Slashdot Mirror
Evergreens: What The RIAA's Doing Wrong
Chris Johnson writes: "Recently I've been doing some heavy analysis of actual RIAA sales data, from the entire history of Platinum-certified albums. I've worked out a methodology that compensates for the explosion in CD sales and highlights ability to drive sustained sales over years. There's a top 10 list of albums and a top 10 list of the most commercially important artists in history with definite surprises- and the full lists as well, downloadable as text files, with Perl-friendly index numbering, so the analysis can continue with the annoying work already done! Perl folk, go nuts! The actual analysis takes this data and attempts to extrapolate from it and explain the competitive situation the music industry is in relative to Internet music in general, and what they are missing in their assumptions and plans. Should be interesting to see what people make of the Evergreens project! -Chris Johnson" Nice analysis of the history and future of music promotion.
I'll add that I have been kicking myself since halfway through for only noting down the product of those two fields and not keeping the fields themselves. The 'alphabetical' list is literally what I was typing in, record by record- at first because I wanted to get a quick look at the multiplatinum level, and then I was stuck either going back and starting again, or going on... Please do show up my inadequacy as a data analyst by going to http://www.riaa.com/Gold-Intro-2.cfm and making nice downloadable text files suitable for data munging that contain ALL the information you can get from that source, i.e. artist, album, release date, cert date of each certification. I would love it if you did that! Would have saved me days of exhausting and nonprofessional work. I'll further warn you that loads of the information is either missing or flat wrong, and must be corrected through cross-referencing with other sources- _this_ I did, well I thought.
There is no original files- what you see is literally all I have. http://www.riaa.com/Gold-Intro-2.cfm is the source for mining. If you mean it and have the facilities to do a better job at this, may I please, earnestly and in _total_ _sincerity_, beg you to do it? I'm not a data analyst: I'm a sound engineer dabbling in many other fields. If I'm not mistaken my own limited analysis is _the_ _only_ compiled data source of this information out there and publically available. Now that I've taken a shot at it, maybe someone or some open-sourcey group can get some people together and do it right? Nobody would be more pleased than I. Eighteen-hour days of data entry sucked...
Now, Chris doesn't make any real bucks on this. But someone you might have heard a bit more about does... Heather Alexander, a Celtic fiddler/singer/songwriter, actually makes a living at this, both for herself and her hubby/agent. No, they're not rolling in it. But they're doing what they want to be doing, and making a living at it. I would call that success.
The point is, the music industry outside of RIAA is not dead. There are a lot of small labels working outside the box, and bands who get airplay on alternative stations and rackspace in mom and pop music stores... and a growing audience that does NOT listen to Top 40 anymore. When enough people figure out that there's more to life than "Oops, I did it again," RIAA will lose its stranglehold on the business, and the world will change. A very quiet revolution, but I think it's already happening, given the amount of noise going on in the courts....
No, I didn't get a dime for those shameless plugs, I'm just a fan. Deadhead-style band promoting is alive and well.
My job is doing analysis using variations on Kohonen Self-organizing-maps and other unsupervised classifications techniques. We do classification and data-mining all many different types of data. I read this article and immediately thought I might be able to use the data to try and find some of the correlations mentioned in the article. Unfortunately, the data doesn't contain enough information. The only files that I saw had some magical number which is the the sum of each product platinum times the number of years since the album came out. The unsupervised learning algorithms should determine the calculated numbers, not a person. Where are the original numbers?
If the evergreen project wants any serious analysis to be done with this data, then they need to include the raw, initial data. This would be the year of each platinum album, the year of the record release, and any other relevant or irrelevant data. Please re-release these files with the original data so that a thourough analysis can be done.
I'd be happy to test out a few algorithms on the data.
Keeping
- Of the bands/albums listed there as "important", none are all that important any more. The success of Beatles "1" aside, hardly any of those artists have released what I would call a "hit" album in the past 10 years.
- No band will ever "come out of the blue" any more and hit it huge. There is no such thing as a non-manufactured band anymore. U2 is currently in the midst of a sold-out tour, but if they were starting today, it's likely they wouldn't exist. They owe their early success to a core groups of fans and a bunch of radio stations that were willing to play their music. Those stations are now essentially run by the RIAA. See these articles:1,2. Bands today such as The Dave Mathews Band are the rare exceptions.
- The RIAA's business model is focused around promoting a small group of artists through the radio and other media which they mostly control. Having "free" music available over the internet breaks this model. The RIAA companies will no longer be able to promote the artists they want to promote (ie, the ones that they feel give them the greatest chance of profit).
- The RIAA's biggest fear is that artists will be able to promote themselves and leave the RIAA out of the loop entirely. It's only a matter of time before this really starts to happen. Good riddance.
-S
--- What parts of "shall make no law", "shall not be infringed", and "shall not be violated" don't you understand?
The ranking index the author uses to measure long-term selling ability (millions sold x years aon the market) would weight an album that sold 1 million copies 5 years ago (and none in subsequent years) the same in long-term selling power as one that sold 2.5 million copies last year, 2.5 million copies this year, and will go on selling 2.5 million copies every year from now. Whether you agree with the conclusion or not, this is shoddy analysis.