So, this sale gives us a chance to have an interesting insight into the workings of Gog, because of the published XP leaderboard.

Now, I spent ~£6 and got 14000 XP (including all the badges etc. but missing one check in), so that £6 was worth 9100 XP (I think).

I ranked at 26004, which I presume is quite low on the ranking as probably only the people who bought just one or two very cheap games would come below me (excluding those who didn't buy at all of course and just got the free XP)

So at the very least we can assume Gog made £156,000 (6 * 26004) worth of sales. But clearly a lot of people spent much more than I did. The top spender got 2638130 XP which, if you take out the free XP, means he spent roughly 290 times more than me... or just over £1700 (assuming no regional discount applied).

There's probably a much better way to interpolate and estimate and average for the list, but lets just take the average of my purchase and the 10th name on the list (I think the top few are a little high and probably skew the average as using that gets a total sales of over £22 million...) and multiply it by the number of people on the list:
10th Place had 986220 XP so probably spent around £600 (again, barring any regional discounts)
The average is then £303 (which still seems a little high, unless I'm just really cheap XD), which means sales of about £7.9 million, give or take a little.

Even if we say 25% of buyers were Russian and spent 70% less you're still looking at £6.5 million (ish).

And the reality is probably a little higher (even if we assume the average I've used might still be a little on the high side) as the leaderboard only includes those who opted in and we have no idea how many people didn't.

With a few more data points (and a better mathematician) we can probably get an even better idea.

So assuming Gog takes a 30% cut of that (£1.95 million) they've done pretty well in the last few weeks I'd say.

And in conclusion, I'm clearly in the wrong business XD
I wonder how difficult it is to open an online games store....

Hm, I was somewhere around the 21000th place in the ranking list, with about 21000 XP. I spent about 13 Euros, so we can assume that 21000 people spent more than those 13 Euros per person.

You forgot something important: the leaderboards were opt-in. I never did, so I don't think my purchases even show up there (feel free to check if they did, I ended up with around 45000 XP). That means there's no real way of knowing how many more sales there were beyond those on the leaderboards.

Oh of course, but I was just looking at a kind of bare minimum (though I think max's added data point suggests the average is much lower, it would be interesting to get someone who had spent ~£100 to see where they fell on the list) to see what kind of numbers Gog is selling in and how many copies of games they're shifting these days.

I'm fairly certain those who didn't opt in weren't included at all (how could they, if the highest ranking person didn't opt in then there would just be a blank spot at the top of the leaderboard) so we're probably only looking at half the picture...

Yeah, with some more people sharing their spendings, XP and rank on the list, I guess we could paint a realistic picture of the amount that was spent in the whole sale by all the people here.
Taking into account, that I surely was not on the last place and that there are a lotta people who didn`t opt-in for the leaderbord, I guess it`s safe to say, that there are at least about 100,000 people who spent at least 10 € or the local currency aequivalent.

half a curly wurly and an apple

Rofl, my first thought when I saw the thread was https://www.youtube.com/watch?v=l91ISfcuzDw

Maybe I wasn't that far off xD

Calculating with averages is the wrong way of estimating their sales, because averages aren't accurate. It would be better to calculate it with ranks:

(- The amounts of XP mentioned are paid XP, meaning that I removed 5,000 XP from the real amounts.
- I'm assuming that 1,000 XP = $1 for simplicity's sake.
- The amounts are fairly recent, no more than five hours old.
- Of course, the numbers are rounded down to make it easier, and to get a floored approximate.
- I chose not to link to the posts I used as references because some might not want to appear directly in this list.)

26,000 - 22,000 users who amassed more than 9,000 XP but less than 16,000 XP = $36,000
22,000 - 16,500 users who amassed more than 16,000 XP but less than 27,000 XP = $88,000
16,500 - 2,200 users who amassed more than 27,000 XP but less than 92,000 XP = $386,100
2,200 - 2,050 users who amassed more than 92,000 XP but less than 95,000 XP = $13,800
2,050 - 1,630 users who amassed more than 95,000 XP but less than 107,000 XP = $39,900
1,630 - 340 users who amassed more than 107,000 XP but less than 230,000 XP = $138,030
340 - 100 users who amassed more than 230,000 XP but less than 370,000 XP = $55,200
100 - 45 users who amassed more than 370,000 XP but less than 531,000 XP = $20,350
45 - 10 users who amassed more than 531,000 XP but less than 981,000 XP = $18,585
And the ten biggest spenders = $16,705

All that totals to $812,670. If we assume as you assumed that 25% of buyers were Russian and spent 70% less, that totals to $812,670 * 0.75 + $812,670 * 0.30 * 0.25 = $670,452.75.

