What horrendously interesting data as usual, thanks Shadow.

I will give this a proper study tomorrow!

Why, it's right here!:

So.

Hopefully there will be another one of those hils that suddenly go up and there is a slope going down.

:*(

Too sad. WHHHHYYYYY!!!! D':

Dat Vita doe.

True, but to be honest, that's kind of the low-hanging fruit.  I guess what I'm trying to wrap my head around is whether it's possible to use existing shipment data to piece together the likely distribution of sales.  Like, if you look at a system over its lifetime, you have a series of known data points - shipments.  Now, if you take into account things like their frequency and size, is it possible to infer what came between those data points?

It would definitely help if we knew where manufacturers set the threshold for shipments, but then I suppose this, itself, would depend on sales.  If there are a million units stagnating on shelves, for example, then they're hardly going to ship a million more.  If, on the other hand, shops are selling out left and right, then those million units suddenly seem very insignificant, and a fresh shipment would be bustled out the door as quickly as possible.

Now, the number of units shipped gives us our upper limit on sales, but couldn't things like the timing, frequency, and relative size of shipments also be used to infer a lower limit?  We can't determine the fate of any given shipment in isolation but surely it can tell us something about the last shipment, can it not?

To use a really obvious example, let's say that three million units are shipped, and then a month later, another three million are shipped.  I think it would be fairly safe to deduce that most of the former shipment has sold during that month.  Again, exactly how many would depend on a number of factors.  If sales were consistently high and showed no sign of slowing down then you might argue that they got in early before stocks could drop to dangerous levels.

On the other hand, maybe there was a sudden spike in sales.  Perhaps they had only sold a few hundred thousand and then suddenly (due to a sale, popular game release, etc), two million or so flew off the shelves just like that.  In that case, perhaps they were biding their time and, getting caught off-guard by the rapid reduction of stock, hurriedly pushed out another shipment of units to meet the sudden rise in demand.

So, you could certainly argue that the distribution of sales would be influenced by a number of variables, but the one thing I think you could be sure about in the above scenario is that the first shipment sold massively.  It wouldn't make any sense to ship an additional three million units after just a month if the first three million had only sold a few hundred thousand, or perhaps even a million.

I suppose the question is where do they set the bar?  Unless sales are absolutely booming, I can't imagine manufacturers making additional shipments until they're at least inside their last million units.  And during slow sales periods, I would think they'd wait even longer, like under 500k, or maybe even just 200k - 300k.

I admit that these are just guesses on my part, but I don't see why they would make more shipments unless units were really moving.  The only exception I can imagine is if they were to make a big shipment in anticipation of something, like a massive release or holiday sales.  Still, I'd be interested to hear if there is any actual info on how shipments relate to sales.  Or are we pretty much left guessing on this one?

Dude, let us know how that works out!  I'd be very interested to see it!