• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
  • Hey Guest. Check out the NeoGAF 2.2 Update Thread for details on our new Giphy integration and other new features.

SteamSpy - Approximate LTD sales for every game on Steam (Updated Daily)

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
Folks, decision time...
Would you be interested in actual weekly saleownership data for all Steam games? Like, top-25 selling games thread with additional data provided online?
I think I found a relatively hassle-free way of getting that data off of SteamSpy, are people interested in that though?
Also, galyonkin, would you be okay with having said data posted on these forums?
I'm absolutely ok with it :)

I'd advise to use monthly or bi-weekly sales though, as weekly changes in owners might be not as reliable for smaller games.
 

Knurek

Member
May 15, 2013
11,532
2
0
I'm absolutely ok with it :)

I'd advise to use monthly or bi-weekly sales though, as weekly changes in owners might be not as reliable for smaller games.

Could you add an option to sort entries by SteamID (with additional field for that data, of course)?
I've just tried my method, and unfortunately, it doesn't seem results sorted by release date will suffice - I'm getting some entries switched around and new ones added, even for 2014 released games, where no change should happen...
For example, data from 2014 released games, sorted by release date and then by game name and pasted into spreadsheet.

Here are the results scraped from 06.08.2015:

And here the results from today:


As you can see, rows 23 to 31 are different, which makes matching the updated results much harder than I expected. :\
 

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
Could you add an option to sort entries by SteamID (with additional field for that data, of course)?
I've just tried my method, and unfortunately, it doesn't seem results sorted by release date will suffice - I'm getting some entries switched around and new ones added, even for 2014 released games, where no change should happen...
For example, data from 2014 released games, sorted by release date and then by game name and pasted into spreadsheet.


As you can see, rows 23 to 31 are different, which makes matching the updated results much harder than I expected. :\

Are you doing it manually? I'm afraid some games are going to be on and off because my site only displays games that had enough data points in the last three days. For smaller games it means they might appear "blinking".

Oh, well, I have an idea, give me one minute :)
 

Knurek

Member
May 15, 2013
11,532
2
0
Ok, here is a new stat page. It requires some URL fiddling for now, but I think you can manage it. Also, you need to be logged in.

Works like this /sale/FROM/TO where TO is optional

for example:

http://steamspy.com/sale/2015-04-01/ - sales from 1st of April to current date
http://steamspy.com/sale/2015-05-01/2015-05-31/ - sales for May.

Neat, thank you.
Any particular reason it doesn't show sale data for our Indie darling, Rocket League, no matter the sale duration used?
 

ashecitism

Member
Jun 29, 2013
17,989
22
0
Hungary

Knurek

Member
May 15, 2013
11,532
2
0
It was a bug, fixed it :)

Nice, will try to do a thread in main tomorrow.
Another thing I've noticed - SteamSpy doesn't seem to handle FreeToKeep games correctly.
Notice Race The Sun. The game had a three day period where you could add it to your library for free and keep it - Steamspy doesn't show any price decrease for that duration.
I'm not sure if you can get that data from Steam API, but such situations have been so far infrequent enough so that a manual fix whenever that happens might not take too much of your time.

Well, people now own this game, so "owners" data is right. The only problem is pricing and discounts, right? I can set the price to 0 for the weekend in question.

Oh yes, of course, sorry if I wasn't precise.
 

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
Nice, will try to do a thread in main tomorrow.
Another thing I've noticed - SteamSpy doesn't seem to handle FreeToKeep games correctly.
Notice Race The Sun. The game had a three day period where you could add it to your library for free and keep it - Steamspy doesn't show any price decrease for that duration.
I'm not sure if you can get that data from Steam API, but such situations have been so far infrequent enough so that a manual fix whenever that happens might not take too much of your time.
Well, people now own this game, so "owners" data is right. The only problem is pricing and discounts, right? I can set the price to 0 for the weekend in question.
 
Nov 27, 2013
1,932
0
0
So Witcher 3 has sold 724K copies on PC? Or is that just for steam and it gets sold other places too?

As doesnt seem that high considering the amount of steam users.

edit..just seen its since april 1st
 

Nzyme32

Member
May 23, 2013
18,286
1
0
So Witcher 3 has sold 724K copies on PC? Or is that just for steam and it gets sold other places too?

As doesnt seem that high considering the amount of steam users.

edit..just seen its since april 1st

Witcher 3 sold more in it's first month on GOG than on Steam. All copies outside of Steam, both digital retailers and physical are for GOG. The number of 724K from Steam means GOG is likely to be more than that. To date, you are going to be looking at well over 1.5 Million PC sales.

