Evaluating Google’s I/O 2015 announcements

I wrote a piece a few days back for Techpinions about the challenges Google needed to address at I/O this year. Now that the I/O keynote is over (which I was fortunate enough to attend in person), I thought I’d revisit that list of challenges to see how Google did.

Retaking control of Android

There were a couple of things that Google did related to retaking control of Android at I/O: reinforcing the value proposition of its Android One initiative, and announcing Android Pay. The former is obvious: it’s Google’s attempt to get a close-to-stock version of Android as the default version in emerging markets. But the latter may be less obvious. However, by launching Android Pay, Google nixes its OEMs’ efforts to introduce their own payment systems, as this post outlines. Samsung is likely to be the hardest hit by this, since it’s the only Android OEM that had launched a payment service. But I wonder how much Google will spread this model of Google Services Mandatory classification for key Android features to other areas, squeezing out OEMs from offering competing services.

Defending the web against apps

This year’s I/O was a mixed bag in this respect, with a couple of initiatives clearly aimed at just this, but some others moving things in the opposite direction. On the one hand, Google announced Chrome Custom Tabs, an alternative to in-app browsers for displaying web content, which brings users back into the web, where Google can better track their activity and otherwise capture data. It also reiterated its app indexing and deep linking projects, which recently began to roll out on iOS too. These efforts are both aimed at making the web more relevant in a world where apps are becoming dominant. Google Now on Tap is an interesting mix in this respect – on the one hand, it allows users to stay in apps when they have Google Now queries, rather than having to exit out of them, which could be seen as favoring apps rather than the web. But on the other hand, this inserts a Google layer between users and apps, in some cases recommending other apps for users to open, but in others using the Google Knowledge Graph to disintermediate those apps.  On balance, Google seems to be serving its own needs pretty well with these new announcements.

Convince developers Android users are worth targeting

There was remarkably little at I/O about why developers should target Android users exclusively or in addition to iOS users. There was no update on the total number of Android users, which remains at “over one billion”. And Google did very little to argue why these users might be attractive, rather emphasizing the fact that many of the new users on Android will come from emerging markets, where there are lower incomes and less propensity to spend. There was no mention of carrier billing at all, and the only mention of monetization was in relation to better ad products within apps rather than paid apps or in-app purchases. I don’t think Google has given up on the paid apps route, but it was hard to escape that impression from the keynote, which I find baffling. To be sure, Android will never have the same attractive demographics as iOS, but it can still do much better than it has in the past, especially in mature markets.

Take Android beyond personal computing devices

The big announcement here was the Brillo project, which takes Android and strips it down to a barebones version for use in Internet of Things devices (and for today at least home devices specifically), together with the Weave communication protocol, which will be baked into Android at the Google Play Services layer and therefore make compatible devices instantly discoverable from Android devices. This is exactly what I was getting at in my preview piece on this topic – a version of Android that’s optimized for these devices, which have no need for the full version of Android but have other specific needs Android in its current form doesn’t meet well. There’s a lot of work still to be done here – though Brillo will launch in Q3 (and Weave in Q4) I sensed many decisions about Brillo still haven’t been made (not least the final name for the OS itself). More broadly, I was disappointed that we didn’t hear more about operating systems for the car, another area I highlighted in my preview piece, and one in which Google has been reported to be making some interesting progress. I wonder if we’ll see more on this in the months building up to the Android M public release.

Demonstrate a clear value proposition in TV

This was the other big area where I was expecting much more from Google’s keynote at I/O than we got – it got barely a mention in the keynote, and the expected announcements were made either not at all or in press releases. Nvidia announced its Shield device built on Android TV, and bought up almost every outdoor advertising spot within a few blocks of the Moscone Center to advertise it, but there was nothing in the keynote to demonstrate meaningful progress in this area. The only other concession to this challenge was the announcement that HBO Now would be coming to Android and Chromecast shortly. Speaking of which, Chromecast has now sold 17 million units, putting it in the same ballpark with Apple TV and Roku. There continues to be an odd disconnect between Chromecast and Android TV in Google’s TV strategy which I hope it can resolve in time. Chromecast certainly seems to be the more successful model so far, though the total number of casts (1.5 billion) relative to the number of Chromecasts sold (17 million) implies relatively low usage of those devices.

