#Homescreen2014
What can 1000 homescreens tell us about the way people use their phones. Excerpted from the betaworks shareholder book
At betaworks we aim to build apps that people love: the essential apps that people use every day and that they obsessively want to have on the homescreen of their devices, one touch away. Yet, measuring progress against this goal is a challenge. We have internal analytics, tools, and KPIs that give us an indication of progress. We are obsessive users of Chartbeat, which we helped design specifically to track real-time social engagement. We use Twitter and social channels to measure the scale of engagement and its depth. Twitter is especially good at giving us a sense of depth: examining the language, the influencer clusters and the sentiment that people use to describe our work. When people talk about Dots as an obsession they love or a Tapestry story as something that moved them, we take these as indicators that we are accomplishing our goal. However, it’s just an indicator and the world we build in today is balkanized. More often than not, we can’t get enough visibility into many of the platforms on which we build experiences. Whether it’s the App Store, Facebook, Instagram, Snapchat, most platforms today are opaque in terms of metrics and data. But at the start of each year there is an elegant hack we apply.
Each new year, people share pictures of their homescreens on Twitter, Instagram and other social sharing platforms. If you search Twitter for #homescreen2014, you will see a stream of pictures of people’s homescreens — the primary screen of their phone with all the apps they choose to keep there. It is fascinating to browse through this stream of images — analyzing it is even more interesting. Right after the new year, we culled 1000 homescreen images from Twitter, cut up the images and tabulated the apps on the homescreens vs. those in folders. Admittedly, it’s a hack, and the sample is skewed: among all smartphone users, we’re biasing completely for people who use Twitter, and among Twitter users we’re selecting for the type of person who is willing to share a homescreen image. But, caveats aside, the data are fascinating. Eighty-seven percent of homescreens shared in our sample were iOS and 12 percent were Android (1 percent was Windows). For the sake of consistency, we focused the analysis below on iOS — the 87 percent.
The first metric that we pull from the sample is the percent of people who have apps we are developing at the betaworks studio on their homescreens. We then look at the investments we have made. These are our KPIs, so let me start with them and then offer up some data and perspective beyond betaworks.
Our results. Betaworks apps we are building at the studio are on 17.3 percent of people’s phones, up from less than 5 percent at the start of 2013. In terms of the investments that betaworks has made — that haven’t exited — they account for a further 15 percent. That is a significant jump in presence and share.
More often than not, we can’t get enough visibility into many of the platforms on which we build experiences. Whether it’s the App Store, Facebook, Instagram, Snapchat, most platforms today are opaque in terms of metrics and data.
In this picture you can see my homescreen. I’m right handed and I organize my homescreen with the app’s that I want to access with one thumb touch on the left hand column.
Here are some of the other results we drew from the analysis we did in January of 2014. Most everyone has either a texting/messaging app on their homescreen. Eighty-nine percent have the standard Apple texting app and 86 percent of people have the phone app. More people regard these devices’ primary use as messaging rather than calling and 14 percent of people in this sample don’t view their smartphone as a phone: it’s a computing device. Telephony is an app that doesn’t even make it to their homescreen. Alternative, non Apple, messaging apps are starting to appear on the homescreens of users and I bet that if we sampled just non US users the numbers would be much higher. Facebook is Messenger is on 14% of people’s homescreen, Whatsapp is on 12 percent, Snapchat is on 11%, Path on 5 percent (while snapchat and path arent straight messaging app’s, worth noting them here for comparison), Groupme 4.7%, HipChat on 2.6 percent, Line on 1.5 percent, Viber 1%, Kik is on 0.5 percent. What’s interesting is that these alternatives dont replace the standard texting app. As I outline below in many categories the standard Apple app is replaced by users on the homescreen, in messaging its complemented.
After texting and calling browsers, photos and camera are the next most popular apps for homescreens. Sixty-five percent of people have Safari on their homescreen, 18 percent have Chrome, Opera Mini is at 0.5 percent. For 7 percent of people the browser is so important, they have both. For 9.8 percent of users there was no browser whatsoever on the homescreen. 63% of people have the Apple Camera app on their homescreen. And what about the dock, the permanent menu dock on the bottom of the iOS interface? The most popular apps on the dock are mail, phone, and browser.
