Mixpanel: How Addictive Is Your App?
At the end of the day, developers all want the same thing: To build products that are useful, beloved and seamlessly integrated into people’s daily lives.
At Mixpanel, we’ve always dissuaded customers from relying on “bullshit metrics” like DAU and MAU. In our view, a truer indicator of success is retention, which will tell you (and your investors) how sticky a product is, and how many users return to it after their first experience. Whether you’re building a game, a social network or a CRM tool, your ideal customers are the people who engage with your product at least once every day — better still if they’re using it constantly.
We’ve taken this a step further by introducing a new analytics report — Addiction — which will be available to users soon, and which tells companies how frequently people use their apps throughout the day. For example, if you’ve built a game, and a customer plays once in the morning, once at lunch and once on the commute home, they’ve engaged with your product during three separate hours: As such, they’ll fall into the three-hour bucket. (For the record, about 22 percent of actual game players fall into the three-hour addiction bucket, which makes sense — only a very lucky few can spend the entire day gaming.*)
We’ve taken a look at addiction across verticals, and highlighted five in particular: Hardware, Enterprise, Social, Messaging and Media. All five illustrate unique patterns of user behavior, with Hardware, Enterprise and Social the starkest examples of how addiction can vary.
Social apps have a stable, consistent and thoroughly addicted user base, with 50 percent of people engaging with social networks for more than five hours a day, and even a small percentage logging time during every waking hour. This makes perfect sense, as social networks are increasingly becoming a primary way that people talk to each other throughout the day. For some, social apps have replaced email and telecommunications (if you happen to know a teen, ask when they last made a phone call). People don’t think twice anymore about checking email on an hourly basis, so it’s not surprising that the same would hold true for social networks.
While people seem to be engaging with social networks from breakfast straight through to bedtime, Enterprise apps are clearly used most during office hours. The majority of people using enterprise products use them for two to four hours throughout the day, with a minuscule percentage engaging for more than 10 hours. Enterprise products seem to be more necessary than they are addictive, which does align with expectation: Mixpanel is an enterprise product itself, and as addicted as we know our customers to be, there’s still a significant decrease in use after the nine-hour mark. That’s fine; once happy hour rolls around, most people aren’t thinking about their Funnel reports. (We know through Segmentation, however, that there’s a regular spike between 10 pm and 12 am, indicating that people check in on their reports before settling in for a good night’s sleep.)
Hardware had the most unusual addiction pattern, with a bump between 13 and 18 hours; the six-hour bucket and the 16-hour bucket are almost identical in size. At first glance, this seems odd, but it’s actually in keeping with what hardware apps and wearables are designed to do: Be on your body all day long. (This is the age of the quantified self, after all.) That said, patterns of addiction will reflect the device itself; there’s a difference between a fitness band that isn’t Internet-connected but syncs a few times a day, and Google Glass, which is Internet-connected all day long.
Somewhat surprisingly, messaging apps have the top 20 percent percentile of users engaged for more than five hours a day. That said, messaging apps are less addictive than those in the Social vertical, with about 80 percent of users only messaging three or fewer hours a day. This only serves to back up the argument that people are leaning on their social networks as an all-in-one communications center. (And it can’t strictly be because it’s easier to browse social networks on a computer than it is to text under the desk at work or school.)
Apps in the Media vertical are notably less addictive than their social counterparts; media apps do enjoy a few hours every day of high-volume usage, but users tend not to check back on them constantly — just over 80 percent of users are active for only three hours a day. To be fair, most media apps are used in a concentrated way; you wouldn’t drop in on a news app in the same way you might check your messages or social network updates. While users may read the day’s top stories on their commutes to and from work, this behavior would still fall into a two-hour bucket.
Addiction adds an extra layer of insight that allows companies to analyze user behavior on an even deeper level. One thing that’s clear is that addiction is inextricably linked to function: If your product is a social app that people don’t use more than once a day, that’s a red flag — and not one you would have previously been able to catch if you relied solely on Retention. (Similarly, if you work for a SaaS company and most users are lighting up your Addiction report by using your app for 10 hours every day, you’re doing something very, very right.)
Most importantly, Addiction allows companies to now see how embedded they are into people’s daily lives — which in turn allows them to finely gauge the immediate impact that product changes, new features and marketing efforts have on their customers.
* In previous reports, we’ve discovered that gaming is popular throughout the day — especially in the afternoon. What Addiction reveals is that while gaming is a popular vertical overall, it’s not the most addictive: The majority of unique users are gaming for three hours or fewer every day, rather than continuously throughout the day (as a regular trend line might suggest).
Suhail Doshi is the co-founder and CEO of Mixpanel, an advanced analytics platform for mobile and the Web. More than 1,800 mobile apps and websites use Mixpanel to analyze more than 17 billion actions every month. Follow him @suhail.