Lower your price, lower your ratings. Lower ratings, lower social proof. Lower social proof, lower sales. That’s my theory, and I’ve got data to support it.
I’m a pattern matcher. I like extracting the hint of a signal from noise. For a while now, I’ve had a hunch that pricing an app higher would lead to better ratings. Been reading Predictably Irrational and Influence: The Psychology of Persuasion.
So when, in the closing days of summer, I saw Pete Schwamb tweet about the ratings decline he saw for Cricket Song (his clever app that tells the temperature by listening for cricket chirps) when he starting giving the app away I filed it away as a data point. A short time later, I read Andrew Binkowski’s blog post A Week At $0.99 and noted that the ratings for his baby naming app Stork Drop dove after cleaving 2/3rds off the price.
I sometimes bury the headline; so I’ll state it now, then back-fill the supporting data:
Don’t Discount Your App At Launch
Conventional iPhone buzz-building wisdom tells you to launch your app at a fraction of your real price. Don’t.
During the early days of an app’s life, social proof is critical. Social proof is how you decide whether or not to jaywalk in a new town: you look to see what others do. More formally, social proof is:
a psychological phenomenon that occurs in ambiguous social situations when people are unable to determine the appropriate mode of behavior. Making the assumption that surrounding people possess more knowledge about the situation, they will deem the behavior of others as appropriate or better informed.
Applied here, social proof means that ratings provide prospective buyers a peephole into the reasoning of the assumed-to-be better informed earlier buyers.
So, it’s important to maximize your rating early on. Discounting your app won’t maximize ratings:
Start with the Christmas tree lights test: to ensure even light-coverage step back from the tree and squint at it. Squinting blurs the lights, causing them to fill in. It’s easy to see — and fix — gaps. Try the same thing with the ratings-distribution graph in the App Store; squinting blurs the specifics and presents the general shape of the curve. On the whole, you’ll see that the graph’s weight for free apps shape sags towards 1 star. As you move up in price the weight shift to the top of the chart where the better ratings live.
|Free App||$2.99 App|
Compare the the top-10 paid and free apps. The average rating for top-10 paid apps is 3.9. Free apps average 3.4, half a star’s difference!
What’s more, when you break the apps down by price you’ll see that every price increase is accompanied with a ratings increase:
Furthermore, the likelihood of a 1-star review goes up as the price goes down: 1-star ratings made up 23% of the total number of ratings in the free group; 16% in the paid group. The percent of 1-star ratings dropped at each price increase:
Pete’s tweet hit the nail on the head when he labeled the folks who got his app for free as uninvested users. Customers with some skin in the game cary a psychological pressure to feel that they’ve been wise in their purchases; they’ll tend to over emphasize their positive feelings.
What’s more, I’d bet that the likelihood of a user uninstalling an app goes down with its costs. Uninstalling apps triggers the negative review bias.
I’ve got two somewhat related items that I couldn’t fit in elsewhere:
1. The “Sex Jokes Lite” application is clever/manipulative about their ratings. They push people to give 5 star reviews with this little bit of a social contract: in their app-description they say, “IMPORTANT: If you think the jokes are TOO dirty, please give a 1-star review. If you want them dirtier, give a 5-star review. This way we know what direction to take in the upcoming updates !”
2. I’m interested in the number of ratings and number of reviews as compared to total of installs on a category by category basis. Games have a small ratio of reviews and ratings to total downloads. I’m curious about this more broadly. Have data? Leave a comment or email me.