Sunday, 11 August 2013

Spoonfed by robots: What do we get from music streamers... and what do we give?

Step aside, Radio. The web is the hot way to get new music. Or is it? How we relate to music and performers isn't a commodity you can easily exploit. 

2010 Facebook traffic. Look - that's me, Liking The Destroyers
A question: where do you go for new music? I follow tipoffs, or I catch a promising support act; I dig around. I use the web, but I'm not happy.

I rarely get any leads from radio now. After decades listening to people putting themselves before the music, or relegating music to filler status, there’s only a few shows I can stand to listen to. In truth, that's probably how it should be. 

And it's fun following my nose to live and local new stuff, before it becomes product. There are inventive, fresh performances to discover, played to interested audiences. Do this, and you’re a grassroots part of something unique, priceless and intangible, that’s at the very core of the music industry. That's what this post is all about.

Here’s an interesting stat: Sales of single songs are at an all time peak. Sounds good? Actually, no. Units are up, but unit price (through downloads) is down. Most sales go to the really big hits; the rest earn much less. And albums aren't selling, because people don’t download sets of songs. 

When the record industry could fix public demand laughably easily, the quality of its output dipped. Now, the emphasis is on shifting mega units globally by mass marketing. All the stuff on the side - for me, that's the new, experimental and interesting stuff - has to pretty much fend for itself. The web is the main method of delivery. Big radio is, also, off on one side, trying to grab on to the most obvious coat-tails. 

The argument ought to be that because we have the web, we benefit from a greater music diversity. But we don’t, do we? 

A Google server farm - full of YouTube videos 
Start Spotify, everybody's favourite (after YouTube). Millions and millions of tracks, all free or really cheap. You see the Discover page. That's supposed to steer me to new music; I need to check some acts to get started: Daft Punk, or Avicii, or Jay-Z, like that. But these are monster acts. You get a choice of 20. 

So what am I giving, and what am I getting? 

If I fill in my choices from this list of predictable acts, none of whom thrill me to the core, Spotify learns quite a bit about me, and you can bet that's going to show up in the next online pitch they throw at me

There's also stuff that is trending near me. Right now, that’s Rihanna. I really don’t care about this information, either; I’d like something new, please, without jumping through hoops. 

Maybe there's some leads on the Follow page. Look! Spotify can even tell me what that red-hot and influential taste-maker Wayne Rooney might be listening to. Wayne Rooney? I am totally not thrilled. Clearly Spotify are, because he must have thousands of followers, and that's what counts in their world..

Maybe Spotify could let me check what my Facebook friends are listening to? That falls down too. I don’t know why they’re listening. A lot of them are music professionals – they might be doing research. It’s wrong to assume that because someone listens to a song, they approve of that song. 

Further down is Featured Apps, loaded with big names that are, again, not remotely of interest to me. Ironically, they even invite me to look at’s playlists. One digital medium pushing another digital medium. Isn’t this getting a bit incestuous? 

There has to be something I can relate to. Aha – here are all sorts of apps with specialised functions, or which purport to be your essential guide to Jazz/Classical/Indie… or which offer you playlists for all situations. These are guides. Lots of guides, most of them crowd-sourced. Endorsement tools. But do they work like, say, Tom Robinson on 6 music? Or John Peel, back in the day? 

How does this work? Who sets up the endorsement tools? Do all these specialist app businesses hire a lot of staff? I doubt it. I think it’s probably very few, and most of them use data scraped and filtered from sites across the web. That’s simple to set up; the data is cheap or free. But to do this well, you need to dig deep: you need really big data, and that data has to be really well indexed. That’s not working here, judging by the results. 

Last FM – the very name derides old-school radio – is pretty much the same. The front page gives me top plays this week – why, there’s Daft Punk and Robin Thicke. And Kanye West. Gosh, what a surprise. Oh, look, here’s a list of events near me. Except they’re not, they’re a hundred miles away. At least they offer a promise of discovering music, based on my tastes – that’s their USP. But they seem to think I adore Slipknot. Hmmm. Bit old for that.

OK, let’s see if Last FM can throw me some ideas based on bands I like. Let’s go West Midlands local and drop Goodnight Lenin into the mix. Aha! Look, here’s a link for Boat To Row! … and five other artists. Why, that must be because there’s a Folk tag in there somewhere. But BTR are really good at tagging themselves – they’ve got about 20 tags, because they’re diligent types, while nobody’s doing much tag work over at GNL. So, probably a coincidence. Again, not impressed. 

A couple of years ago, Facebook was pretty good. I hooked up with local bands, got useful gig info, and followed interesting leads. But then Facebook started being selective about what to show me. If I choose to follow / like a band or a musician I respect and admire, then I’d like to see their posts. Guess what? That’s just not guaranteed to happen anymore. Instead, I’m sent posts people have paid for. Last time I looked, most local bands are flat broke or in debt, and just can’t afford to do advertising. And I'm certainly not getting the info I used to get even this time last year. 

Let’s go back to that big data thing. Last FM expect contributing musicians to pile vast amounts of search information in; their site visitors are expected to do the same, and give Last FM access to their listening habits. Spotify, as we've seen, also wants to know stuff about you before it will give you anything back. 

And let's not forget those masters of data mining, Google, whose YouTube service I use all the time, including regular embeds in posts on this blog. Google know, better than anyone, where you search, what you look at, and what's in your gmail. And they are about to launch their own music streaming service. If I were Spotify, Last FM, or Pandora, I'd be terrified. 

So you do all the work to build someone else's database. Then their algorithms can swoop in and maybe thrill you with a treasure you don’t know about. That’s a bit rich. 