Continue to unify Android and Chrome OS

There was even less on this theme at I/O 2015 than there was last year, where Google at least talked about Android apps running on Chrome OS. Chrome OS was barely mentioned at all in the keynote, and there was no news about it at all. This continues to be an area where Google has to do a much better job telling its story and bringing the disparate threads together.

Differentiate against Amazon and Microsoft in the cloud

The whole enterprise space, and cloud in particular, got incredibly short shrift in the keynote too, after a lengthy session in last year’s keynote. There was nothing here to indicate that Google was going to make meaningful progress in this department in the coming year beyond its existing strategies.

Beyond the keynote, and beyond Android

Even though the pace of the keynote and the volume of individual announcements was somewhat overwhelming, as always, the substance of what was announced really wasn’t. The Android M release looks like a very modest, incremental, improvement on Lollipop. This was a little disappointing, but partly reflects the fact that Google is slowly extracting functionality from the operating system and putting it in more easily upgradable layers like Google Play Services instead. The new Photos app is a great example of this. This does mean, however, that the news around new Android version releases is going to become less interesting over time, while the announcements separate from Android will become more interesting in many ways. But this year’s I/O also looks like introducing much more news outside the keynote – I’m writing this on day two of the event, and several announcements around wearables, Google Loon, and other projects have been made separately. As an attendee, I’m grateful that the keynote was trimmed down to just two hours, but it does make it harder to follow all the news that’s coming out of I/O as Google starts to fragment the more notable announcements across sessions.

Jony Ive’s promotion

Yesterday, CEO Tim Cook announced to Apple employees that Jony Ive had a new title and role within the company. The change has been described as a promotion by Apple, but I’ve seen some skepticism about this characterization externally. I have no inside information here, but based on what I’ve seen from Apple and others, I had a slightly different take on what’s going on here.

Freeing up a creative person to be creative

I’ve been an industry analyst for 15 years at this point, and I spent the first 13 years of that time working within a firm of industry analysts. One of the things that I learned during that time was that the best analysts often made the worst managers, and vice versa. It was a rare thing to find a really good analyst who both enjoyed being a manager and was good at it. Why do I mention this? Well, creative work and management take fundamentally different skill sets, and they’re rarely found in the same individual. And yet, the longer we’re in our professions, and the more senior our job titles become, the more likely we are to be placed in roles where part of our job is managing other people.

I see no evidence that Jony Ive is a poor manager, but the reality is that he’s first and foremost a creative person, a designer, and not a manager. By virtue of his longevity and seniority at Apple, he’s ended up managing the design team, but I would suspect that this isn’t how he wants to spend much of his time. Ive can’t simply be off in a room by himself designing things if he’s going to continue to have the lead design role at Apple, but he can be freed up from some of the day-to-day management, delegating a lot of the minutiae to others in the team, and especially to his two lieutenants, Howarth and Dye. This delegation is all the more important given how many new design projects Ive is taking on, in relation to the retail stores, the new campus, and who knows what else. As others have pointed out, the recent profiles of Ive have portrayed a man stretched very thin, and I would guess that, given the choice between yielding some management responsibility versus yielding design projects, he chose the former.

Does this mean he’s staying, or going?

As I mentioned at the top, some have seen these changes as a sign that Ive is moving on. To my mind, there are two possibilities here: either Ive is now going to be rejuvenated by relinquishing some of his responsibilities and will gain a new lease of life, or this represents his last hurrah, with redesigning the stores and designing the campus as his valedictory efforts before moving on. Perhaps even he doesn’t know yet – maybe the change in his role is an attempt by the company to find a way to keep him happy even as he’s feeling the pull back to England and out of a day-to-day role at Apple. And maybe neither Tim Cook nor Jony Ive nor anyone else knows whether this will work yet. But I suspect (and hope) that it means Ive will be free to focus on what he does best – design – and that we’ll get lots more good stuff from him over the coming years.

One quick postscript: I’ve seen a couple of people suggesting this new role for Ive is a sort of Jobs 2.0: consolidating power at the top of Apple. But I think that’s the furthest thing from the truth. Whereas Jobs was the consummate micro-manager, injecting himself into every detail of Apple’s operations, I suspect Ive’s temperament is the opposite. Yes, he obviously cares deeply about every aspect of design at Apple, but in the profiles that have appeared over recent months, it’s become clear how many other people already have a hand in the details of design, and it’s clear that Ive has little interest in management beyond the realm of design. In some ways, Ive is the anti-Jobs, focused on his particular sphere of expertise with little interest in going beyond it.

