Having said some nice things about Last.fm lately, I've been disappointed to find their system down at least twice in the past few days–as I write this, it's been down all evening. Given that all they do is keep track of the songs I play on iTunes, I'm a lot more disappointed than you'd expect. I'm pretty pissed, actually. What gives?
There's an interesting dynamic at work here. When we begin to capture our attention data, like my iTunes playlist, we unlock its potential value. Not monetary value–I'm not getting a dime from Last–but the inherent value in seeing for ourselves and showing others what we're paying attention to. And the great thing about an attention service like Last is that value is created while we're busy going about our ordinary activities, like listening to music. But when the service goes down, and we realize that our attention data is no longer being captured, we feel the absence of the value that's no longer being created. Even though the actual experience is unchanged–I'm listening to Dirty Three right now–the meta-experience is gone.
And when it's a dynamic stream of attention data that can change dramatically in a short period of time–like the music I listen to–then uptime becomes very important, because you never know when a unique burst of attention data is going to come down the pipe. I spent a recent evening listening to a lot of old Naked Raygun. The next morning I realized that Last had been down the previous night, and I was annoyed because data that I assumed was being captured was gone; a meta-experience that I thought I was having didn't actually occur.
Two lessons here: First, if you're an attention service like Last, get your act together. Once users begin to think of your service as something more than a science project (to quote Steve Gillmor), they're going to have a very low tolerance for downtime.
And second, if you're a user who really values your attention, you're probably going to want to roll your own attention service at some point. In an RSS world, Last isn't doing anything that I shouldn't be able to do for myself, and sooner or later some geek is going to show non-techies like me how to turn my iTunes playlist into a feed that I control, without relying on a third party.
tags: Last.fm attention attentiontrust
4 Responses
Ed: Well, the big issue with Last is that they try to offer the service for free indefinitely. That keeps it from scaling well. I happily give them money to allow them to keep operating.
The value of Last.FM is not just in your attention; it’s in aggregating everyone’s attention to pick up trends. Everyone doing this on their own in a distributed nature would reduce the value.
But then I’m a pretty passionate Last.FM user. 🙂
Geof, I agree that aggregation is key to unlocking aspects of attention data’s value. But the choice we face isn’t (or shouldn’t be) between mass anonymous aggregation through a service such as Last and totally distributed data capture and analysis through individual data recorders.
Of course there are many advantages to be gained by sharing data with a large anonymous user base. But I’d much rather do that and also have the option to capture and analyze my own data for myself, and share it with specific individuals of my choosing.
By relying on Last to do it all for us–capture and analyze the data, aggregate the user base, and build services on top of that data–we’re letting them hold us hostage. Bad idea.
And Last’s model has flaws, even when it’s up and running. What if I’d like to share my data with people who aren’t Last users? What if I’d like to analyze and share not only my iTunes playlist but also other types of attention data? Etc.
More fundamentally, what is Last doing to allow me to own and manage MY data? Nothing, as far as I can tell–they’re a black box. That completely violates AttentionTrust’s principles, and it makes me uncomfortable.
I also don’t think it’s sustainable. I should have added a third lesson to my original post: As users come to realize the value of their attention data, they’re going to want more control over that data. And Last is failing on that count right now.
I grant you that they’re not getting it wholly right. You might contact them—I think they’d be amenable to the concept. Their software has always been open source, so I can see AttentionTrust’s principles as something that they’d embrace.
I don’t think there’s anything stopping you from building your own data store for the data you’re shipping them—just not anything that uses the output from the Audioscrobbler family of plugins.
I’ve contacted them twice, hoping to get them involved with AttentionTrust. As you say, it would be a natural fit. Unfortunately, I haven’t heard back from anyone there.
And Geof, you might be able to build your own data store, but I can’t 😉
To meet the needs of ordinary users like me, services like Last are going to have to become a lot more data-sharing-friendly, or they could find themselves losing out to attention data intermediaries like Root Markets.
Full disclosure: Root provides a range of support for AttentionTrust, and Root’s CEO, Seth Goldstein, is a member of AttentionTrust’s Board of Directors.