We all know the importance of tracking changes in your CR, so you can make changes in the demo and see if it sells better or not. Steve's A/B experiments come to mind.

The problem is I can't do it. No matter how well I calculate the number of downloads from my site, it doesn't have much relation with the number of sales, since every time a major distributor starts selling the game, I get many sales from the site, too many to be caused by my demo downloads.

If I had much more direct sales, I think this effect could be neglected; but as it is now, it has a great impact on my statistics.

Any ideas on how to deal with this?

You need some way of distinguishing orders from demo downloads and those from these major distributors. I'm guessing your problem is people going directly to your site, hence not giving you any chance of working out where they got your demo.

Maybe you could ask them on the order page?, where they got the demo from?. Not all would enter the correct info but it might help.

I think that this is a major effect of using publishers, and you can't have a valid solution.
I too don't have control on USM downloads from 5-star, for example, and got quite few orders from it: but the CR could be 50% as wel 0,00001% since I don't have download data :D

There isn't a solution to your question, I'm sorry :) I don't think publishers are willing to give you exact download numbers...

What if your demo were to detect an internet connection, and if it exists, it would ping a "hidden" web page? -- only the first time the game was played. You could then count the number of hits to the web page.

Taking this one step further, you could give different builds to different publishers/portals. Each build would ping a different web page. That way you'd be able to approximate the count of downloads for each publisher. Not sure how legal this is, but I think it could work.

Well if you think people are coming to your site from the demo link, then you could give each publisher a build that links to a particular page. For example:

On your "Buy Now" link, it could be www.foobar.com?game=X&action=buynow&referrer=DistributorX

Make the CGI script keep track of the referrers and possibly set cookies. Also make it so you only have to change the referrer value in the url string to indicate a different distributor. This way when you modify the actual demo build, you're only modifying a tiny piece of code (ie. referrer=DistributorX).

I think this will be the most accurate since they'll most likely be visiting your website through a link in the demo. While this doesn't tell how many downloads you're getting from DistributorX, it will tell you how many people are purchasing the full version from those demos and thus you can count them out while doing your CR calculations for your site (your original goal).

If you're interested in download counts from distributors, you can always have your demo "phone-home", as ConanCurrie mentioned, but some people are wary (including me) when applications attempt to connect to a website for no apparent reason. Though Cas will tell you otherwise. :) So it's a touchy subject and regardless you still would have to do something like I mentioned above about tracking website links to know who's purchasing from you.

So there's a way to do it, it just requires some pre-planning.

To:ConanCurrie

The internet connection doesn't work well.
Not all peopl have broadband connection or simply don't want application that connect to internet without their consent.

To:Chris_Evans

Publisher will require ABSOLUTELY NO links to external pages (expecially the BUYNOW, they may allow highscore table like Betty, but for sure not the buynow!) beside the one of their site. At least this is my experience with BFG.

I'm confused now...

I thought ggambett's original problem is that he's getting sales to his site, which he's unable to determine where they came from, so he's having a hard time figuring out his actual site CR.

If the demo and full version are handled by the Publisher (like Real Arcade and etc.) then I don't see what the problem is. It doesn't affect his site CR.

I'm assuming (correct me if I'm wrong) that certain other sites are hosting his demo, but the final purchase actually happens at his site. But he's not privy to their download statistics, so it's messing up his CR calculations.

If the final purchase is made at his site, I would think he could add any type of link in the demo he wants.

Yes his problem is that one: he get sales from his site because ppl see his game from Real or BigFish, but for some reason go buy the game from his site.
He can't track that: because Real, BigFish, etc don't allow your games to connect to internet in external sites (except their own site). So he can't know exactly how many downloads he get :)

Originally posted by Chris_Evans
I thought ggambett's original problem is that he's getting sales to his site, which he's unable to determine where they came from, so he's having a hard time figuring out his actual site CR.

If the demo and full version are handled by the Publisher (like Real Arcade and etc.) then I don't see what the problem is. It doesn't affect his site CR.
You're right about my problem; you're not right about this last part. It does affect my CR. For some reason, some people who download demos from the big portals end up buying directly from our site, as Jack explained.

Originally posted by ggambett
For some reason, some people who download demos from the big portals end up buying directly from our site, as Jack explained. Blessed be those people! If anybody could come up with a scheme for promoting this, that'd be great. Although the portals would probably disallow it in an instant.

When I spoke to some Real dude, he said that you're allowed to present your company logo/name in the game, but not to have any embedded links to your page. Are you allowed to write your url ANYWHERE (like in the readme or with graphics in your game), or is that prohibited?

Different builds prior to launch are the key way to keep accounts seperate. There are other ways, but different builds are the easiest. Plus it could make it harder to crack your game.

Moving on to the issue: Does everyone understand negative space in art? Its drawing what isn't there to make a picture look like it is.... uh, I doubt that made much sense but I suck at art.

The trick I would use to factor this in would be to use negative space in download and purchasing terms.

Have a seperate build for people who download directly from your website. Their "buy" screen A) Sets a cookie, B) directs them to a different buy screen than someone who is visiting your website and clicking "buy" (the two URLs may APPEAR the same, but can be different for tracking reasons).

Track the people with cookies to see if they leave the buy screen, go back to your website and click buy AGAIN (meaning they go to the OTHER buy screen) and see if they buy it there.

So what you end up with is this:

If someone downloaded your game from bigfish and went manually to your website and purchased, they would buy it from purchase screen A. Lets say 10 people do this in a month.

If someone downloaded your game from YOUR site and went through a click in the game "buy now" it takes them to purchase screen B. Lets say 100 people do that a month.

If 80 of those people buy the product, 10 leave without buying, and 10 more leave, go back to the main website, and click buy again you get the following figures

20 sales from site A, 80 sales from site B. You then check to see how many of those 20 had cookies from site B on them, and you find that 9 did (and 1 jerk was blocking cookies).

You now have a reasonably correct statistic that 11 sales came from outside sources and 79 from your own personal downloads.

Now its up to you if this amount of work is WORTH improving the accuracy of your CR. You could always ditch the cookies part pretty easily I think and still get a more accurate reading.

Joseph Lieberman

Originally posted by ggambett
For some reason, some people who download demos from the big portals end up buying directly from our site, as Jack explained.


Ah, I understand now. I didn't realize you were getting spill-off sales from the big publishers (that's a good thing!). I thought it was just about other sites hosting your demo.