The same is true for GTA V - all copies outside of Steam (digital and physical) are for Rockstar Social Club. Again the Steam numbers are not representing the full scale of the PC sales and could well be more than double what SteamSpy can report.
 

Shantom

Member
Oct 20, 2011
2,030
0
0
Witcher 3 sold more in it's first month on GOG than on Steam. All copies outside of Steam, both digital retailers and physical are for GOG.

The same is true for GTA V - all copies outside of Steam (digital and physical) are for Rockstar Social Club. Again the Steam numbers are not representing the full scale of the PC sales and could well be more than double what SteamSpy can report.

I would have thought a big chunk of the GOG copies would have been from GPU promotions and retail copies. The Nvidia deal is over, and retail copies have less legs on PC than digital, so I wouldn't be surprised if Steam has caught up to the GOG numbers. Edit: However GMG have had some pretty aggressive deals on the GOG version recently, so it's entirely possible that version is still a fair bit ahead.

As for GTAV, we know it sold about 2.8 million on Steam, and Rockstar said today it's now over 54 million across all platforms, which isn't that much higher than the last number they gave, so it's possible Steam makes up a very large chunk of the sales for it on PC.
 

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
A quick note.

I've fixed some bugs in sales calculator.

It wasn't calculating variance properly and used 1-day samples instead of 3-days samples as everywhere else.
 

Qassim

Member
Jul 6, 2012
6,854
2
590
United Kingdom
qassim.uk
As for GTAV, we know it sold about 2.8 million on Steam, and Rockstar said today it's now over 54 million across all platforms, which isn't that much higher than the last number they gave, so it's possible Steam makes up a very large chunk of the sales for it on PC.

Timeline of GTAV recent sales:

May 2014 - 33 million
November 2014 - PS4/XBO version releases
February 2015 - 45 million
April 2015 - PC version releases
May 2015 - 52 million
August 2015 - 54 million

I think it'd be a very reasonable assumption that the PC version has sold at least 5 million of the 9 million copies sold since February, as that was the only new version released of it in that time frame. But personally, I would be far from surprised if the non-Steam version had sold more than the Steam version. The Steam version is full priced and hasn't had any real sales, it was often the most expensive option to buy the game. Buying the game physically and from other key sellers was a cheaper option, sometimes by a significant margin.
 

alexandros

Banned
Jan 30, 2012
7,045
0
0
Timeline of GTAV recent sales:

May 2014 - 33 million
November 2014 - PS4/XBO version releases
February 2015 - 45 million
April 2015 - PC version releases
May 2015 - 52 million
August 2015 - 54 million

I think it'd be a very reasonable assumption that the PC version has sold at least 5 million of the 9 million copies sold since February, as that was the only new version released of it in that time frame.

Yeah, I would expect PC sales to be the majority of those 9 million sales.
 

CertifiedFP

Member
Mar 25, 2013
5,385
2
470
Can't get data on Type-0 HD, but I can for the rest of the series. Did SE request Type-0 data be hidden? Just asking since I can get preorder data on other games like Tales of Zestiria.
 

Shantom

Member
Oct 20, 2011
2,030
0
0
Timeline of GTAV recent sales:

May 2014 - 33 million
November 2014 - PS4/XBO version releases
February 2015 - 45 million
April 2015 - PC version releases
May 2015 - 52 million
August 2015 - 54 million

I think it'd be a very reasonable assumption that the PC version has sold at least 5 million of the 9 million copies sold since February, as that was the only new version released of it in that time frame. But personally, I would be far from surprised if the non-Steam version had sold more than the Steam version. The Steam version is full priced and hasn't had any real sales, it was often the most expensive option to buy the game. Buying the game physically and from other key sellers was a cheaper option, sometimes by a significant margin.
Sorry for the late reply to this, but you're right. I was remembering the February figure as being 50 million for some reason.
 

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
I'm surprised to hear this because when I preorder a game it's added to my Steam Library. I assume it has some sort of different flag which prevents it from showing up?
Some games are different, but most games don't show up in Steam WebAPI on user accounts until they're released :(
 

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
I have a question for everyone here.

I've been thinking about making some extra money off SteamSpy as some people suggested on Twitter.

I'm thinking about introducing paid subscription to SteamSpy.

The basic stuff will be free, but paid accounts will give you more data and you'll be able to see stats for longer periods - not 60 days like it's now, but for a full year. Although I don't have stats for a full year yet.

