Mobile users know the drill when it comes to free apps—decline to pay anything, and you’re likely going to be peppered with tap-through ads and auto-play videos periodically to offset your activity. It’s a trade-off that can be annoying but in many cases still feel worth the cash you’re saving on a free version. But that’s not all you’re paying when opt for free: you’re actually forking over hours off your precious battery life.
Apps powered by advertisements can eat up as much as 33% more energy than those without while draining significant amounts of a phone’s central processing unit and network data, according to a new study from the University of Southern California and Rochester Institute of Technology. The drain from the ads also contributes to higher costs for phone users with limited data plans, as well as key drops in average app rating scores for developers.
“The whole app economy is based on an idea that these free apps cost you nothing,” says William Halfond, a professor of computer science at USC who co-led the study with Meiyappan Nagappan of RIT. “I hope that users become more aware that when they choose an app, it’s not exactly free.”
Many phone users are at this point well aware of their role in an advertising-based transaction, of course—ads are ubiquitous on the Web and in apps from Facebook to many games. The eye opener that Halfond, Nagappan and their teams found, however, is that users are actually paying in multiple ways. Using ad-based apps drains an average of 16% more energy, which shortens the battery’s life cycle from 2.5 to 2.1 hours on average, or 1.7 hours during constant use. The apps are taking up 22% more memory use and 56% greater CPU utilization, or an average of 48% more CPU time. And the data usage from an ad-based app can be twice as much as one without—and 79% more on average. According to Halfond, that’s 1.7 cents per use at an average data cost from AT&T.
But are developers aware that they’re taxing your phone in ways other than the eyeballs and clicks of traditional advertising monetization? Halfond believes the answer is, not so much. Use of ads in apps correlated with an average drop in five-star ranking of .003 per app. That may sound small—and Halfond admits he himself was unimpressed when he first heard the amount—but just three one-thousandths of an average ranking is enough, the team found, to drop apps from the front page of an app store into the zombie ranks.
That’s a significant unseen cost in advertising in apps that developers need to keep in mind, too. “I hope developers realize they can overdo it,” Halfond says. “Somewhere in there, it goes from bringing in money to hurting your revenues.”
How Halfond and Nagappan conducted the study was to take 21 apps from a list of 10,750 apps that had been in the top 400 of Google Play’s 30 categories for the first 8 months of last year and then study their use on a Samsung Galaxy SII phone. Halfond and a student tinkered with the apps’ code packages to isolate out their ads, and compared use over several minutes for the apps with ads versus without. So for the app Restaurant Finder, for example, they searched for a type of restaurant and then looked at their top result. Naggapan, meanwhile, worked with another student to mine data from reviews and app store stars to study the difference of ad-based apps’ scores from those without. The students came from USC and Queen’s University in Canada.
Halfond hopes that smartphone users will become more aware of the extras they’re giving away when they tap through an ad—for many of us, short battery life and data usage is a constant gripe—and that as word spreads, developers will become more careful about over-saturating their apps with ads, too. To further that effort, the researchers now plan to build techniques to correlated ratings and reviews with app usage in a way that will be able to predict for a developer whether their new app version goes too far with ads.
“Developers still don’t know what that point is,” Halfond says, but to tell them could take a while. “It’s a pretty a hard problem.”