Why an Apple television doesn’t make sense (and does)

It appears some sources at Apple have this week indicated to Daisuke Wakabayashi at the Wall Street Journal that Apple is no longer actively working on making a television. This doesn’t surprise me in the least – the project never really made sense to me as I’ve repeatedly written and told reporters over the past several years. It may seem like odd timing, but I thought I’d outline my thoughts as to why this is so, and at the end talk briefly about a couple of reasons why it does make sense.

Cost, margins and differentiation

If Apple did make a television, there are several things we can be fairly sure of: it would make it out of the same premium materials as almost everything else it makes, and it would want to make sure margins on such a product were in line with the rest of its product line. The challenge here is that Apple would be starting at a very small scale, so would enjoy none of the benefits of economies of scale that current TV makers have, and current TV makers already operate at razor-thin margins. Consumer electronics generally is an incredibly low margin business – single digit operating margins are the norm when companies make any money at all. For Apple to come in, raise the cost significantly because of both premium materials and its lack of scale, and then to try to recoup its supra-normal margins too would drive a price at least twice as high as televisions with similar specs, if not significantly higher. And of course we have a precedent for this in similar products: Apple’s 27” Thunderbolt display retails at $999, while Dell’s equivalent product retails for $599, Asus has one for $430, and low-cost brands go significantly lower. (I’m even completely ignoring, for now, the emergence of 4K televisions – which would magnify all these issues significantly, putting an Apple television into the stratosphere in TV pricing terms).

So why couldn’t Apple do this again in the TV space? To my mind, it comes down to differentiation. The Apple display is differentiated at least in part on the basis of its materials and its look. Arguably, the presence of the Apple logo is also a great signal in a workspace that this is a premium product – for the kinds of creatives who are likely to use these displays, this is an important signal to clients and others about the kind of work they do, and the products they use to get it done. But think about TVs and how they’re evolving. They’re mostly either attached to walls, on stands up against walls, or hidden away in cabinets much of the time. Bezels are shrinking and even disappearing. The prominent logos which once sat under the screen are disappearing with them. To a great extent the television is becoming the purest version of the black rectangle in our increasingly black-rectangle-filled lives. How would Apple differentiate on hardware here? Would it turn back the clock and increase the size of the bezel? Would people even notice if the tiny bezel were made of aluminum instead of black plastic? Would they care? Differentiation in TV hardware today is primarily about making everything but the screen disappear, and this seems totally at odds with Apple’s hardware differentiation.

How, then, to convince customers to part with double or more what they’d pay for an equivalent TV from competitors when the differentiation in hardware will be largely invisible? One option, of course, would be to add additional functionality to the hardware – a camera and microphone for FaceTime calls, for example, with the microphone doubling as an enabler of Siri for the TV. But these things have been tried and failed – FaceTime on personal devices works, but no-one has ever been able to convince families that they should be paying lots of extra money for a TV they can use as a videophone. It appears from Wakabayashi’s piece that Apple did indeed tinker with some of these things, but clearly concluded much the same thing.

Integration vs. a single input

The other way Apple could have differentiated a television is through software, and of course the vast majority of Apple’s products do differentiate through a combination of beautiful hardware tightly coupled with easy-to-use software. So, how would Apple differentiate an Apple TV through software? Well, the problem here isn’t so much that Apple couldn’t do this, but that if all the differentiation is in software, why can’t it be fed to the TV from a companion box like today’s Apple TV? What’s the difference, ultimately, between software baked into a TV and software baked into a box which directly connects to the TV? The challenge with companion boxes and traditional pay TV set top boxes today is that you often need more than one of them to meet your needs. TVs (and accessories such as receivers) come with more and more HDMI ports to cater to the range of devices the average individual or family wants to connect to them: pay TV set top box, Blu-Ray player, game console, a streaming box or stick, and so on. In such a world, it’s easy to imagine an Apple television providing a better way to manage all these inputs in a way a companion box simply can’t solve.

But what if Apple’s vision for the TV space involves more than just being another input plugged into another HDMI port? What if Apple’s plan is to take over the HDMI1 slot and convince you to dump all the other boxes you have historically plugged into your TV? To be clear, this is exactly the strategy I expect Apple to pursue with a revised Apple TV box and the Apple TV service. Under this scenario, input-switching goes away as a problem, and there’s very little meaningful difference between an Apple television and a generic third-party television fed by an upgraded Apple TV box. The only real differences are the need for two remotes and the lack of any audio integration with the TV hardware for Siri and other related functions. Both problems could easily be solved with the use of a better remote for the Apple TV, acting as both a universal remote and as an audio input device (much as Amazon’s Fire TV remote does).

