Instagram Information Saver Mode. We just lately shipped Information Saver Mode, a… | by Cristina Acha

We recently shipped Data Saver Mode, a new feature on Instagram for Android that allows the app to use less mobile data. In this post we will explain why we decided to work on this feature, how we developed it, how the engineering was implemented and how it affected people.

The motivation for creating the data saver mode was threefold:

First, research suggested that many people felt limited when using Instagram because it was consuming a large percentage of their data. We conducted a survey in various countries (US, UK, India, Indonesia, Brazil, France, Germany, Japan, Argentina) in early 2018 asking how much data Instagram uses. More than 50% of the respondents chose “a lot more than I expected” or “a little more than I expected”. Additionally, 30% of respondents said they run out of data every month and 28% of respondents said they would use Instagram more if there was a data save feature.

The second motivation was that we noticed that Instagram was consuming more data than we originally wanted. This made sense intuitively: the more someone uses Instagram, the more content they consume. However, we have found that overall data efficiency (i.e. the time it takes to use the actual app for every megabyte of data consumed) could be improved. To illustrate, you can see below how IG data efficiency is below that of Facebook, Facebook Lite, and WhatsApp.

The third motivation was the native data saver function of Android, which was increasingly popular with users. Starting with Android 7.0, users can enable Data Saver for the entire device. When this option is activated, the system blocks the use of cellular data in the background and signals the app to use less data in the foreground. We used IG with the native data saver enabled. Since our app is media intensive, this native implementation leads to a disruptive user experience as photos and videos either load very slowly or not at all.

Google provides APIs that can be used to verify that someone has enabled Data Saver at the system level, but does not allow developers to change the native Data Saver implementation at the application level [1]. Users can also restrict apps in native data saver. So if someone enables Data Saver they can restrict IG and use our custom Data Saver mode instead for a less disruptive user experience.

The following are the levers we used for our Data Saver implementation:

Currently, upcoming videos are prefetched in a person’s feed and story viewer so that the videos can be played when a user calls a video on the screen. We hypothesized that this behavior uses more data, especially if the user doesn’t scroll to upcoming videos later in the feed. Hence, we can turn off video prefetch so that we only fetch video content when the user has paused scrolling on a video to indicate that they are viewing the video. This would reduce data usage as a person might not want to watch all of the videos they are scrolling through on their feed.

We currently automatically play all videos when they become visible on the screen without user interaction. We assumed that this behavior was using more data, especially if the user didn’t intend to watch every single video they scroll through. That way we can turn off autoplay and show a play button for users to play videos manually. This is a more drastic version of disabling video prefetch, as additional user interaction is required to get in touch with a video medium.

Currently, we decide the image and video resolution based on constraints like a user’s connectivity and bandwidth. Rendering high-resolution media consumes more data than low-resolution media because of its larger file size. This can be of great concern for users in unique connectivity situations that may be difficult to see at the application level. For example, people using Pocket Mobile will appear as if they have Wi-Fi. In reality, they are using cellular data. This allows users to specify which connectivity setting they want to view higher resolution media at. That way, users can still browse the content they want without using too much data.

The downside to activating the above three levers is that we want to make sure that users continue to have a consistent browsing experience with reasonable media load time, since Instagram is a media-intensive app.

At Instagram, we know that a significant proportion of our users are in markets where connectivity can only be accessed via cellular cellular data, as opposed to home connectivity (i.e. WiFi). The demand for affordable connectivity has increased, making data cost a key factor in a user’s decision to engage with online content. While we can mainly look at emerging markets, there are also developed countries where a significant portion of the population is also data aware due to the high cost of data. With this in mind, we tested in Indonesia, India, Argentina, Germany and France. The team tested different variations of Data Saver with the above three parameters and with variations that showed user-visible options to disable autoplay and control media quality.

In this initial country test, we found that disabling video prefetch strikes a good balance between reducing data usage and browsing appropriately. Predictably, users would consume less video by not automatically playing video content. However, we also saw that people appreciated having explicit control over media quality and automatic playback.

We found that there were two variants that were considered the best performing test variants:

  1. Deactivate video prefetching. Option to select when high-resolution media should be received (“Never”,
    “Only via WLAN” or “Both via WLAN and via cellular network”), whereby the selection is set to “Only via WLAN” by default.
  2. Deactivate video prefetching. Option to select when high-resolution media should be received (“Never”,
    “Only via WLAN” or “both via WLAN and via cellular network”), whereby the default selection is “via both WLAN and cellular network”.

The variants are the same, the only difference being that the default selection for the “High-definition media” option is different (“only for Wi-Fi” and “for both Wi-Fi and cellular”). Each country did better in certain countries, so we decided to do another country test in Canada and the UK (as previous research has shown, these countries are also data aware). This influenced our decision to finally test globally and start for everyone.

For our global launch, we’ve decided to keep the high-definition media user option, and by default “WiFi only”.

For our testing in CA & GB, the best performing version was the one that turned off video prefetching and defaulted to high definition media for both Wi-Fi and Cellular. Since this is an opt-in and only ~ 10% of the people in the test group chose to do so, the results we saw on our A / B test were pretty watered down. Even so, we saw a significant drop in data usage in cellular communications. We also saw increases in the number of interactions, the number of media created, and other metrics of engagement. These were significant gains, especially when you consider that only 10% of the users in the test group drove them. Finally, we only saw regressions on the video loading metrics that we expected from disabling video prefetching, but they weren’t too big.

In our CA&GB test, 1.6% of users switched to standard high definition media over Wi-Fi only to high definition media on both Wi-Fi and Cellular, and 10% of users switched to standard high definition media on both Wi-Fi and Cellular have only switched to high-definition media over Wi-Fi. While we’re not entirely sure why we saw this big difference, one of our hypotheses is that people who are very aware of their data usage are more likely to actively use data storage options.

For our global test, the best performing version was the one that turned off video prefetching and defaulted high definition media only over Wi-Fi. This test was conducted through a less targeted lens, but we could still see improvements in engagement and data consumption on mobile devices. Given all of these positive results, we concluded that this feature would be very beneficial for users with limited data, and we shipped it worldwide in June 2019.


The function of the data saving mode would not have been possible without the cooperation of research, data, technology and product. Many thanks to Elisa Lou, who wrote this post with me and is the engineer who worked on the data saver mode – without you this would not have been possible. Many thanks to Kat Li, Jeff LaFlam, Michael Midling, Colin Shepherd and many more.

If you would like to learn more about this work or would like to join one of our engineering teams, please visit our careers page, follow us on Facebook or Twitter.

Comments are closed.