- ARPDAUPosted 4 years ago
- What’s an impressive conversion rate? And other stats updatesPosted 4 years ago
- Your quick guide to metricsPosted 5 years ago
Game Theft Affects the Bottom Line
This is a guest post by Matt Warren of HalOtis.
In my last post I wrote about some strange numbers that I was seeing in my app that I just had to investigate.
But first just some background information. My first iPhone Game (UFO Invader) was released just one week ago as a $0.99 paid app. It’s also a bit of an experiment so I included an ad at the bottom of the the menu screens as an extra way to make some money. With the App I can see the sales in my reports from Apple and I can see my ad impressions and clicks in reports from AdMob.
On Tuesday, things started to get a bit wacky.
I noticed my ad impressions (and clicks) spiking massively. However there wasn’t a corresponding increase in sales to account for it. Either something was not counting properly or I was missing something. This was a big discrepancy on the order of 1000x what I would have been expecting for the day.
After looking at the numbers it became obvious that more people were playing the game than had bought it. A lot more.
At this point with limited publicity and 4 days after launching the game I had registered 30 sales. But there were over 200 ad clicks and over 1200 ad impressions (1000 in one day). There were 400+ different users with top scores registered in Game Center.
A Google search revealed what I suspected. A handful of forums and automated blogs were linking to cracked versions of my game. (I won’t link to them because I don’t want to give them any SEO Juice)
I wasn’t sure how to feel about this. On the one hand as a small one man developer shop getting my game into the hands of 1000 kids is awesome exposure and it might help drive sales. The other hand I’ll explain in a bit.
Effects of Piracy
First, here’s a couple of graphs showing my sales and ad revenue for the first week on the App Store.
You’ll notice there were negligible ad impressions until the 9th. The 30 people who bought the game at that point were registering 5-10 ad impressions per day. It very quickly spiked up to 1000 per day. This correlated with an increase in ad revenue. Unfortunately there has not been any positive effect on sales (yet) as a result of being pirated.
There were some other numbers that caught my attention as someone who is familiar with both side of the advertising fence – advertiser and publisher. I had run a test ad campaign to see if I could profitably make sales by advertising my game. Unfortunately only 1% of clicks converted into buyers, and even at a very cheap $0.03/click it wasn’t possible to run advertisements at a profit. That sucks. If it were possible to profitably advertise then my money worries would be over. Alas.
The advertising that I did run on AdMob reported about a 0.5% CTR. The ads that my game was serving was reporting a 18% CTR.
To me, 18% CTR is clearly evidence of click fraud. Nobody clicks ads that much.
What was going on? Well it seems that people who download cracked copies of the game are clicking ads in order to “help” me, the developer, make some money. Unfortunately click fraud has the opposite effect. It reduces the value of those clicks below the minimum bid allowed for everybody. Meaning that I and every other indie game developer cannot effectively use mobile advertising to find customers.
Here’s a concrete example with my numbers:
$10 in advertising buys 334 clicks at $0.03/click. My cut on a $0.99 game is $0.70 which means I need to make 10/0.7 = 14.3 sales on average for every $10 worth of ads or a 4.3% conversion on clicks in order to break even on the advertising.
4.3% for a $0.99 impulse purchase should be EASY. Without knowing any better I would expect to see closer to 10-20% for something like this.
Unfortunately it is impossible to know what percentage of the ads I paid for were actually fraudulent but based on experience running ad campaigns for the last 4 years on the web it seems like mobile ads are terrible and my best guess is that AT LEAST 75% of clicks were fraudulent.
I was seeing a 1% conversion, but removing the fraudulent clicks might have bumped it up into the realm of profitability.
Effects of Click Fraud
This level of fraud creates a huge problem for developers.
- The value of a click is so low that using ads to make money doesn’t work. (My stats show a click pays just 2.8 cents)
- All the Advertisers who have good things to sell won’t buy ads because they’re paying too much for fraudulent clicks.
- Developers can’t profitably advertise their Apps with mobile advertising
- The high volume of clicks that I saw with my Game could potentially get it flagged, accounts locked, money taken back, and prevent any money being made now or in the future from advertising.
To those people who download cracked iPhone apps and want to help the developer here’s the best thing to do:
- Don’t click any ads unless you are actually interested in what they’re advertising. Ever.
- Tell your friends to check out any cool apps you find.
Living with Piracy
You know I’m cool with people pirating my game. They wouldn’t have bought it anyway. And with the right bits of marketing in place I could still communicate with them, promote future apps, sell in-app-purchases to them and ask them to tell their friends about my games. But destroying the advertising ecosystem makes my job more difficult. It eliminates what would be the simplest way for me to promote my game to people who might actually buy it.
What are my options as a game developer to deal with this?
- Release free to play games and be more creative about monetizing them.
- Ignore pirates and forget about trying mobile advertising to advertise.
- Ensure you have ways to communicate with users through push notifications, email, or in-game messages.
Unfortunately, There’s pretty much nothing that can be done to thwart hackers. If they want to crack software bad enough there’s always a way to do it. What sucks most of all is that the click fraud puts me at risk of losing a lot of money if my account gets flagged.
But this is also very early on and it will be interesting to see how things change over the next few months.