I just completed an analysis of the registration rates for our games (i.e. sales divided by downloads). They do vary a bit from month to month, but what's interesting is to see just how big the range is: 0.13% for the worst-converting game and 2.9% for the best.

So out of every 1000 downloads, one developer's game is getting about 29 sales, and another developer's game is getting just 1 sale.

But wait... it gets better....

Of the 19 games we have in our product line, the top 10 converting games are all puzzle games... not a single arcade or strategy game made the top 50%.

These registration rates aren't perfectly accurate, but the relative span should be pretty close to reality. We're currently implementing a system whereby the calculations can be made much more accurately by the use of cookies. I.e. we'll track when specific visitors download a game and then come back to buy, and we'll calculate the registration rate just from those visitors who can be tracked. This will allow us to rule out people who double-download the same game or who use download managers that create multiple hits on the same file. We'll also be able to track things like what % of visitors who view a game's product page actually download it, and then what % of them come back to buy. This will give us the data needed to optimize our own sales pages as well.

Food for thought....

Thanks for sharing that information. Everyone is always trying to get some real numbers out of the successful guys, so I'm sure they'll appreciate it :)

That being said, I'm not sure if any conclusions can be made regarding 'puzzle vs. action/strategy' games. I think it's more likely an indicator of the type of customers that frequent Dexterity than an indicator of the type of customers that prefer those types of games.

(Not to mention the fact that many of the puzzle games Dexterity publishes are just plain excellent games)

Interesting. Perhaps its because puzzle games aren't made by the big companies but arcade and strategy games are. What I'm trying to say is the puzzle game customers haven't been dazzled by the flashy graphics of the big companies and so they are more open minded about buying an indie puzzle game.

On the other hand it could be that the vast majority of your customer base are puzzle game fans and they are downloading the arcade/strategy games because they are bored or curious.

P.S.

Steve: I just noticed that you put your games into separate genres. I think this will really improve your sales so let us know if it works.

Statistics are always fascinating to look at, but one has to be very careful about making any sort of extrapolation or comparisons.... it's easy to compare registration rates between games and between companies, but people should be very careful when doing this.

Simply put, there are way too many confounding factors here (and way too little data). As a simple example:

- after trying a few advertising campaigns I found I received a very healthy dose of downloads, but with a low conversion rate. Makes sense, because many people will follow an ad out of curiosity, but not really have any intentions to buy.

- on the other hand after an e-mailing, I receive relatively few downloads but with a stellar conversion rate, since often these are customers who've bought and enjoyed previous games and hence will be much more likely to buy the next.

There are many other problems with comparing such statistics... e.g. what download sites do you post on (some, such as download.com, are said to yield lower conversion rates), how big is your file (the bigger, the more people might use download managers), etc, etc, etc.

I know that the 1% rule in conversion has been quoted ad infinitum, but I actually think somebody long time ago brought it up, and people have been quoting it until it became "established fact". Yes, 1% is somewhere in between 0.13 and 2.9%, but how useful is that information if you have a possible factor of 100 in there?

I guess what I'm saying (and slightly ranting) is that a single number as a conversion rate can be a dangerous tool to use for comparisons.

However, if you are reasonably sure that most variables are the same for a number of games, you can certainly look at comparisons for such games - I think Steve definitely has some interesting data there, since he's comparing his own games (which all presumably are marketed in a similar fashion). I too have found that for my own games Wonderland (a character driven puzzle game) converts much better than ReCharge (an arcade shoot'em'up).

But once you bring in a couple more unknowns, I think it is impossible to compare such numbers.

Opinions?

One more "unknown" to add to the list:

Many download sites host the files themselves... hence you don't know how many people downloaded from them. Add to that demos obtained from shareware CDs, and you're denominator can be thrown off altogether.

All the points about accuracy and the cautions about drawing conclusions are valid, so you can't overgeneralize from these numbers. Nevertheless, I wanted to share this just to show that the spread can be very wide... suggestive evidence that if your demo isn't converting well, you could be leaving more than 90% of your potential sales untapped.