Finally, by assuming GOG takes 30% of the total revenue, it amounts to $201,136.

There is uncertainty about whether not users who chose to not opt-in are taken into account but, in any case, the real amount should be greater than this one. There is also the annoyance of currency differences and Fair Price package, but I don't think it's realistically calculable, so let's hope it doesn't really matter.

It would greatly and positively impact the total if we had access to more intermediate amounts, like between ranks 22,000 and 2,200 and between ranks 1,630 and 100. For example: assuming that the 10,000th rank has more than 50,000 XP, it would replace the $316,800 spent by users between rank 22,000 and 2,200 by $192,000 + $390,000 = $582,000, almost double the original amount.

Edit. Fixed two minor typos.
Edit. Added one level of data.
Edit. Added another level of data.

I earned around 100 000 XP, while spending a little less than 90 euros. No idea at what place I was on the leaderboard. Is there still a way to check that ?

EDIT : OK nevermind, I saw the list here ^^^

I believe that should be 1,000 XP = $1 :) Let's say you bought Pillars of Eternity Hero Edition for $14.99. The XP you earn for it would be 15,000, not 15.

I've got a couple of suggestions here if you actually want to have a more accurate answer: I haven't opted-in, and I have spent more money than you (owning way less games here, I guess :-D). I suppose I can still opt-in because the big button is still there in the page, and you should go down in the ranking. It should be at least 26005. That way we can be certain that those who didn't opt in weren't included at all.

Actually, maybe you can confirm it already if your position has changed since you wrote your OP. If there's a proper mathematician in the forum that would be great, but I think that given enough data points we can do it quite easily. My calculus is a bit rusty, but I think we only need to plot the graph and find the area? The integral, that's it.

It can be done very easily with numerical methods. It's pretty much what Tyrrhia is proposing but without using arbitrary ranks. You can just use every data point for greater accuracy.

Edit: Typos.

No, no, that was the simplicity I was talking about . . . *cough cough* Yeah, simple mistake. Fortunately, it didn't affect the calculations. :P

What do you mean by arbitrary ranks? I mean, they were all taken from posts within this thread and the official summer sale thread. How can you calculate the sales without data?

Or maybe I'm misunderstanding what you're saying. It's been a long while since my math days, and even then, I learned them in French, I don't know the English terms, except for the super easy stuff. :(

You can say that again... your answer is about 10% of mine...

So if we even assume half of people opted in and half didn't then you're looking at about $1 million (queue more links to Austin Powers) based on that estimate, though it's hard to know how accurate that it is.

That's fairly respectable sales.
Yeah, my rank has dropped to 26067, so the opt outs definitely weren't included (or a lot of people are only now claiming the bonus exp for stuff, assuming that is still possible)

I can certainly see ploting a graph helping to find missing values but I can't say I know enough about that kind of maths to know how to use it to figure out a total value for the list...

Oh no, my bad. I missed the part where you said that "I chose not to link to the posts I used as references", meaning you actually got the data from different posts, and I thought it was just a fictional example you were using to illustrate your method. Sorry.

Still, I'm a bit confused. Are your XP intervals right or I'm reading it wrong again? For example, in the first interval, 26,000 - 22,000 users, the final XP shouldn't be 16,000, which is the start of the next interval? Not that it matters much, because I see you use the same floor value (the left one) for a whole interval, but just to agree on everything.

Also, it might be that I'm the one using the completely wrong English terms, so don't think it's your fault if there's a misunderstanding :-D

You don't really need any advanced math. It can be done easily in a spreadsheet if you know your data points: http://www.statisticshowto.com/how-to-find-the-area-under-a-curve-in-microsoft-excel/ (Could be done as easily in Google Sheets and it would be online.)

Anyway I'm not saying "let's go ask everybody and find the exact figure". That's probably a lot of work and probably you're not that curious? I'm just saying that it could be done if somebody is really curious.

In the end it's what Thyrria has shown, but automated in case there are lots of data points, and a bit more accurate because it uses an average value for each interval instead of the lowest one.

I meant no disrespect. It's just that an average means a potentially super wrong, potentially accurate estimate.

For example: Let's take this situation where five people spent a certain amount: $100, $0, $0, $0, $0. If you make an average of the two extreme values, you will get $50. So, by only using the average, it would mean that the total spent was $250. That's way less than the actual amount, $100.

Of course, my method isn't accurate, either. But it gives a minimum value: it shouldn't drop below that amount, but it could be way higher. I chose this method to not get a high amount in which we could be disappointed once we know the actual amount (if we ever do).