Quote:
Originally posted by Bibliotheque
Except in iTunes you're not. iTunes compares the sales of songs to the sales of the #1 song.
Mediabse updates compare one song's daily airplay to the airplay the same song received seven days ago.
Rolling mediabase updates are caused by a decrease in 7-day airplay, even though the airplay may be increasing on a day-to-day basis. Hence rolling mediabase updates are not a bad thing.
I'm not sure you can assume the same happens in iTunes.
Can total sales go down while hour-to-hour sales go up?
Can this be shown in iTunes updates?
Can we assume either of these, even if they are theoretically possible? We simply don't have enough information.
Even if the answers to these questions are all yeses I think there are too many variables in sales to see the same, almost automatic, rolling effect in iTunes as we do in mediabase  |
I don't think we're on the same page. Each song essentially has a running popularity score which is determined by a rolling window of approximately 24 hours by some algorithm based on primarily sales rate, but also some other secret factors.
That rate can be thought of a single number--not a percentage--which is the popularity of the song. For the first 24 hours, the initial window, this score can only accumulate, but after that each "measurement", or arbitrary period of update determined by iTunes (you can see this because iTunes updates every several minutes, and is not completely real-time), the rolling for the popularity value of a song begins.
The song with the highest popularity value is the #1 song. The popularity percentages iTunes displays are basically the popularity value of any song in relation to the #1 song.
So, it's the popularity value of the song that's rolling, and its percentages change as an extension of that.
The reason I said you're getting compared to yourself is because your popularity value does get calculated with a rolling window and is THEN weighted according to the number 1 song.
There is sufficient evidence to see that iTunes has an initial window of ~24 hours, but much of their algorithm is kept secret.
It's more complex than media base, but still easy to approximate, which is why we can guess sales from it.
And in fact, a rolling window makes perfect sense. Without one there would be massive fluctuations in itunes chart positions because of tiny sample sizes. This method allows songs to reach the high points of the chart only after continuous high sales, and the purpose of having top charts is so that iTunes can sell more copies of popular songs.