The degree that users are switching core Apple apps — calendar, email, tasks, notes etc. — with alternatives is something that interests me a lot. I discussed this in last years’ shareholder letter. Startups and companies other than Apple made significant inroads here in 2013. Fifty percent of people who have a mail app on their homescreens in the sample have a non-Apple mail app. For task-related apps the number is 57 percent, calendaring 46 percent, weather 44 percent, maps is 54 percent and for podcasting the number is 65 percent. The discovery process in mobile is still nascent, yet Apple has a huge advantage over other app developers by installing their default experiences and not letting carriers change them pre-sale. The large percentage of people going through the hassle of switching suggests that even in an iOS7 post-skeuomorphic world Apple’s apps are often not best in class.
Some specific findings by company:
Facebook is the leading non-default app on people’s homescreens. In total, 68.6 percent of homescreens had a Facebook app on them (either Facebook, Messenger, Pages or Instagram). This number is still remarkable but if you consider that more than 50 percent of DAUs for Facebook are mobile only, it isn’t surprising. The more important question for Facebook is, how is the unbundling experience going? In the sample set, 18 percent of devices had just Facebook installed, 12 percent had just Instagram. Thirty-six percent had Facebook and Instagram. Unbundled, single-purpose apps are the way to expand share on the homescreen and Facebook is pulling it’s core desktop experience apart as their audience moves to mobile. The acquisitions of Instagram and Whatsapp are part of this strategy, yet while this strategy makes sense, it comes with execution risk. Case in point: Poke. One person in the sample had Poke installed on the homescreen.
A full 62 percent of users have at least one Google app on their homescreen and the median Google user has 2 apps on their homescreen. The most popular by a long shot is Google Maps: about 42 percent of users in the sample had Google Maps on their homescreen (Waze is on an additional 5 percent). After maps, the drop off is steep — the next most popular Google app is YouTube, which is on 17 percent of homescreens, Chrome on 17% and then the standard Google app, which is on 11 percent. Google+ is on 8 percent of homescreens in the sample. That is higher than I expected, and it’s interesting given that the primary discovery model on the phone is still under development and it’s likely to be social: Google’s (and Apple’s) relative lack of presence in social is important. Google Drive was on fewer homescreens than I expected (4 percent) — given that Dropbox is on 16 percent. I wonder if the rebranding of Docs into Drive confused users. Owned and operated apps aside, it’s important to add that Google data underpins a lot of core experiences on mobile devices: web search, yes, but also mail, calendars etc. I.e., I would guess that most Mailbox users (now part of Dropbox) use Gmail IMAP — and Mailbox is on a full quarter of people’s homescreens in the sample.
Twitter related apps are on 85.5 percent of homescreens. Given that the sample was based on Twitter users there’s sample bias to the Twitter number, but despite that there are some interesting conclusions to draw out of the data. Seventy-nine percent have one Twitter app on their homescreen, 6.5 percent have 2 or more and 14 percent have none — presumably these users use Twitter via the browser or an app not on the homescreen. Vine is on 12 percent of people’s homescreens, which is impressive. But Twitter’s client app is only on 37% of homescreens and third-party clients are on a whopping 55 percent of devices, with one client, Tweetbot, making up a full 49.5 percent of the sampled homescreens. It’s remarkable that a non-Twitter owned client has more market share than Twitter’s client. It’s a byproduct of the early adopter sample bias, but I think it points to the fact these users — myself included — prefer using a different, and more advanced, workflow for Twitter.
Twitter continues to fascinate me. It is an app company with a primary experience that is mobile first. In another time, Twitter might have integrated the Vine service into its web site and mobile in a quest to become a bigger “portal,” but today it is operating it as an unbundled app. I suspect we will see more of those. In 2014, we will see if Twitter can extend the mobile experience and take advantage of the data messaging bus that underlies the platform, build out cards for atomic distribution of content and apps and monetize, as Wall Street expects it to do. It has a lot to do, but, in the long run, I’m long on Twitter. It’s platform is so different to others, so simple: it’s like the Rorschach test of software, simple and open to re-interpretation and adaption. And unlike many of the companies mentioned here Twitter has mobile DNA.
Yahoo apps (branded and non) are on about a quarter of homescreens sampled — but it is acquisitions that have driven this presence. Yahoo-branded apps are on only 11.3 percent of homescreens. Take out Yahoo Weather the number falls to 2.47 percent. Yahoo non-branded properties have a high presence: Flickr is on 6.9 percent of homescreens and Tumblr is on 9.8 percent. This illustrates just how important Tumblr is to the future of Yahoo’s mobile experience; prior to the Tumblr acquisition, the share that Yahoo had of the homescreens was similar to what we have at betaworks.
