What's-a-meta-you? Hey!

Originally published: October 21st, 2009

Earlier today a selection of "mild-mannered geeks", myself included, had a discussion on Twitter following the removal of data for the Bauer-owned Kiss 100 radio station from One Golden Square (OGS) Labs' Compare My Radio project.

At the time of writing the website still states that Kiss "has either removed their now playing information, or is outputting incorrect data." Taking a closer look at the output of the bot that powers Compare My Radio it is clear to see something went awry from around 10:00 this morning. Around this time the sound of a fun, young London began to bang out the hits including 'Lorem' by Ipsum, 'Nunc' by up-and-coming artist Tincidunt and the hot "flavas" of Eget with their new track 'est'. Anyone with even the briefest amount of knowledge on 1550's typsetting should be able to spot a familiar pattern to these artist/title combinations.

Far be it for me to speculate what's going on here, but there is a possibility the team at Bauer may have casually fed some slightly erroneous data to the Compare My Radio bots. Why would a service provider take these kind of steps? Surely the nature of the technology landscape these days is that of an open and honest culture to create, share, mix and mash up and re-create? This is true, but it can be seen that there is still plenty of value in meta data.

The most obvious value in a radio station's now playing data is that a third party could provide commercial links to download the tracks played from providers like Apple's iTunes Music Store or 7Digital with no return of accumulated commission revenue to the originating broadcaster. It could be argued that the originating station should therefore provide a simple and enticing user experience via their own website to ensure that a user never need venture elsewhere to spend their money. But if the listener is unfamiliar with the service or its website, or a third party website dedicated to download music heard on radio stations has sufficient SEO magic up its sleeve the user may never stumble across the original broadcasters beautifully crafted now playing page and the revenue is lost.

But for that single argument it would be easy to dream up half a dozen new and exciting uses for aggregated now playing data from a selection of radio stations. In its current form I don't personally disagree with what Compare My Radio is doing. Its a clever use of the data and provide a useful and innovative tool.

What may have got the backs up of some broadcasters is the way OGS Labs implemented the technology without first consulting those being indexed about their inclusion in the project. What if the service did change overnight to offer commercial links alongside the tracks? Without any form of agreement in place this could be a risky venture to freely hand over all your data to. I'm not suggesting OGS Labs would ever go down this route, but without any agreement in place it is always worth considering the possibilities.

If a simply defined transport was agreed between the interested parties, bound by terms on how the data is to be used, these innovative projects could and should continue to thrive. I very firmly believe that no matter how hugely talented the in-house development team of any broadcaster be, someone out there could be able to mash up your data in to another application twice as amazing that you would never consider. This is a good thing. But it needs a layer of trust, not unlike the API keys implemented by the likes of Twitter and Flickr to safeguard from abuse the access to the content they host. Outside of this protection I don't see it unreasonable for services to suitably obfuscate their data to reduce the risks of those who don't want to play nice exploiting that information.

The now playing / last 10 tracks page on a radio station's website is always going to be a useful resource contributing to a great user experience. No matter how hard a broadcaster tries to obfuscate its data on the website, someone somewhere will have the determination to mine that data. If you were really crazy you could even sit, pen and paper in hand, crafting your own now playing list to republish outside the "evil clutches" of the originator. But the key to obfuscation isn't to attempt to make it impossible to exploit, it is to make the costs associated with obtaining the data in a usable format far more expensive than those gained from utilising that data.

With simple terms of agreement in place, I believe it absolutely right for open meta data to flow freely between interested parties. It's ensuring the trust between those parties before throwing valuable information out there as a free-for-all.

If nothing else today's discussion shows how important it is that across the industry we talk about these relevant challenges. Despite the number of times the importance of meta data has been presented at conferences and in papers, never before have I seen such a vibrant exchange centered around the openness and exchange of meta data. Hopefully there's lots more to come.

Photo credit: stebulus @ Flickr.

Comments

1. A well thought-out post.

   Where I think it falls over is the thought that the "we've just played" data, which is, by its very nature, in the public domain, is somehow protectable by broadcasters and should be treated as privileged information.

   Radio stations are happy to continually promote 'a better music mix' as a description of their music, but you appear to be arguing that the constituents of this 'music mix' - once played - should be protected and nobody should know what songs a radio station plays.

   The only way that a radio station can adequately protect their "list of songs played" is by stopping broadcasting. Which may be the unintended consequence, of course.

   - James Cridland, October 22nd, 2009 at 12:17

2. The "content" that is a list of a radio station's currently playing and recently played tracks should be open for all to see. I believe it is reasonable to attempt to protect the "data" which can be processed by machine, in order to reduce commercial impact.

   - Andy Buckingham, October 22nd, 2009 at 12:30

3. But ultimately we are in the age of creating our own Spotify playlists - do I as an average end user actually care how many times Kiss played Bulletproof by La Roux??

   Also Andy, you said "last 10 tracks page on a radio station's website is always going to be a useful resource contributing to a great user experience." How? In what way? I can't listen to the last 10 tracks again through any traditional radio website, unlike last.fm which publishes my own 'radio station' for all the world to see and play.

   When are traditional radio stations ever going to catch up?

   - Jamie, October 23rd, 2009 at 11:32

4. Jamie, I think the list of 10 recent songs is extremely useful for people who weren't sure what they were just listening to. Then the user can go ahead and 'consume' that song however they see fit - be that through Spotify, iTunes, Amazon or other ways. Of course, if Absolute provided a list of the songs a station played, with affilliate links, and profited from it, then there would be more to worry about.

   I think the way Compare My Radio presents the 'main' artist data is nice - a few searches have led me to discover that I actually enjoy Absolute Xtreme and Q a lot more than I thought I would. So, if a tool out there makes people re-assess what they listen to, then that's a Good Thing.

   However, what I'm not sure I agree with is showing the number of unique songs played, and the variety. Absolute must have been comfortable with their figures in order to release them, but other stations may very well be less impressed.

   The BBC are seemingly happy to post up the playlists of most shows on several stations. It'd be great to know the reasons why some other stations decide not to put that kind of thing online - as I'm sure most stations get e-mails asking "What did you play last Wednesday, at about 4:30?"

   - Sam Starling, October 24th, 2009 at 00:43

The views expressed here are my own and not those of my employer.