Don't miss
  • 2,007
  • 5,500
  • 5,764
  • 120

What’s an impressive conversion rate? And other stats updates

By on July 26, 2012

From editorial assistant Zoya Street


Part of my job here at Gamesbrief is to regularly update the stats pages for the Free to Play Spreadsheet. Recently I’ve been particularly interested in conversion rates: namely, what makes a good conversion rate? We know that the common industry benchmark is 3-7% monthly conversion, but that some games have audiences that spend considerably better, with lifetime conversions of around 20%. Their developers then go running to Edge magazine and the like, declaring their conversion rate to be the largest in the world. I take that as a challenge, so I’m tracking the high-flyers to see if their claims to glory measure up.

I’m also interested in how low conversion rate can go when it falls short of the widely touted industry benchmarks. Majaka reported a conversion rate of 0.1% on Ski Champion, admitting that they hadn’t focused on a sales strategy when launching their game. Benchmarks aren’t the natural resting point of online player behaviour – it takes work to reach them.

I’m collecting a lot of other KPIs as well, so that you can sanity check the projections in the free to play spreadsheet.  Here’s some of the figures I’ve added in the past month:

Conversion rate

Autoclub Revolution, Eutechnyx: 9% lifetime conversion (source: The A List, 6/28/12)

Ski Champion, Majaka: 0.1% lifetime conversion (Source: Majaka, 6/8/12)

AI War, Arcen Games: 15% lifetime conversion (source: Cliffski’s blog comments)

Note that for forecasting, it’s much better to use monthly conversion rates – lifetime conversion is likely to increase as retention increases and time passes since launch, so it’s not really a clean statistic.

ARPPU (Average revenue per paying user)

Giant: $15 (Boing Boing)

Auto Club Revolution, Eutechnyx: $24 (The A List)

Playdom: $20 (Lightspeed Venture Partners)

EA Social Sports: $56 (Lifetime: AllThingsD)

CPA (customer acquisition cost)

At Game Horizon 2012, Torsten Reil announced that CPAs on iOS have not risen to $1.80

DAUs (daily active users)

Temple Run, free: 7 million DAUs (source: GAMESbrief)

Dark Orbit, Bigpoint: 100,000 DAUs (source: Social Games Summit)

About Zoya Street

I’m responsible for all written content on the site. As a freelance journalist and historian, I write widely on how game design and development have changed in the past, how they will change in the future, and how that relates to society and culture as a whole. I’m working on a crowdfunded book about the Dreamcast, in which I treat three of the game-worlds it hosted as historical places. I also write at Pocketgamer.biz and The Borderhouse.
  • Tom Hudson

    Umm, AIWar isn’t F2P, it’s $10.
    http://www.arcengames.com/w/index.php/aiwar-buy 

  • http://twitter.com/markjnet Mark Johnson

    I’m confused over engagement and retention measurements. Everyone seems to measure it differently. People say for example ‘weekly retention’ and give a number, but I have 3 different ways to measure this. (We actually started using all 3 approaches.)

    A) strictly by that day:   2nd day, 7th day, 30th day: ie % of people who we saw playing on the nth day after install. (day after, week after, month after). But in this measurement, we had to see them exactly on that day to count them, so how frequently they play effects too, not just if they continue to use the app.

    B) strictly by a time period:  2nd day, 2nd week, 2nd month. % of people seen during that period. ie ANY time during that period. So if they install on day 1, and we see them on day 5 and day 9, they would get counted as a 2nd week retained user, because we saw them at least once during the 2nd week. These % are of course higher than testing for an exact day. But you don’t count them as a 2nd day user, because you didn’t see them on the 2nd day. I have a feeling that is the best way to measure retention.

    C) any time during of after a time period:  2nd day, week, or month, including any time after that period. So these % measure retention as the highest %. The difference with this approach? Taking that same example as above where you see someone again on day 5 and day 9, this time you count that person for 2nd day and 2nd week, because you saw them at least once any time during or after the time period day. This also seems like a reasonable measure of retention, except that it skews high by people who might occasionally fire up the game to take a peek at an update the app store pushed them, without really still being an engaged player.

  • http://www.gamesbrief.com Nicholas Lovell

     Mark, that is extremely helpful. Would you be interested in turning it into a guest post, with some real data showing how they skew differently.

    As you rightly say, people use the same words to mean different things all the time. We do our best to report what we hear, to sense check it with our experience and to be consistent in our usage.

    Perhaps the best we can say is that we may not always have data with great definitions, but we’re still the place where you can find more data than anywhere else :-)