In a “Project Logo” seminar about radio in cars (June 2015) Nick Piggott said “without metadata we won’t have a better radio experience”. I’m sure we would all nod in agreement – but what are these metadata, where do they come from and why does it matter?
It’s a complex topic, but I find it helps to think about two distinct use cases; “production metadata”, information which is needed to support the business of making, broadcasting and archiving content, and “public metadata”, which helps the public find content they love – YOUR content. Of course there are some data which are important to both production and the audience and this can change during the life-cycle of a piece of content.
The best known public metadata are “now playing” information, which not only scroll across the screen of your FM or DAB radio, but which can also drive your app and build playlists on your web site. “Now Playing” data can in theory drive listeners to your station through search functionality in the UK Radio Player or through third party aggregators, though the value of this is probably over-stated. If I am using a connected device and want to listen to a particular song there are far easier ways of doing it than searching for a radio station that’s likely to play it! Some radio stations make the data clickable, BBC Radio’s “Playlister” is one such example, though you need to think carefully about the benefits and risks of sending your audience to another service where they can hear the same music as on your radio station! Far more important amongst “public metadata”, I would argue, are objects relating to your brand, such as your logo and station description. Nick Piggott made the point that some car radio manufacturers include station logos in the firmware of the radio, which can be very bad news indeed when your station is rebranded. You can watch his presentation on YouTube. Station descriptions matter too; if I type “sport” into the UK radio player search box, I don’t see Absolute Radio; if I type “football” I do see Absolute Radio, but I don’t see Talk Sport. I would argue this is about your brand and matters more than whether or not you are currently playing the latest hit from Taylor Swift.
Speech is Special.
Speech is what stops radio – even “all the hits” radio, from being a jukebox. Think about the last time there was a really great interview on your favourite radio station – how easy would it be for me to find it, by (for example) typing the name of the interviewee into Google? Even if the programme comes up in the first page of results, can I hear the actual interview? Or is there a link to the whole three hour programme and I have to search through it to find the three minutes I want to hear? Much of radio is live, and the radio industry has not historically managed “segments” within programmes well. Our schedules don’t usually get more granular than programmes. If you want listeners to find that interview, somebody at the radio station has to create a clip and put it online with all the appropriate tags, then link to it on social media. This takes more effort than most production teams have available whilst on the air, and if it isn’t done within a few minutes of live broadcast, it is ancient history as far as social media are concerned.
The importance of making speech findable is just as true for speech in between the records on a hit radio station (or at any rate, on a decent hit radio station with interesting DJs) as on the Today Programme on BBC Radio 4. Every time you play a popular hit is the same as every other time it is played by you or any other radio station; when something unique and amazing happens on your radio station, I bet it’s speech, so make it findable.
It’s all about people…
We all know, I think, that it’s a good idea to enter data as early as possible in the production process. It can then be entered just once and re-used by all the downstream systems. The challenge is that the people involved at the start of a production are not generally those who will benefit from complete and accurate metadata downstream, so motivating them to enter it thoroughly is hard. It’s also a culture clash – many very creative people are really not comfortable with structured data, even once they do understand its importance. It’s very easy for technologists to leap to a technology solution, make the fields mandatory, for example. This doesn’t work – creative people will soon figure out that a full stop in the “Contributor Name” field will get them to the next field. You then have a system full of bad data, which is far harder to clean up than missing data. Having drop-down lists of controlled data from which people can pick a value is fine, provided the data set doesn’t change often. For example a fixed list of genres works, but it’s hopeless for “Contributor Name”. Yet “Contributor Name” can’t be a free text field, if it is, you can’t tell one “John Smith” from another. If you want your content to be found, you need high quality metadata and to get that, there is no substitute for winning hearts and minds. I’m very pleased to see that the senior management of BBC Radio have made the team responsible for their metadata system “Proteus” the “Team of the Year”, underlining the importance of metadata in the thinking of the BBC Radio management team.
Who are the people who look after metadata in your radio station? Show them some love.