The addressable market

The third reason why an Apple television makes far less sense than an upgraded Apple TV box is the addressable market. Were Apple to sell TVs, it could only target those willing to swap out whatever television they have for a new one, and at a significantly higher price than they’re used to paying. However, an Apple TV box, at a fraction of the price, has a significantly lower ASP but a vastly bigger addressable market – anyone who has any HD TV today and sees the value in adding an Apple experience. Now think about the potential revenue stream from an Apple TV service tightly bundled into the Apple TV box, and suddenly the overall addressable market and the associated revenue becomes significantly larger for this combination than for a television set. Factor in refresh cycles for televisions and the effect is magnified still further – a single purchase every 5-10 years versus more frequent upgrades on hardware and monthly recurring revenue from TV services becomes a no-brainer.

The counter-argument

Having spent most of this post talking about why a television doesn’t make sense, I’d like to briefly review a few reasons why it might, despite all these objections:

  • Control and integration: Apple’s standard model for product development is to approach hardware and software hand-in-hand, and create complete, end-to-end experiences. The current Apple TV flies in the face of this model, because it sits in the background behind a TV built and branded by someone else. An Apple television would be much more along familiar lines, tightly integrating hardware, software, and services, and creating an end-to-end Apple product.
  • Feeding the base: the reality is that many of Apple’s most ardent customers, who likely view Samsung as an inferior brand, nonetheless have Samsung TVs in their living rooms. For those used to buying high-end, well-designed hardware that works together seamlessly, having a relatively inferior product as one of the most visible pieces of consumer electronics in their homes may be irksome. Feeding the Apple base by providing them with an Apple product for this prominent piece of hardware must be tempting. There are no doubt those who would pay the massive premium to have an Apple television set, even if the total number is small.
  • Shutting out others: as long as Apple only makes a companion box, its role is essentially the same as other boxes plugged into the TV, and it has no control or leverage over them. With both the pay TV set top box and the television itself getting smarter and incorporating more functions, there’s a risk that the Apple TV slowly gets pushed out. But turn the model on its head, with Apple making a television, and suddenly Apple is the one calling the shots. It could gain huge leverage over the pay TV providers and how their content shows up on the television, for example.
  • Visible differentiation: one of the interesting things about the Apple TV is that it’s the only one among Apple’s product line today that’s made substantially out of black plastic rather than its usual premium materials. The reason for this is simple: it’s far cheaper, and the device in many cases will be hidden away in a closet or TV cabinet, especially when not in use. A television set, however, would allow Apple to be far more visible in the living room.

I don’t think any of these today come close to overcoming the objections I outlined above, but I can see why Apple at least wanted to explore the category for these reasons and others. Over time, it’s possible that the relative dynamics I’ve outlined above could shift such that the reasons for making a television start to overpower those for holding back. But for now I’m confident Apple has made the right decision.

Why Verizon’s AOL deal makes sense

Note: I’ve added an addendum at the bottom of this post about the content angle specifically.

Verizon buying AOL isn’t a complete surprise – there were rumors of a deal back in January, but now they’ve been confirmed. But it also makes good strategic sense for Verizon as part of a broader strategy that’s been emerging for some time now.

Verizon’s traditional business

Verizon, of course, has its roots in a very different business – landline and wireless telecommunications services. Those make up the vast majority of its revenues today, and its landline business has been going through an interesting transition recently (as I outlined in an earlier post), and in the landline business this means shifting the base from old copper lines to fiber. To be clear, there’s growth left in this business, but it’s highly dependent on the combined value proposition of broadband and pay TV.

