What he's saying may be consistent, but I think his conclusion is wrong. I think that the plan directly affects how the account scored (although again there isn't enough data for a strong arguement in either direction). Looking at the final chart it seems that there are 4 tiers that accounts are placed in:
1 very low usage (movie per month ~2.75*limit = poor availability (score 35-45)
They probably still make money on medium accounts, but not much and the ~2.75 is probaly the point at which they start lowing money on an account.
I think authors mistake was in looking at the availability on the A account in the second period. At that point, though it was a low usage account in the previous month, it still had medium availability. He took that as showing that this 5-movie-limit account was being rated the same as a 3-movie account with the same last month usage. I think it more likely means that they average a 2 or more months rather than just look at the last months.
Slashdot Mirror
Step 5) . . .
Step 6) Profit!!
my god, can you tell I'm a newbie or what.
one more try(this time using 'code':
1 very low usage (movie per month < movie out limit) = highest availability (score < 7)
2 low usage (mpm > 2*limit) = high availability (score 7-17)
3 med usage (mpm > 2.75*limit) = med availability (scor 20-30)
4 high usage (mpm > ~2.75*limit) = poor availability (score 35-45)
sorry again
Ack forgot to switch to plain text posting
Here's the 4 tiers:
1 very low usage (movie per month ~2.75*limit) = poor availability (score 35-45)
sorry for the double post
What he's saying may be consistent, but I think his conclusion is wrong. I think that the plan directly affects how the account scored (although again there isn't enough data for a strong arguement in either direction). Looking at the final chart it seems that there are 4 tiers that accounts are placed in: 1 very low usage (movie per month ~2.75*limit = poor availability (score 35-45) They probably still make money on medium accounts, but not much and the ~2.75 is probaly the point at which they start lowing money on an account. I think authors mistake was in looking at the availability on the A account in the second period. At that point, though it was a low usage account in the previous month, it still had medium availability. He took that as showing that this 5-movie-limit account was being rated the same as a 3-movie account with the same last month usage. I think it more likely means that they average a 2 or more months rather than just look at the last months.
That is amazing. Both yahoo and lycos knew that what I really would want to see is that cool intelliCam widget.
Google is just way off base thinking that one might actually be interested in companies with the *name 'widget'.
The beauty about letting advertisers control the search engine is that advertisers know what people *really want to see.