Misc.: Health care claims automation & Music discovery applications
On health care automation, I say (among other things):
"Automating the processing of health care claims is going to happen."
i-node responds:
"Claims processing is already automated. As is the process of posting remittances to patient accounts."
I have a preteen daughter who doesn't like to brush her hair and I have a conversation with her fairly often that unfolds similarly to this one. I say "you need to brush your hair" and she says "I *DID* brush my hair!" even though her hair is still a mess. Then I tell her "it's not enough to brush your hair a little … you need to get ALL the tangles out. Once you do, brushing your hair again later will be a lot less work."
Similarly, it's not enough to say that health care claims are already automated and leave it at that. If you aren't already in the health care business (I haven't been for 25 years, a time when claims were a lot less automated than they are now, but there was also a lot less documentation required) you can check out some in-depth studies, like this one, and find that there remains a great deal more to do:
National Center For Biotechnology Information | The Healthcare Imperative: Lowering Costs and Improving Outcomes: Workshop Series Summary.
One of the more constructive corporate mantras I have been exposed to over my professional career was "USA", which stands for "Understand. Simplify. Automate." Some entity is going to prioritize this in coming years, at something faster than the speed of bureaucracy, and that entity stands to make a lot of money lubricating what is now a very sticky, slow, and highly manual process.
On curated music discovery applications, I say:
"The key to a good music app isn't to listen to what the record companies want. It's providing avenues to let listeners find music like the stuff they already like, and then the music like that, and then the music like THAT. It's providing easy-to-follow but not necessarily structured paths. It's reading what reviewers (Curators! DJs! Critics! Musicians!) have to say about what the music reminds them of, and then what they have to say about what THAT music reminds them of.
It's to show what inspired the artists you like, and then what inspired THOSE artists. It's not formula, it's the opposite. Because musicians listen to *a lot* of different stuff. It inspires them across predefined genres, which often *are* defined by record labels, but many musicians are not happy being stuffed into any single one of them, though some can sometimes be grateful that their eclectic music can sometimes be classified into a category that has a defined audience. I've heard Randy Newman and Dave Alvin both express gratitude to certain radio programmers, both before and during the Internet age, for deciding that their music fit into a given category and *playing* it, because it otherwise wasn't being heard, and wouldn't be supporting the life they'd chosen to follow. But that doesn't mean they were comfortable in those pigeonholes.
A good music app doesn't just let record companies dictate what listeners should listen to, or buy. It provides links that allow listeners to explore. To LEARN. Apple can build one, but hasn't built good a music exploration capability as they could, or that other companies have. I hope Apple will. They won't need Beats to tell them how, but they should listen when Beats tells them what's needed. Hopefully Beats knows.
Check out Radio Paradise if you haven't already, either from iTunes or its own site You might be surprised what a great DJ can teach."
Doren replies:
"If you do what you did above you will be stuck in genre."
Nate Silver's website posted a very nice riff on the sort of phenomenon that you seem to be inferring from my postings on this subject (FiveThirtyEight.com: Why Classic Rock Isn’t What It Used To Be), but which is the polar opposite of what I have described as being a good music discovery app. Excerpt:
"In fact, radio stations are using data to make their selection decisions. [Eric Wellman, the classic rock brand manager for Clear Channel] said any radio company with the resources conducts regular studies in its major markets to find out what its listeners consider classic rock. And so it’s you, the consumer, who’s helping to define the genre.
“The standard in the industry these days is an online music test or an auditorium music test where you just gather a sample and have them rate songs based on the hooks — the most familiar parts of the song — and you just get back a whole slew of data,” Wellman said. The stations find a cluster of people who like the music that makes up the core of classic rock, and then finds out what else they like. They like R.E.M.? Well, R.E.M. is now classic rock. “It’s really that simple,” Wellman said."
…
But do radio stations rely at all on the institutional knowledge of their DJs to decide what to play?
Nope. The role of the song-picking DJ is dead. “I know there are some stations and some companies where if you change a song it’s a fireable offense,” Wellman said, cavalierly ruining the magic."
The key is the curator, not the algorithm, at least at the moment. In future, the relationships between bands, songs, musicians, etc. might be coded rather than just remembered, potentially allowing an algorithm to make the sort of associations that we take for granted from expert music makers or curators. But this is not low-hanging fruit, and genre music experts are not expensive. Heck, if you're a parent of a teenager, you probably have one in your house that won't cost any more than you're already paying to house/feed/clothe/educate him or her there. |
