« New Releases for October 26, 2004 | Main | Blockbuster Q3 2004 Investor Conference Call »


Jim Biancolo

Thanks Mike!

My progress on the distribution center codes is in this report:


There are a few other reports there as well, and I should able to get some more online this week. With enough data, some of them could be interesting. I'm particularly looking forward to having enough data to do "average turnaround time" reports, to see how proximity to your renewal date affects turnaround (if at all, as some claim), and to see how activity in the previous month affects turnaround time (another theory).


I've been doing this sort of thing in a notebook since July. It's rather interesting to look back, trying to spot trends and patterns. I must say, though, it's a bit of a shock to find out there's 2 more distribution centers in my state that I've never heard about before (Bowling Green and Paducah).

Takuya Murata

I have added a table to Netflix entry in Wikipedia. (It's essentially a copy from the site mentioned above) Also, I put a link to this site in external link section.


I'm not clear on whether the turnaround time is really tracking time-to-replacement, though. What's the query measuring? I'm just not sure it has enough information to track time-to-replacement, which is what really affects your viewing schedule. It can measure the time it takes for nf to receive (based on your reported return date), and the time it takes for the user to receive (based on nf's reported ship date), but it's not matching movie A shipped by the user on Monday to movie B received by the user on Thursday. Right? (or am I missing something?) It's still a really, really cool tool, though.


It has a date field for NF shipped, received, returned, NF received.


Right. That would enable it to measure how long shipping takes. But in order to measure time-to-replacement, it would have to assume FIFO -- first in, first out -- that each movie you receive is a replacement of the earliest you sent (using the # of movies in your plan as the total out). But that's not what happens when you have 8 out and you're dropping a movie daily. A Monday return likely gets replaced Thursday, but a Saturday return often takes a week (i.e., the queue gets out of order). In order to measure total turn time (time-to-replacement), the db would have to match movie received w/ movie shipped to replace it. In other words, there would have to be dedicated "slots", which the NF interface doesn't honor -- I've been tracking this manually. Hope this makes more sense.


I think I kinda follow...and since I have a free minute.

I don't think it would necessarily have to match everything up. If Netflix receives two movies, you only need to track the time it takes them to send replacements. If one movie is shipped it doesn't really replace a specific return, it is just in response to an empty queue slot.

If you send 2 movies back and and it takes them 1 day to send the first replacement and 2 days for the second replacement, then you have 3 days total delay. It sounds like you just want to keep track of delays and add them to the shipping times.


As it applies to listology, as an example, it tracks that 1 movie was received on Monday and then another one on Tuesday. Netflix doesn't send a replacement on either day. If Netflix then ships out a replacement disc on Wednesday it looks at the records of the 2 previous returns to NF, assumes it is a replacement for the oldest record (disc NF received on Monday), and calculates the delay-to-shipment based on that.

The comments to this entry are closed.


Third-Party Netflix Sites