What do you mean by "leaving 90% of sales untapped"? Do you think the games that aren't converting well have potential to do so? I was assuming that those low conversions were do to weaknesses in the game, and that higher conversion couldn't be expected.

That's a very interesting statistic Steve, thank you for sharing it! Do the puzzle games also get the most downloads? That is, if you look at downloads/month for each of your games are the puzzles also the most downloaded? I'm wondering if you have any games that have twice the number of downloads as another game that has twice the conversion rate (thereby making up the difference in sales)?

Actually some of the arcade games get quite a few downloads -- one of them was #1 in terms of downloads for a while.

Since there are many non-gameplay factors that can hurt a game's registration rate, I do believe it's possible for many games with a 0.1% registration rate to reach 1% and higher. This depends on the game of course, but there are lots of ways to kill the sale.

Steve,
Sale precentage is redundant when you do
a cross-genere comparison. i believe that
puzzle gamers habbits and action gamers habbits
are quite different when you put the in terms of
buys/downloads.

There is also the factor that Dexterity was known for puzzle games. And I bet a large portion of dexterity's mailing list are puzzle gamers who signed up when dexterity only offered Dweep or Fitznik. So right now, the target audience would be highly skewed towards puzzle games.

I have been doing tracking for quite a while, and here are some random observations:

1. People take a long time to register. Sometimes weeks or months.

2. With a simple redirect, you can cookie user downloads, even from other sites, like download.com. However, I don't do that because I am not sure about the ethics of it, plus the download accelerators do not save the cookie. And since they also download a game in hunks, the number of downloads in the log file is useless. You have to go by bytes.

So here's the formula I use to evaluate the performance of a particular game:
g = megs downloaded
s = number of sales
m = game's file size (in megs)

s / ((0.8 * g)/m)

I multiply the megs downloaded by .8 to account for people who abort downloads. Although it could easily be less than .8....

3. Different download sites perform far better than others, with my own site being the highest performing referrer.

4. The average customer comes to the purchase page 6 times before they actually buy.

5. Get a good log analysis program!



-Mike

This range of conversion rates is very similar to my own spread, and I do strategy and arcade games, so I would say these are the standard rates for downlaodable games at the moment. I had assumed that puzzle gme rates would be generally higher than that, but its still not bad, especially if the games are small you are earning good money per gigabyte of bandwidth paid for.

Another figure that's more significant to me than sales/downloads is revenue/downloads ... and also profit/downloads. Since games have different prices and some are bought with a multi-game discount, there are different dynamics here. For instance, one game that gets a decent conversion rate is bought in a bundle most of the time, while a different game is usually bought solo (and thus at a higher price). So you have to be careful that you measure what you want to optimize.

Have you tried charging $ 9.95 instead of $ 19.99 across the board for those highly downloaded arcade demos? You might not make that much per unit but if they are getting download alot, you got get more conversions and make more money ultimately.
I think you might be suprised at the conversions if the games have any lasting appeal on the hard drives of downloaders.

Strayfire would be a seller if at that price and if been giving the spotlight for downloads. I found it slightly amusing and I'm not into shoot em ups in space.

I tried pricing experiments with A Snake's Life. It's my only arcade game at the moment and it converts lower but downloads a lot more copies than the my other titles, which is consistent with what Steve has mentioned above. Anyway.. at 10$ it sold pretty much exactly the same as it did at 20$.. I have it on at 14.95 now but really it's not much different than 19.95 so I'm probably just losing 5$ every time I make a sale. It really seems like below 20$ people care more if they like the game than if they save a 5er

I've found the following indicators work well for determining if your price is too high or too low:

1. If you get a few comments about your price being a bit high, then you're probably spot on.

2. If you get no comments about your price at all, you're probably too low.

I don't have a 3...I've never priced something "too high" yet. Not sure if there is any indicator of "too high", besides a sudden drop in sales.

Like most people, I'm sure I undersell myself frequently. So I try to push my price up to the point where I feel like it *might* be too high, but where I can still justify it. That seems to keep the price at least within range of "just right."

Just my thoughts.

-David