Microsoft-branded apps are 0 percent of people’s homescreens. Microsoft-owned (i.e., Skype) are on 11.5 percent. In the context of a different era, when Windows competed with OS X, this made sense. In the world we live in today, it doesn’t. It highlights just how irrelevant Microsoft has become for people whose primary computing experience is off the desktop. Furthermore if you consider that Skype’s traffic now represents almost 40 percent of the international calling business. The 12 percent number suggests that Skype is predominantly a desktop service.
Then there are other startups with large presence on homescreens. Evernote is on 24 percent, including mine. Evernote confounds me: on one hand, it’s a great service, on the other hand, when one of my colleagues last year said, “Evernote is the only Windows app on my iPhone,” the analogy made immediate sense. The app is just good enough. It often crashes or fails to sync, and I will be interested to see if it holds share in 2014. What about local? Foursquare is on a full 23 percent of homescreens. I know traditional marketing people will argue this sample is small, and skewed to early adopters, but in doing so they miss a fundamental aspect of networks. The most active users are the most densely connected members of networks and their participation and data isn’t merely an early indicator of success; it’s an unfair advantage. As Foursquare moves into passive recommendations and passive data collection it has the potential to become a very important data set and its presence on a quarter of homescreens speaks to the fact that for local information, for this sample, its brand dominates.
Then there is the list of incredibly well capitalized businesses and huge brands that are still predominantly web apps. AOL has almost no branded presence on homescreens. Its non-branded apps have a small presence, as Huffington Post and Techcrunch were on 0.74 percent of homescreens. Amazon’s position is small; its branded app is on about 3 percent of homescreens in the sample, while the Kindle app is on 6 percent. Pinterest on 6.1 percent of homescreens, eBay on 2.2 percent, Fab on 1 phone, Fancy on 1 phone, and Gilt is on no homescreens in the sample. Linkedin is only on 11 percent. You can argue that media and commerce and job search are just starting to shift to mobile, but given the installed base on mobile today, these companies have work to do, and when the shift happens it will happen fast.
Moving beyond a company perspective, a bit of data around important categories of apps.
Start with gaming. Interestingly, most users in the sample didn’t have any games on their homescreen. They don’t view them as primary to their workflow or they want to hide them away and avoid or guardrail temptation! One percent had Angry Birds on their homescreen and there was not a single instance of Candy Crush, Clash of Clans, or Temple Run. Our game, Dots, shows up on 2 percent of homescreens. In context of other games, that’s great — the only other games on homescreens were all on sub 1 percent of screens: Lumosity (0.9 percent), Zombies (0.7 percent), Minecraft (0.25 percent). Interesting.
Music is still a primary service that people have on their homescreen. The default Apple music app is on 49 percent, Spotify is on 21 percent, Rdio is on 17 percent, and Pandora is on 7.4 percent of homescreens. These are big numbers, bigger than I expected.
Newsapps have a strong presence on the homescreen, but branded News sites don’t. I suspect this is because many news sites haven’t invested in app experiences, they are focussed on the mobile web and some news sites live within apple’s newsstand app. On the Newsapp’s side Flipboard and Reddit are on 10% of people’s homescreens, our Digg app is on 3%, Zite is on 2.7%, Prismatic and Pulse on 0.7%. CNN and Huffington Post are on 0.74%, The NYT is on 0.5% of homescreen’s and Apple Newsstand is on 3.9% of people’s homescreen.
What about payments? Similar to gaming, commerce isn’t core to the workflow that people have on their phones, yet. Venmo (formerly a betaworks investment) is on 2 percent of homescreens, Paypal is on 1.3% and Square’s app (Cash) is on 0.5 percent of homescreens sampled.
There are 3.5-4bn phones in the world today and 1bn of them are what we call smartphones. If you consider that the replacement rate of these devices is approximately every 2 years, it suggests that sometime in the next 2-3 years we are going to surpass 3bn+ smartphones worldwide. And some time early this year, the installed base of smartphones in the world will exceed that of PCs. These mobile devices are swiftly becoming, for most people, their primary computing experience, and the homescreens are the hubs of primary navigation for users. Yet, as this analysis confirms, its a misnomer to call these devices smartphones. The phone is an app — these are computing devices. This sample data says to me that as users move their primary computing experience to a smartdevice, it’s opening up opportunities for startups to re-define those experiences. Call it mobile first, mobile product design or unbundling — app user after user and developers alike are seeking to re-define computing by building products and services that are powered by smartdevices, are data rich, cloud-connected and as simple as possible.
#homescreen2014 was written for the 2014 betaworks shareholder book, the post above is an excerpt.