Comparable AppAnnie subscription starts at $599 per month and it is worth it if you're developing big f2p titles. AppAnnie also has a cheaper option, where you can pay $59 per month to see limited data for one (yes, one) game.

So, how much would you be willing to pay for a monthly subscription for SteamSpy? Provided that it would still have basic free version similar to what it has now.
 

Shengar

Member
Jun 27, 2013
13,101
0
0
I have a question for everyone here.

I've been thinking about making some extra money off SteamSpy as some people suggested on Twitter.

I'm thinking about introducing paid subscription to SteamSpy.

The basic stuff will be free, but paid accounts will give you more data and you'll be able to see stats for longer periods - not 60 days like it's now, but for a full year. Although I don't have stats for a full year yet.

Comparable AppAnnie subscription starts at $599 per month and it is worth it if you're developing big f2p titles. AppAnnie also has a cheaper option, where you can pay $59 per month to see limited data for one (yes, one) game.

So, how much would you be willing to pay for a monthly subscription for SteamSpy? Provided that it would still have basic free version similar to what it has now.

So you'll restrict geography spread as a feature for paid user? Well, that sucks but understandable and I can accept it somewhat. But I wished to see Celestian Tales: Old North distribution first at least :(

On your question, with the amount of data wealth that Steamspy can generate, a sum of $599 seems modest. But considering there are many indie devs around the world who have different standard living cost, that amount might be very prohibitive while at the same time they clearly need it to make their games better. In my opinion, you should have a flexible or multiple price point offering different set of option each.
 

CertifiedFP

Member
Mar 25, 2013
5,385
2
470
You also have to remember that AppAnnie covers a market infinitely larger than the confines of Steam as far as userbase and capital being moved around goes, so $599 is a lot more bang for one's buck there than it would be with just Steam stats. IMO $599/m for only Steam data would be like AppAnnie being in the $10k/m (NPD subscription price IIRC) despite only covering App Store/Google Play data.*

Perhaps somewhere in the $50-150 range? Seems right relative to the size of the market it's analyzing. $599/m is way beyond the means of the majority of indie devs who make their living on Steam.

*Though I don't know what their Enterprise-tier pricing is though.
 

dude

dude
May 18, 2007
6,988
1
1,290
Tel Aviv, Israel
So, how much would you be willing to pay for a monthly subscription for SteamSpy? Provided that it would still have basic free version similar to what it has now.

I think about 50$-100$ seems about right. The data is valuable, and I think it's very good to charge for it. As a developer, that wouldn't be too bad to pay monthly. Anything north of that will be a problem with smaller indies, which makes a large chunk of Steam (and thus your potential customer base.)
 

Saty

Member
Sep 4, 2009
7,403
0
0
So, how much would you be willing to pay for a monthly subscription for SteamSpy? Provided that it would still have basic free version similar to what it has now.

Depends what will be free and what will require a sub, but honestly i don't like the idea.

The big deal with Steamspy is that everybody can see and study the data. Information isn't locked behind hundreds and thousands $ worth of subscriptions that only publishers can afford. Having had this breakthrough i don't want to get back to the 'dark ages' of sales data as it is now in the console industry and even worse.

Also, what's the 'justification' of charging money? This data is freely available to anyone who scrapes account data. While all the graphs and distributions are great, it's still just a presentation of the same data that is available to anyone. Isn't the playtime data more easily available than owner data that requires scraping and extrapolation?

Moreover, wouldn't this kind of move to make money of data that is technically tied to Steam make Valve reconsider it's approach of silent acceptance to Steamspy? Wouldn't publishers call for Valve to do something about it now that Steamspy charges a subscription?

Additionally, what stops a subscriber sharing info with free users? 'Here's a screenshot of the 1 year history of this game'. Either very few people will subscribe or somehow you'll need to police this kind of activity not unlike 'NPD ninjas', which in my mind goes against the spirit of SteamGauge \ Steamspy services.

And ultimately, what stops someone else mining the data on his own and to put up a new site with the same data SteamSpy will offer to subs, free of charge?

I understand that server and database necessities might require Steamspy to be able to financially sustain itself but i think a subscription plan should be the last option. Dunno if the GI.Biz 'sponsorship' is bringing you any money but maybe you should look into these kind of opportunities and ads generally before you go down the subscription path. Also, you could add a 'donation' section on the site.
 