A shift in TV viewing requires a hedging strategy

However, the writing is increasingly on the wall for traditional pay TV – disruption is coming, and this quarter marked the first time in recent memory that the major pay TV providers actually saw a decline year on year in TV subscribers:Screenshot 2015-05-11 12.31.47

Verizon recognizes this, and in a previous post I wrote a little about Verizon’s response to this threat to the pay TV business. That response takes the form of a hedging strategy, which allows it to take advantage of the video business even if the traditional pay TV side begins to suffer. I outlined in that post a three-part strategy here (quoting now from that earlier piece):

  • Sell classic pay-TV services to as many of the 15 million households where FiOS is available as possible. It’s only sold about a third of those so far, and the number is creeping up pretty slowly, so though there’s growth left, it’s not going to be huge. Competition from cable and satellite remains fierce, so there’s incremental growth here at best. Verizon will continue to evolve the offering here to make more and more content available on more and more screens, but this will be largely a competitive differentiator rather than a source of significant revenue growth.
  • Sell a range of wholesale content delivery and related services to third-party content providers like HBO. This is unbounded by the FiOS footprint or even Verizon’s overall broadband footprint, so it could grow significantly from where it is today, though it will always be a much smaller market than consumer video services.
  • Sell over-the-top video services to consumers independently of the FiOS offering. This is exemplified by Redbox Instant today, but could well expand into something more. Verizon already has great relationships with all the major content providers through FiOS, and through broader licensing agreements could easily create a sort of unbundled FiOS TV offering to be sold nationwide.

Where AOL fits

AOL fits firmly in the third pillar of this strategy, but also broadens it in several ways:

  • It takes the over-the-top content strategy beyond video into news and other forms of content, including Huffington Post, TechCrunch, and so on (it remains to be seen how AOL’s news-focused employees respond to the acquisition, especially given the strong negative reaction to Verizon’s own “news” site a few months back)
  • It takes Verizon beyond subscription content and heavily into the advertising sphere, which both provides a more varied set of revenue streams around content but also offers opportunities to provide converged advertising campaigns, retargeting and other attractive elements of a multi-screen advertising platform. It’ll take time to build these linkages, but in time they could be quite powerful for advertisers (see the Cablevision ESPN deal this reported this week).
  • It takes these content services beyond the Verizon brand – though Verizon has a national brand, it’s not associated directly with quality content, and though it owns FiOS and therefore a video service, it’s not national. It also doesn’t have a position yet in shorter-form video content. AOL extends it into some of these new areas, and using a different brand that may be more familiar to some potential customers.

The AOL brand

Now, the AOL brand is in some people’s minds forever going to be associated with yesterday’s technologies (my wife’s first reaction to hearing about the deal this morning was “what does AOL do anymore? I just think of “You’ve Got Mail”). But the reality is that AOL remains one of the top online brands in the US in particular, and one of only a handful of companies that reaches over half the US online population with its content each month. AOL, both through its own brand and through powerful sub-brands such as Huffington Post and TechCrunch, is a much more powerful content player than Verizon on a national and especially international basis than Verizon. Yes, there’s some baggage that comes with that, but for the most part the AOL brand family is a great boost for Verizon.

Risks and uncertainties

Despite the strategic sense behind the deal, there are of course risks and uncertainties. There will undoubtedly be a strong culture clash between the two companies – Verizon’s an enormously conservative company in many ways, and although there are pockets of startup mentality, it’s far from the norm across the company. And it’s an absolutely enormous company, much closer in size to AOL Time Warner in its heyday than AOL today. It will be easy for AOL and its culture to be lost or squashed in the course of the acquisition. And although there are lots of synergies (including those described above) on paper, the devil is in the details and it will take a lot of work to identify how these pieces really come together in practice to provide value for the combined company, its customers, its advertisers, and its shareholders. However, on balance I think it’s a good thing that Verizon is willing to take risks as it seeks to navigate uncertain waters in the TV and video space, and this is certainly a much bigger bet than its peer AT&T’s joint venture with Chernin around Otter Media. That, of course, could end up being brilliant or very costly.

Addendum: not just an ad tech play

I’m seeing a lot of people assuming that this deal is entirely about advertising and/or ad tech, and that the content side is either incidental or will be sold off down the line somewhere. However, I’m not as convinced about that, or that the content business at AOL looks the same under Verizon ownership as it did on a standalone basis. Here’s why:

  • Verizon is building up to the launch of an over-the-top mobile first subscription video service in the summer. Huffington Post content in particular seems like a great fit in such a service (it may well have been one of the content partners even without an acquisition). With over a hundred million of its own wireless subscribers to market such a service to, Verizon has a great new shopfront for some of the AOL content
  • Verizon also provides FiOS within its landline footprint, and some of AOL’s video content could be a great fit there too, as part of traditional bundles, as part of Custom TV (should it survive legal challenges), or in some other form.
  • Verizon’s insight into its own users and the 70% of Internet traffic that traverses its network at some point could also allow it to add a very important targeting layer to AOL’s advertising around its own content – the challenge for all online advertisers is how to extend their reach beyond the sites they own – Verizon provides a significant depth of insight about users AOL could never glean itself. This goes the other way too – Verizon can gain insight about users from the time they spend on AOL’s content, even if  they can’t monetize that usage directly.

