Are pre-installed apps more power hungry?

Carriers as well as handset manufacturers pre-install several apps on mobile devices that they sell. In this post, we exhaustively study pre-installed apps by leveraging the rich energy profile information that Eagle provides — our analysis below is based on data collected from 70K+ devices. One primary concern that users have about pre-installed apps, which leads them to be used less often, is about their battery usage being high. This concern gets magnified given that users quite often can’t delete these apps unless they root their phones, etc. In sharp contrast to this preconceived notion, we show that pre-installed apps actually have similar power draw as apps with similar functionality from the Google Play store. Our conclusion: power usage of these pre-installed apps is not worth losing sleep over!

Our battery management app, Eagle, has a patented technology to accurately keep track of both the time and energy spent inside each app. For this study, we analyzed data collected by Eagle from three different Android device manufacturers: Samsung, HTC and Motorola.

We first averaged apps we see across all the users of Eagle and compared the number of pre-installed apps that users have on average compared to the total number of apps. Amongst the three most popular Android device manufacturers, Samsung phones have the highest percentage of pre-installed apps (25.6%) followed by HTC (23.3%) and Motorola (12%) (see Figure 1).

Figure 1: Number of pre-installed vs. total apps across device manufacturers.

The Table below shows the average daily usage across the different device manufacturers, with Samsung devices being used the least (88 mins. per day) compared to Motorola and HTC. In terms of usage across different pre-installed apps, Samsung devices have the least usage at ~6 mins per day compared to 12.6 mins per day for HTC devices.

 Screen Shot 2015-02-14 at 3.08.29 pm

The next question we ask is how often are pre-installed apps actually used compared to apps that users install directly from the app stores (see Figure 3). Interestingly, HTC phones have the highest usage of pre-installed apps at 12% of total daily usage across all apps. Samsung is second at 7% followed by Motorola at 6%. Next, looking at the breakdown in time spent across the variety of pre-installed apps for these manufacturers (see Table above), a few other notable trends emerge: The pre-installed Browser on Samsung and HTC phones is used for about 2 minutes per day whereas Motorola users use the pre-installed browser very infrequently. However, given that Samsung devices are used less often on a daily basis, hence, we next compare each pre-installed app’s usage in terms of percent time spent across all pre-installed apps (see Figure 4). In terms of Camera apps, Samsung and Motorola phones seem to have a popular pre-installed Camera app whereas the pre-installed Camera app on HTC devices seem less popular. For Motorola devices, the most popular pre-installed app is a Gallery app at 50% usage. Samsung devices also have a non-trivial usage of the Gallery app at 17% across all pre-installed apps.

Figure 3: App usage on Samsung, HTC and Motorola devices.

Figure 4: App usage across pre-installed apps on Samsung, HTC and Motorola devices.

The low usage of pre-installed apps compared to user installed apps should come as no surprise, as they were not actively downloaded by the user to serve a specific need — hence user may sometimes be not even aware of them. Moreover, several myths prevail about pre-installed apps: (1) overall they draw more power, (2) and they draw a lot of power in the background, i.e. even when they are not used — yet they can not be deleted easily.

In Figure 6, we compare the foreground power drawn by the six most popular user-installed apps versus the six most popular pre-installed apps. We see they are remarkably similar, suggesting that pre-installed apps should not be discriminated in terms of power efficiency!

Figure 5: Average foreground power (mA) for user-installed and pre-installed apps on Samsung devices.

Next, we pick groups of pre-installed and user-installed apps with identical functionality, e.g. Samsung Clock and Android clock, and perform a head-on comparison within each group. The first set consists of five such groups of Samsung pre-installed apps versus user-installed apps. Figure 7 shows the power draw of the two competitors within each group is largely a wash — there is no clear winner or loser. This further validates our previous finding that pre-installed apps are not drawing more power than their user-installed counterparts.

Figure 6: Foreground power between pre-installed samsung apps and user-installed apps on Samsung devices.

So far, the pre-installed apps seem to be vindicated as not being overall “power hungry” when they run in the foreground. But in the background — do they consume more power? To answer this, we analyzed the “background” power usage of a few of the most popular pre-installed apps and user-installed apps on 8K devices. Figure 8 below shows a scatter plot for these apps as the average daily background power over the app’s average daily usage time. Amongst the user-installed apps as well there is no clear trend as there are apps that draw more power in the background (e.g., Facebook Messenger) and those that do not (e.g., Facebook). Moreover, on comparing two different browsers: Chrome which is user-installed and Samsung Browser which comes pre-installed, we find that they are quite similar in terms of background power usage. Samsung Video Player when compared to YouTube draws more power though and hence it appears that there may be room for power-improvement in Samsung Video Player. Hence, background usage corroborates what we found about foreground usage earlier — that pre-installed apps seem to be consuming as much power on the average as their user-installed counterparts.

Figure 7: Background power draw of a few apps compared to their daily usage on Samsung devices.

So far, we’ve focused only on pre-installed apps as installed by device manufacturers: What about pre-installed apps as provided by carriers? Are those more or less power-hungry? Figure 9 below, groups similar apps by functionality, while including a user-installed app, an AT&T pre-installed app and a Verizon pre-installed app in each group. Once again, based on the data below, we can’t argue that Verizon or AT&T pre-installed apps are more or less power hungry!

Figure 8: Foreground power usage of groups of apps with similar functionality while comparing user-installed, AT&T and Verizon apps per group, on Samsung devices.

Conclusion: In this article, we’d set out to examine whether pre-installed apps are indeed consuming more power than their user-installed counterparts. Contrary to preconceived notions about pre-installed apps, they actually do not consume any more power in either the foreground or the background. Hence, our recommendation to smartphone users: if battery usage is your concern, there’s no need to root your device to delete the pre-installed apps.