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
So you'll restrict geography spread as a feature for paid user? Well, that sucks but understandable and I can accept it somewhat. But I wished to see Celestian Tales: Old North distribution first at least :(

On your question, with the amount of data wealth that Steamspy can generate, a sum of $599 seems modest. But considering there are many indie devs around the world who have different standard living cost, that amount might be very prohibitive while at the same time they clearly need it to make their games better. In my opinion, you should have a flexible or multiple price point offering different set of option each.

I will probably restrict resource-hogging stats. I feel it's more fair than arbitrary restrictions :)

Also, I don't intend to charge $599 per month, the market is vastly different. It was just to give you an idea how much does services like this cost :)
 

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
The big deal with Steamspy is that everybody can see and study the data. Information isn't locked behind hundreds and thousands $ worth of subscriptions that only publishers can afford. Having had this breakthrough i don't want to get back to the 'dark ages' of sales data as it is now in the console industry and even worse.
Not my intention. I want to provide something similar to what I provide now for free. But people keep asking me for more advanced stuff that requires both my time and processing power.
Also, what's the 'justification' of charging money? This data is freely available to anyone who scrapes account data. While all the graphs and distributions are great, it's still just a presentation of the same data that is available to anyone. Isn't the playtime data more easily available than owner data that requires scraping and extrapolation?
Well, I'm planning to restrict graphs that require a lot of processing resources. Calculating owners is easier and faster than calculating playtime distribution :(
Moreover, wouldn't this kind of move to make money of data that is technically tied to Steam make Valve reconsider it's approach of silent acceptance to Steamspy? Wouldn't publishers call for Valve to do something about it now that Steamspy charges a subscription?
It's possible, yes. But does it change anything to publishers whether service is paid or not?
Additionally, what stops a subscriber sharing info with free users? 'Here's a screenshot of the 1 year history of this game'. Either very few people will subscribe or somehow you'll need to police this kind of activity not unlike 'NPD ninjas', which in my mind goes against the spirit of SteamGauge \ Steamspy services.
Nothing.

I don't mind people sharing the information. It would be kinda like piracy and I don't mind piracy as long as service pays for itself :)
And ultimately, what stops someone else mining the data on his own and to put up a new site with the same data SteamSpy will offer to subs, free of charge?
Nothing, of course. I even have an API to make it easier.
I understand that server and database necessities might require Steamspy to be able to financially sustain itself but i think a subscription plan should be the last option. Dunno if the GI.Biz 'sponsorship' is bringing you any money but maybe you should look into these kind of opportunities and ads generally before you go down the subscription path. Also, you could add a 'donation' section on the site.
GI.biz isn't sponsoring the site, they're selling ads :) And they haven't sold any yet :)

I don't believe in donations for a service. Donations are good for a single cause, not for something that has to be sustainable.

Patreon is an interesting idea, but it's the same as subscription - you pay monthly and you get some benefits from it, when everyone else gets a bit simpler version free.
 

Alo81

Low Poly Gynecologist
Jun 24, 2012
13,143
1
710
steamcommunity.com
I see why you'd want a subscription for access to the more resource intensive statistics.

Those prices listed sound absurd, but also I'm just a single person who likes knowing stats on games for personal knowledge so I'm definitely not the target for it.

I mean, personally I'd prefer everything to be free for me forever with everything, but that's obviously not realistic so do what you gotta do to sustain yourself. If I or other users can't get the info we want any longer, we're free to either suck it up and move on, pay, or try to create our own alternative.

I'd hope that the basic statistics would stay freely accessible, because I would be bummed out to lose sales numbers on games.
 

wrowa

Member
Jul 26, 2006
14,351
1
1,170
Germany
I feel like introducing paid subscriptions might be a can of worms you'd rather not open. You are using publically available APIs in order to calculate your data, and I'm not sure if Valve and co would be okay with a commercial use of such data.
 

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
I feel like introducing paid subscriptions might be a can of worms you'd rather not open. You are using publically available APIs in order to calculate your data, and I'm not sure if Valve and co would be okay with a commercial use of such data.
I've asked them, but they're silent as usual.

It's funny, because they do reply to work-related emails :)
 

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
I'd hope that the basic statistics would stay freely accessible, because I would be bummed out to lose sales numbers on games.

Of course. I have no plans becoming next AppAnnie or anything. Basic stats will be free. But it's either I don't add advanced stats and graphs at all, or make them paid.

