allows in the very crude demo to say in which locations people read more about say sport, or culture, or business, or restaurants. Obviously publishers can use this data to inform their thinking about location based interfaces (floating the most often viewed content in a location up higher in a page). The could use it to inform where they buy print advertising (fish where the fish you know will bite are). The completely obvious thing is geo based advertising, but I have to admit I’d like to see a really good example of that, I’m still a sceptic. Most of all it would let you know more about your most precious asset, your readers.

I’ll be fleshing out some richer prototypes in the next week or so and I’ll keep writing up progress here for now before transitioning over to a blog of it’s own in the next few days. If you’re in Stockholm and want to play with what’s there (and give me more data to play with, please do let me know. 

Right, plane about to board at Arlanda, time for adding an outlier, I can’t resist before I’m totally out of range. 

Powered by Tumblr