This reminds me of working with lamentably bad Selector music databases at radio, where nowhere near enough thought had gone into how and why the music mattered. If you don’t do that basic, diligent, geeky groundwork, you get lousy results, even with a tiny music library. Scale that process up into the millions at any web music service, while letting the website try to kid you their algorithms are your personal friends, and you’ve got a recipe for disaster.

Boris waves a flag for The Clash,
but I can't begin to explain why
This is where it falls down. If you’re a music provider, you really should know why something is working before you start waving a flag about it. You can’t crowd-source that knowledge.

That's why it's all so maddeningly familiar. Familiar is where the money is. Familiarity suits mass media. It’s not there for individuals who have decide for themselves.. The more businesses scale things up in their quest for market domination, the more the unfamiliar becomes a risk. The bigger the canvas, the less likely it is that whatever they use to try to detect your tastes, and so reach out to you, will be accurate. 

What is now familiar always starts as unfamiliar. But you can work in new and unfamiliar material without threatening your overall proposition. You can make a big thing about playing new stuff. BBC Introducing and Amazing Radio do that. Both are delivered by people who care about music. The trouble is, the timeslots and positioning scream out ‘specialist’, and this in turn calls for listeners to make a special effort. Appointment to listen radio just doesn’t work that well in the 21st century. The younger your audience, the harder it gets. 

But at least there is a conscious effort to bring new stuff to the listener’s attention. That can backfire, though. The other day, I heard Kiss FM go bonkers with a (very) staged first play of a new song by a familiar artist. They trotted out an apparent deluge of tweets and messages, feedback from the video on their website… all of which amounted to hysterically telling the listener that because everybody else really liked the song, they should too. That smacks of defensive thinking. 

The relationship between the audience and the music is at the heart of what makes the industry tick. It’s often horribly misjudged by mega-corporations. It’s researched to death by brand radio, seeking sure-fire commercial appeal. Artistic merit, however you choose to define it, doesn’t come into the picture. 

Any business that makes money out of music essentially exploits that relationship. Radio needs to know about it. Record companies and promoters need to know about it. All the music streaming sites need to know about it - without it, they’re all dead in the water. 

That vital relationship is at its most intense and pure when the act and the audience know each other in some way. When a band starts to grow, it’s there to see. As an act evolves, it changes. It can be a struggle to keep that relationship strong. Some musicians manage it all their working lives. Others lose it along the way. Others still start strong and then find they can’t change the relationship to match their evolving music. 

My point to this? There is, still, despite everything, one medium that is uniquely placed to jump on that relationship and do well out of it: radio. And the more targeted, localised and specialised it gets, the better it works, because the programmers are closer to the music and their listeners. But radio costs, and specialised radio is having a hard time right now, as I’ve said. 

There are two interesting grounds for faint optimism. This month came news of new technology to enable much lower cost digital radio - just as the big companies are closing down regional digital multiplexes. That means there just may be a chance, in due course, for new and specialised operators to graduate from the web to something more substantial. 

And I continue to have high hopes for committed, passionate internet radio delivered on Smartphones. The latest research shows that listening on Smartphones continues to grow. 
Infographic courtesy of RAJAR/Ipsos MORI/RSMB.

Radio is, still, a place where passions can thrive. The internet is a place where niche interests can reach global audiences cost effectively. But the two have yet to really, really, come together at this level. Those people who understand and share the vital relationship between listener and music will be the one to specialise and properly exploit that relationship – not the big boys. The big boys won’t notice until it’s too late for them to do anything about it. That’s how media works these days. In fact, that’s how it’s always worked. 

So here's to the small-scale mavericks chasing a dream. Here's to the people who are passionate about their music, and who share that passion. 

Me, I’m off out to check another act – invited by a pal who’s a fan.

More music business posts on Radio To Go

Get the Radio To Go Newsletter sent to your inbox every Monday.


Chris B said...

"So you do all the work to build someone else's database. Then their algorithms can swoop in and maybe thrill you with a treasure you don’t know about. That’s a bit rich." This is the most important paragraph in your piece, Robin. There will always be a disconnect between algorithms, current 'data', and 3rd-party interpretations of personal taste because of human error (business managers and programmers, to start), human expectations (marketing departments, to start), and attenuation (basic physics) - the further away energy is from its source, the less substance it has to measure, period. Global business marketing models refuse to acknowledge this last point, which only amplifies those inherent human error/expectation issues. Big data is big and getting bigger. We're lucky we can move it and store it without blowing the power grid, let alone glean something truly personal from a construct so ginormous and inherently noisy...So sick of seeing ads from pre-fab pop idols and gewgaws I'll never click on. I'd much rather hear what YOU have to say about that band your friend recommends, and then make my own choice to seek them out.

David Moore said...

The whole issue of Spotify, streaming services in general & internet radio is interesting.

I use Spotify (and pay for it so I can use it on my phone/tablet) and it's completely changed how I buy music ie I don't. Unless it's a local or small artist and I'm basically making a point of supporting them I don't buy music anymore.

Even though it's excellent on the 'give me music now' front, Spotify's completely useless at pointing me at new music, and with Google doing their 'AllAccess' service I imagine there are tidal waves of panic going through the Spotify offices. With Google's ability to track, analyse and push targeted advertising at people I imagine they'll be a lot better at suggesting new and interesting music to people in a way that makes a lot more sense than Spotify's poor system.

That aside, we host a relatively large internet radio station based in London and we're planning on expanding our web radio work - looking at the stats piles of the listeners are connecting via their phones. I suspect internet radio and local DAB stations are going to do really, really well over the next few years and will need to be very mobile focussed in future.