In short, I think AOL’s content businesses have a better shot and a better role under Verizon than they did at AOL, and I’m not convinced they’re just going to be spun off once the merger closes.

Thoughts on the Fitbit IPO filing

Fitbit, the company that makes a variety of fitness trackers, has filed for an initial public offering with the SEC, and its S-1 filing contains lots of financial and operating data for the past several years. There’s plenty here to dig into, but I want to focus specifically on the user numbers and what they suggest about abandon rates of these devices.

Revenues and margins

The first thing to note is that the headline financials look extremely good. Revenue growth is very strong, especially over the last couple of quarters (note that this is 4-quarter trailing revenue, to smooth out cyclicality):

Trailing 4 quarter revenue $mAs you can see, Fitbit’s revenues have risen from just over a quarter of a billion dollars per year in 2013 to almost a billion in the last four quarters, which is phenomenal growth. And it’s doing this without losing money – in fact, it’s enormously profitable (note that these margins are adjusted for the negative impact of the recall of the Fitbit Force in late 2013):

Fitbit adjusted marginsAs you can see, the company has very healthy net, operating and gross margins, which show no signs of falling. These revenue growth and margin metrics help to explain why the company is going for an IPO now – the numbers are very, very good. I would suggest that the launch of the Apple Watch also creates a trigger for this event: it both brings welcome attention to the sector, while threatening the concept of dedicated fitness trackers, so now is in some ways the perfect moment to IPO, while the sector is hot but before Apple’s entry causes problems. Perhaps even more importantly, the sector is only beginning to feel the effects of the Shenzhen ecosystem, and Fitbit today still clearly commands a significant price premium for its devices, one that will be increasingly difficult to maintain as cheap Chinese trackers enter the market.

User numbers and abandon rates


To my mind, however, the various user numbers Fitbit reports were far more interesting, especially for what they suggest about how long people use Fitbit devices for after they buy them. Fitbit reports two different user numbers: registered users (reported on a fairly patchy basis) and paid active users (PAUs). The latter number is not quite what it sounds like, and that’s important here. Based on the definition in the S-1, a user only has to meet one of these three criteria within the preceding three months to qualify:

  • “[have] an active Fitbit Premium or FitStar subscription”
  • “[have] paired a health and fitness tracker or Aria scale with his or her Fitbit account
  • “[have] logged at least 100 steps with a health and fitness tracker or a weight measurement using an Aria scale.”

In other words, this is really just a measure of active users, incorporating those with paid subscriptions, those who have recently activated a device, and those who have recently used a device (with no double-counting). The only users that are excluded would be those who only use Fitbit’s dashboard without also using a Fitbit device (e.g. those manually entering activity or calorie consumption data).

Registered users vs. device sales

Let’s start with registered users (which is not defined but which I assume simply means those that have created an account at some time in the past). The first interesting comparison is looking at registered users against the total number of devices sold over time:

Registered users vs cumulative sales

What you see here is that the total number of registered users tracks very closely to the number of cumulative devices sold. In some ways, that’s not terribly surprising, but of course one important conclusion is that very few of Fitbit’s users have ever purchased more than one device. The difference between the two numbers at the four points in time we have available grows from under 500,000 to around 2 million over time. That’s probably not a bad proxy for the number of people who have bought more than one Fitbit device over time (though it’s not perfect – some may have bought more than two). That’s actually very small in the grand scheme of things – only about 10% of the total number of devices would have been sold to people who already had one.

Registered vs. active users

Let’s turn now to comparing registered and active users – if registered users are those who ever had a device, and active users are those still using one (plus the small number using a paid subscription), then this is a good indicator of the abandonment rate:

Registered paid users and cumulative salesAs you can see, the number of active users is far smaller than the number of registered users when added to that same chart. In the chart below, I’ve shown PAUs as a percentage of registered users at the four points in time where we have both numbers:

Paid users as percent of registered usersThe number bounces around at about 50%, rising or falling a little over time but remaining remarkably constant. In one sense, that’s obviously fairly bad news – in addition to the fact that very few Fitbit buyers purchase a second device, it would appear that half of those who bought one stop using it after a period of time. However, there’s a flip side to this, if you’re looking for a silver lining, which is that the number isn’t falling over time. In other words, over two years ago, the number was 50%, and it still is. I’m actually a bit surprised by this, because all the early abandoners should still show up in the numbers and drag the overall retention rate down, but that doesn’t seem to be happening. What’s interesting is that this correlates closely with a survey I did last year about fitness trackers. The key question here was the individual’s experience with fitness trackers:

Fitness tracker survey

As you can see, the 50% abandon rate suggested by the Fitbit numbers closely mirrors the results of the survey, with a 9%/11% split between former users and current users.

