Apple appears to be on a campaign to ban the practice of pay-per-install, where developers offer their apps in other games and pay for downloads when players install their titles for virtual currency. Developers often use this tactic of paying per download to break into the top of the app store charts. (Update: We also have a story here as of April 20 that covers how developers are dealing with the crackdown.)
Developers began receiving rejection notices for dozens of apps late last week and yesterday, saying that they are prohibited from having offer walls. Apple looks like it is changing the interpretation of clause 3.10 in its developer agreement, which says, “Developers who attempt to manipulate or cheat the user reviews or chart ranking in the App Store with fake or paid reviews, or any other inappropriate methods will be removed from the iOS Developer Program.” It’s uncertain what will happen to the thousands of apps that have already been approved and have offer walls.
The move comes on top of a ranking algorithm change we reported yesterday. It looks like Apple is considering extra factors in addition to downloads for the top free app rankings in the store, possibly including active usage and ranking in a category. Apple did not reply to requests for comment or publicly confirm the changes, but the biggest ad networks serving thousands of developers and gaming companies acknowledged that positions in the free rankings had shifted in a very unusual way last week.
Taken together, these decisions are going to change the business models for thousands of apps which rely on pay-per-install either for revenue or getting new users. Apple clearly wants to fix the perception that its top rankings can be gamed and its approach to managing the app store is converging with that of Google’s.
A developer (with their name redacted) passed on this rejection notice from Apple late last evening:
“We found that your app, or its metadata, includes features or content that can have an excessive influence in the listing order or ranking on the App Store, which is not in compliance with the App Store Review Guidelines.
Specifically, your application allows users to download other apps in order to receive in-game currency. Please refer to the attached screenshot for further information. This feature can have an excessive influence in the listing order or ranking on the App Store.
If you only have to revise your metadata – including icons and screenshots – and not your binary, your iTunes Connect Application State will still show as Rejected. However, we don’t require a new binary for metadata issues only. Please visit iTunes Connect, Manage Applications, and revise the appropriate metadata values or settings. When you are finished, please return to the Resolution Center and reply to this notice.”
Separately, Tapjoy, which is the very largest of all the networks and reaches 200 million unique users, passed on this statement to us:
“As you may have heard, a number of applications submitted for update have very recently been rejected from the Apple App Store based on the fact that they were running incentivized app installs within their apps. This is something new from Apple and we, along with every partner we’ve talked to, were unaware of this prior to these notices of rejection. Like many application developers, we have reached out directly to Apple and look forward to clarification.
To be clear, there is no new Apple policy that we are aware of. It seems there may be a new interpretation of the existing 3.10 clause, which is a bit surprising, as Tapjoy, AdMob, iAd, Flurry, W3i and others all power various forms of app install advertising. Many of the brands that promote their apps via Tapjoy also do the same on other major ad networks across the mobile advertiser ecosystem, and all of the apps we promote on iOS are Apple-approved. 3.10 reads:
3.10: Developers who attempt to manipulate or cheat the user reviews or chart ranking in the App Store with fake or paid reviews, or any other inappropriate methods will be removed from the iOS Developer Program
Unfortunately, we believe much of this is caused by misconceptions around pay-per-install, the free-to-play model, cross-app promotion and their collective value to the ecosystem. We believe there are significant benefits to the advertiser (only pay for what you get), the publisher (monetize users who otherwise wouldn’t pay), and perhaps most importantly to the users, who not only get to discover new, exciting applications, but receive what is essentially a coupon for ad-funded virtual currency inside one of their favorite apps. All of this, of course, adds up to value for Apple as well, by creating a viable and thriving ecosystem.
Tapjoy has been and continues to be very supportive of the Apple app ecosystem, and we were not surprised about the Top Free & Paid rankings algorithm changes – we’re all for incremental changes that add to the user experience and keep the environment dynamic. But banning the largest and most effective channel for app installs has a significant and long-term negative impact on the user experience, developer innovation and advertiser utility.
As the market leader in application distribution and monetization of free-to-play games, Tapjoy is currently coordinating with a number of our developer partners, as well as others in the market, to get a clear understanding of the issues and to continue to partner with Apple to meet their needs, along with those of app advertisers, developers and users.”
There are many ways to look at Apple’s decision and its fallout:
1) Apple clearly wants to fix the perception that its top rankings can be gamed. As the iOS ecosystem has matured, it’s become common practice for developers to set aside a paid marketing budget to promote apps, whether that is through classic banner advertising or the pay-per-install model. Pay-per-install caught on in the last year because it’s cheaper compared to a traditional CPM campaign. With limited, visible shelf space in the app store, there is a huge drop-off in downloads between the top-ranked apps and everyone else. A top 10 ranked free app in the U.S. usually sees more than 100,000 downloads a day, with the top three slots getting more than 300,000 downloads a day, according to developers who’ve held those ranks. There is a big winner-take-all effect.
Paying for enough downloads to get into the top echelons of the charts has become a way for developers to at least guarantee some visibility. Networks like Tapjoy and Flurry provide downloads to these developers through offer walls in other games, where players can get virtual currency if they download apps. With tens of millions of downloads happening a day, if apps are good, they stick. If they’re bad, they fall off quickly.
