Fun With Numbers: Amazon Tracking vs. Sales (11/23-12/6, 2015)

Originally, I was really excited at the prospect of checking holiday sales data and seeing how anime was affected. However, the process turned into a huge headache due to what I can only assume were mistakes in the data tracking – DVD data skipped 2 updates, and that left us with only BD sales data for those weeks. To top it off, some of those BD release datapoints aren’t even usable.

So I was dealing with a confluence of bad circumstances in figuring this stuff out, piecing together what I could. Still, it wasn’t a complete wash. Details after the jump.

For convenience’s sake, I’ll be calling Black Friday week (11/23-29, week 14) BF week, and Cyber Monday week (11/30-12/6, week 15) CM week.

Raw Data:

Week 14 (11/23-11/29):

w14-chart

Week 15 (11/30-12/6):

w15-chart

Charts:

Week 14 (11/23-11/29):

w14-plot

Week 15 (11/30-12/6):

w15-plot

Comments:

Holiday Season Data, Absent the DVD Figures:

As proof of the aforementioned lack of DVD data updates, I’ve taking this screenshot of how Jurassic World was in the database prior to the 12/7-13 week DVD update. Its BD data is from the 12/7-13 week, but the DVD data is from the 11/16-22 week. Jurassic World is my example because it was in the top-20 charts pretty much every relevant week:

jan13-jurassic

So for these two weeks, all we have and will ever have is BD data. We’ll get how the DVD releases performed overall in that 3-week chunk from 11/23-12/13, which I will process later and may or may not be particularly revealing. So we’ll do what we can with the BD figures.

Individually Negative Weeks:

As you can see from the charts above, a number of releases had their totals decrease in at least one of the weeks: Vampire Hunter D (BF+CM), The Cat Returns (BF), and Naruto (CM).

Beyond those, look, that one Cat Returns value for CM week was way positive, but way lower than it should have been, both relative to its normal sales (normally around 40 copies/week) and the expectation. The series ranked top-4000 in amazon, and it wasn’t out of stock. The expectation for that kind slot is like 25 copies *per day*. Even if amazon only makes up 20% of total sales and it sold nowhere else, that should still be a total of 35 copies. The odds of there being some negative value added to it are fairly high, so I choose to exclude it from comparison with expectations. A similar case can be made for Marnie’s performance in the same CM week – it’s way below expectations, and the BDs sold over 1k copies in each previous week of the records. I’m not quite certain about it enough to throw the values into the shredder, though – there’s just enough plausibility there, and it’s a range over which the formula is less than perfect.

Comparison With Expectations:

I’ve built a formula to approximate the average value for amazon ranks in terms of copies sold per day. Here, I’ll apply it to the rank data above to see what the ‘expected’ amount of sales for each release was, vs. what it actually got, and the see how much the expeced values deviated from the actual increase. Those with an actual sales total significantly exceeding expectations presumably got a boost from holiday sales.

To recap, here’s the formula (S=daily sales, R=daily rank):

#10k-#40k: S=10-(R-10000)*(9/30000)
#2k-#10k: S=30-(R-2000)*(20/8000)
#1k-#2k: S=70-(R-1000)*(40/1000)
#500-#1k: S=220-(R-500)*(150/500)
#200-#500: S=250-(R-200)*(30/300)
#150-#200: S=800-(R-150)*(550/50)
#100-#150: S=1000-(R-100)*(200/50)
#50-#100: S=1300-(R-50)*(300/50)

And here’s that data:

Week 14 (11/23-11/29):

w14-deviation

Week 15 (11/30-12/6):

w15-plot

It would be really nice to not have multiple screwy things happening with the data. But we got what we got, so let’s break it down. Excluding the 5 release-weeks under the likely influence of subtraction mentioned above, we’re left with 25 individual weeks of ranks to compare with Black Friday and Cyber Monday sales weeks. Of those, the biggest deviations came from Marnie on Cyber Monday week (significantly undersold), Attack on Titan on Cyber Monday week (oversold), Poppy Hill on Black Friday week (oversold), and DBZ:F in both weeks (oversold). The other 20 were within a factor of two of their expectation value – pretty much in line with how the formula works, ideally.

So, as near as I can tell, only the Attack on Titan and DBZ:F BD releases got BF/CM sales beyond what they normally would have had. Most of the mid-tier anime, even the newer releases, doesn’t really show significant gains. Based on this, it seems like holiday shopping season isn’t quite as huge for anime as it is for the general blockbuster movie market.

The 3-week chunk of holiday DVD data came out as I was struggling over this post. I’ll get to that the next time I have some time to write.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s