For example, I have graphs that show trends for tags, so you can see, for example, if interest in sci-fi is rising or falling among Steam userbase. It's useful, but it takes up to 60 seconds per tag to calculate, so these graphs aren't available to anyone but me now. I could let paid users access these graphs, because a) there won't be many of them b) I could afford a better server then.
 

dude

dude
May 18, 2007
6,988
1
1,290
Tel Aviv, Israel
I have to say I rather have a service like this be paid than operate on ads. It helps it be more stable, introduce more features etc. As long as the pricing is fair, it is reasonable.
 

Nzyme32

Member
May 23, 2013
18,286
1
0
Of course. I have no plans becoming next AppAnnie or anything. Basic stats will be free. But it's either I don't add advanced stats and graphs at all, or make them paid.

For example, I have graphs that show trends for tags, so you can see, for example, if interest in sci-fi is rising or falling among Steam userbase. It's useful, but it takes up to 60 seconds per tag to calculate, so these graphs aren't available to anyone but me now. I could let paid users access these graphs, because a) there won't be many of them b) I could afford a better server then.

I don't see any problem with this. I'm certain that the paid service will be extremely valuable for those that need more detailed and longer term statistics.
 

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
Can't view the page without an account, so I'm curious, what kind of pricing did you settle on?

I decided to go for tiered pricing.

9,99 (107,99 per year), 19,99 (199,99 per year) and 49,99 (399.99 per year). That's before any discounts

The subscriptions are for people that would like to support Steam Spy (the cheapest option), indies (the middle one) and big developers or publishers (the most expensive one).

So far I've seen several subscribers choose the middle option and one choose the most expensive.

Here is a screenshot. I'm not really good at design, sorry :)
 

Saty

Member
Sep 4, 2009
7,403
0
0
Meh.

This is a tough sell. I mean, Enthusiast is x2 the price of Supporter but the only extra features are tag graphs and additional 9 months for all data graphs. Same goes for the ramp up to Pro.

Will new features only be added to Pro members? Are all of them going to be server-demanding? If so that's disappointing. I was hoping free users will get those and the more advanced usages of the feature will be for subscribers. If you look at all the features you added to Steamspy since April, it would have been pretty lousy if they were only to Pro subscribers if you had subs since the beginning.

It's also a shame that my most wanted feature (data all the way back Steamspy's launch) is behind the Pro subscription. That + Geography distribution in my eyes is the most important for basic\casual analysis. Having that one LTD graph stretching from launch to now and checking where the game had discounts, featured, bundled, picked up be a streamer, etc. and how it did or didn't affect sales.

Anyhow, i hope it works for you and Steamspy gets better and better. Unfortunately, i don't think i'm going to sign up in the near future and i might have to look up 'leaked' data.
Dunno, maybe as a group of GAF-ers we can get together and subscribe for a shared Pro account.
 

galyonkin

Actual Russian Spy.
Apr 3, 2015
205
0
0
galyonkin.com
This is a tough sell. I mean, Enthusiast is x2 the price of Supporter but the only extra features are tag graphs and additional 9 months for all data graphs. Same goes for the ramp up to Pro.
Well, it depends on what you consider to be valuable. I think Tag graphs for extended period are extremely valuable for developers (I'm using them at my day job), but everyone is different.

If you're happy with Supporter of free account - that's great ;)

Pro version is a bit underpowered now, but I'm planning to add cross-game analysis - that's extremely resource-hogging feature that I haven't even tried to add before.
Will new features only be added to Pro members? Are all of them going to be server-demanding? If so that's disappointing. I was hoping free users will get those and the more advanced usages of the feature will be for subscribers. If you look at all the features you added to Steamspy since April, it would have been pretty lousy if they were only to Pro subscribers if you had subs since the beginning.

No. I'll add features as I see fit, as before :) The next one that's coming this week is Twitch tracking for Steam games and it's going to be free.

I will limit features primarily based on how much resources they require. Twitch scan for example is easy on my processor, while playtime distribution is not.

It's also a shame that my most wanted feature (data all the way back Steamspy's launch) is behind the Pro subscription. That + Geography distribution in my eyes is the most important for basic\casual analysis. Having that one LTD graph stretching from launch to now and checking where the game had discounts, featured, bundled, picked up be a streamer, etc. and how it did or didn't affect sales.
Well, some data, like global Steam stats is free all the way back to April. But individual games graphs will be locked, yes.

Anyhow, i hope it works for you and Steamspy gets better and better. Unfortunately, i don't think i'm going to sign up in the near future and i might have to look up 'leaked' data.
Dunno, maybe as a group of GAF-ers we can get together and subscribe for a shared Pro account.
I'm absolutely ok with people not signing up and using free version.

The idea behind paid version is that only developers will use it. :)