Recent device sales and active users

Another way to look at all this is to compare device sales and paid users (which was the first thing I did when the S-1 was released). We’ve already looked at the relationship between cumulative device sales and the active user base, but let’s drill down a bit. The chart below compares PAUs (in blue) with device sales over the prior 6, 9, and 12 month periods:Measures of base sizeThe line that correlates most closely with PAUs is the 6-month device sales number. This suggests that this may be a good estimate of how long people hold onto their Fitbit devices on average. That doesn’t mean every user abandons their device after six months – some clearly hold onto them a lot longer, while others likely abandon them more quickly.

What does this mean for Fitbit’s prospects?

We talked at the beginning about how well Fitbit is doing financially. It’s selling over 10 million devices per year at this point, growing rapidly, and making good margins on them. So, how important is this abandon rate information to our evaluation of Fitbit’s prospects going forward? Well, one could argue that at just 10 million sales per year, there’s tons of headroom, especially as Fitbit expands beyond the US (the source of around 75% of its revenues today). But in most consumer electronics categories, there’s a replacement rate for devices, which continues to drive sales over time even as penetration reaches saturation. The biggest worry in the data presented above is twofold: one, very few Fitbit buyers have yet bought a second device; and two, many don’t even use the first one they bought anymore. Once Fitbit maxes out its addressable market, it’s going to have a really tough time continuing to grow sales.

This, taken together with the threat of Chinese vendors invading the space with much cheaper devices, reinforces my perception that Fitbit is IPOing at the best possible time from the perspective of its existing owners and investors, but that its future looks much less rosy than its past.

The changing devices on the US mobile networks

For much of the last three decades, the terms “wireless carrier” and “cellphone provider” have largely been synonymous.  Your wireless carrier was the company that provided the service for your cellphone, and in the vast majority of cases also sold you that cellphone, often at a heavily subsidized rate. However, the reality is that the wireless carriers in the US are becoming much more than just cellphone providers, and in fact cellphones are becoming a smaller percentage of devices on the major US mobile networks over time.

Phones are growing, but not as fast as in the past

The total number of phones on the four major US mobile networks (AT&T, Sprint, T-Mobile, and Verizon Wireless) continues to grow, but the rate of growth has slowed. Whereas in the past phone subscribers grew by double digits every year, now they grow by single digits at best, and the reason is obvious: the vast majority of people who will ever own a phone already have one.

Screenshot 2015-05-05 15.05.47The small amount of growth there is in phones comes in large part from the generational effect of new users emerging at the younger end of the population scale, combined with a few laggards and late adopters among the older generations. Continue reading

Thoughts on Comcast’s Q1 2015 earnings

Comcast’s results today were notable for a significant transition that’s about to happen: this was likely the last time Comcast will report more TV customers than broadband customers:

Screenshot 2015-05-04 10.57.34As of the end of March, the two numbers were just 6,000 apart, so next quarter the two will have crossed over, and Comcast will have more broadband than TV subscribers. This is a hugely symbolic moment for a cable company, but Comcast is far from the first to make this transition – in fact it’s the last of the major public cable operators to do so:Screenshot 2015-05-04 11.10.55 Continue reading

A week with the Apple Watch

I’ve now spent a week with a couple of Apple Watches. The first of my pre-orders to arrive was the one I ordered with my wife in mind, a 38mm Sport with a white sport band, which arrived on the 24th itself. The one I ordered for myself arrived this week, and that’s a 42mm Watch with stainless steel case and black sport band. I’ve worn that for the past two days, but wore the other for several days first. I don’t normally do reviews here, and goodness knows there are plenty out there from people who do, so this isn’t a review in the “should you buy it?” sense, but rather a set of observations from my use of the Watch.

This is going to be a significantly longer post than I’m used to writing here, so bear with me, and feel free to skim through using the headings as a guide to what’s in each section. Continue reading