That said, the cost of a campaign has been rising throughout the spring as the market has become more competitive. At the beginning of the year, we were hearing that a campaign to break into the top 25 might cost at least $20,000, but the bottom range rose to about $30,000 last month. Some developers are even paying between $500,000 and $750,000 a month to have a consistently-high ranking. The amount varies widely because a well-designed app will need to spend less as users stick and pass it on. Costs are rising as the app store has simply become more competitive. Incumbents from the social gaming and console worlds are taking the iOS platform more seriously and several mobile game developers have raised funding from top-tier venture capital firms, giving them capital to spend on paid campaigns.
As pay-per-install has become common and Tapjoy’s model has spawned more than a dozen copycats, this issue has been getting more attention in the mainstream tech community. Chris Dixon, a prominent angel investor and co-founder of Hunch, wrote a post two weeks ago pointing out a free night vision app that was highly-ranked but clearly didn’t work. It appeared to have climbed the rankings by frequently releasing updates to keep its user ratings high and doing some pay-per-install. The app was a client of iDev Network, an obscure cross-promotion network that says it’s still in private beta. (Speaking of which, that app has somehow made it to #4 again this morning on the U.S. free app rankings list.)
In general, changing the ranking algorithms to factor in more than downloads is a good thing: apps can now get visibility on what really matters, which is active usage. Downloads are frankly a terrible metric for measuring success but these numbers are often widely cited in the press because they’re the only ones developers are willing to share.
A few of the larger pay-per-install networks had anticipated ranking algorithm changes for some time. Flurry, Tapjoy and W3i began moving toward a full-service model earlier this year where they provide game design, marketing advice or even funding to developers on top of install campaigns.
What wasn’t expected was an outright ban on offer walls.
2) Apple and Google’s approaches to managing their stores are converging: Google’s Android Market rankings have long included factors other than installs. While the company has never shared its ranking algorithm (in the same way it has never revealed its search algorithm), we suspect it factors in criteria like the long install rate, or whether users hold on to the app after downloading it, and the DAU/MAU ratio, which is a measure of stickiness.
In moving beyond downloads, Apple appears to be catching up with Google’s approach and is fixing rankings to better reflect apps that hold onto to users.
In an interview we had at the beginning of the year with Google’s group manager for the Android platform Eric Chu, he expressed distaste for paid user acquisition.
“A developer with deep pockets could just buy a lot of ads to get to the top,” he said in an interview with us. “So we try to use other signals to help us understand, do users actually like the app? We’re trying to fine-tune it to make sure that great apps do well.”
3) This swings power back in favor of more traditional advertising networks like Apple’s iAd and Google’s AdMob: If Apple is banning incentivized installs, then developers will likely return to traditional advertising campaigns and pay for banners or interstitial ads. This coincidentally favors Apple’s own in-house solution, iAd. Classic CPM campaigns can cost an order of magnitude more than pay-per-install for the same result.
4) The big, freemium game developers will have to go back to the drawing board and rewrite their business models: This week’s rejections also mean that the freemium model that’s taken off in the last six months is probably broken as it exists today. In the last year, we’ve seen developers shift from relying on paid apps to building free apps that make money through in-app purchases of virtual currency. Almost half of the top-grossing apps in the U.S. today are free.
One prong of this approach is that developers give users the choice of paying in the app or getting currency for free if they download another app. In turn, many of the top apps also buy downloads, spending less than what they anticipate they’ll earn over the lifetime of the player. They earn the spread between the two figures. While the profit margins from this approach are a closely guarded secret, the rejections almost certainly mean every company using this strategy will have to rethink their model.
5) If Apple is favoring active usage, the free rankings may become less volatile and it may be harder to brand-new apps to break in unless they are hand-selected for featured placement in the store:
If Apple moves more closely to the Google model and factors in active usage or long installs, the rankings may ultimately end up favoring older apps that have large, existing user bases. If you look at the top free apps in the Android Market today, more than half of them are well-known consumer Internet brands like Facebook for Android, Pandora, YouTube, Google Maps, Angry Birds and Adobe’s Flash which squeeze out indie titles.
In spite of its flaws, focusing on recent downloads injects volatility into the rankings, meaning a greater variety of apps get visibility. With a featured section of the store and an app of the week placement, Apple can combat this to an extent. But it also means that the company is consolidating its power to curate the store. Having relationships with employees at the company to get placement is especially important.
6) It’s deja vu all over again for Tapjoy and Superrewards. Lastly, again this shows the risk of building on someone else’s platform. Tapjoy ran into a not too dissimilar situation in 2009 when it was known as Offerpal and was building on the Facebook platform. Run by a different management team then, it caught flack for having poorly-vetted offers. Sensing the financial opportunity of social gaming, Facebook later moved to create its own universal virtual currency Credits and selected a rival Trialpay to the be the offers provider. That may eventually crowd out Tapjoy’s business on the Facebook platform.
Around the same time, Offerpal fortuitously bought a small startup named Tapjoy, which revived the company. It later renamed itself Tapjoy and today the company has a completely different management team and business model. Yet again, it’s running into a situation where the platform is exerting control and its revenue model may be fundamentally challenged. Adknowledge-owned Superrewards also had a similar model on the Facebook platform and has since shifted into pay-per-install this spring.