Short Note on App Store vs. Google Play Revenue

As Benedict Evans wrote in his weekly newsletter, Apple reported its revenue from the app store in 2014.

Apple reported that app store revenue net of its 30% commission was $10bn in 2014. This is actually the same as the figure it gave at WWDC in the summer for the previous 12 months – I suspect there’s some rounding going on. Apple’s commissions on this would be about 2% of CY 2014 revenue.

Apple announcement link

Although we will not know exactly what the rounding error was, given the very strong growth of the app store in the past, it is quite difficult to believe that sales have suddenly flattened out. So to get a better picture, I decided to provide an update to my series of articles on the iOS App Store vs. Google Play (most recent one is here).

In that article, I lamented that App Annie was not giving out the actual growth of neither the App Store nor Google Play and as such, it was impossible to see whether the revenue gap was closing in absolute terms.

Since then, App Annie has reported its Portable Gaming Spotlight, 2Q14 report in conjunction with IDC. This report gives us a direct answer.

Screen Shot 2015 01 13 22 12 07

As you can clearly see, iOS App Store absolute revenue growth is larger than Google Play absolute growth. Although Google Play is doing remarkably well on a YoY percentage basis (>+100%) compared to iOS App Store (+70%), it is not closing the gap at all.

Also you have to keep in mind that Google Play revenue is increasingly games. In fact, App Annie notes in their Q3 2014 report that, “Games drive nearly all Google Play Revenue Growth”. This is presumably not the case with iOS which suggest that the gap between the App Store in Google Play revenue is wider and also growing if we include non-game apps.

Actually, my view is that the difference in non-game app category is more interesting than the total difference. It’s likely that by now, iOS has about 6-times more non-game revenue than Android. This means that apps like Byword, RunKeeper and others which are either paid-apps or in-app payment apps are struggling on Android.

Overhyping the Revenue Potential of the Emerging Countries

A beautiful infographic on the state of App Stores by App Annie.

Unfortunately, I can’t agree with the title “Mobilizing the Next 5 Billion”. If anything, this infographic tells us how dominant the “app store superpowers”, Japan, South Korea and the United States are. These superpowers are not only dominant in current revenue, but also have revenue growth that is equal to the emerging countries. This means that the emerging countries are not catching up; instead the lead of the superpowers is widening.

The infographic tells us that the superpowers will remain dominant in revenue for the foreseeable future. The next 5 billion is unlikely to contribute significantly to total revenue from app stores for quite a while.


Infographic GMIC 11182014 2

Material Design and Opaqueness

I’ve been playing around with Material Design for a few days, and the most striking thing for me is the opaqueness. This is in stark contrast to iOS 7 and iOS 8.

Since iOS 7, Apple has gone to significant lengths to make the toolbars at the top and bottom of the screen translucent. To do this without decreasing the legibility of the frontmost element, Apple uses a frosted glass effect which is apparently quite computationally expensive and is not used in less powerful devices (for example, the iPad 2 reverts to simple transparency for some elements). Apple has mentioned that the goal was to make the underlying content stand out and to minimize distraction from the control UI elements. You also see a lot of iOS apps using gray toolbars, which also helps to emphasize content instead of the app.

On Material Design however, they do not use translucency anywhere. Everything is an opaque surface, just as paper is opaque. Although Material Design emphasized depth and the z-axis, this is conveyed solely through the use of shadows.

The feeling of opaqueness is further fortified by the use of bold colors for UI controls. Whereas iOS apps often use only a humble gray color as the background for the UI controls, Material Design encourages bold colors. The effect is that the toolbar and buttons make a very strong impression and feel omnipresent, a sense of opaqueness.

The new iOS design and Material Design feel very different. iOS feels very light. Material Design looks a bit heavy on the top.

Although opaqueness itself is neither good nor bad, in terms of the history of computer UI design, it does feel a bit backward. It feels like going back to the Windows XP days since every PC-OS since then has used translucent elements. The bold colors of Material Design are also reminiscent of the vivid toolbars of that OS, which fell out of fashion since. Similarly, iOS moved from an opaque design to a translucent design. I am wondering if the absence of translucency in Material Design is a result of having to make it work well on cheap devices which don’t have good GPUs.

iOS WWDC 2014 app

Note the translucency on the leftmost image, on the bottom toolbar.

NewImage

Google I/O 2014 app

Running on KitKat but using Material Design.

NewImage

Ad Spend Versus Time Spent on Mobile

Flurry Analytics released a report titled “Apps Solidify Leadership Six Years into the Mobile Revolution” back in April 1st of this year which I hadn’t analyzed in detail. Looking back, I certainly should have; It is a trove of information.

Here, I would like to look at the final graph that they show.

NewImage

Here the assumption is that “time-spent is the timeless currency”, an old saying in the world of advertising that means that advertising revenue distribution follows time-spent distributions. If this holds true, then the above graph suggests that ad revenue share for Google will trend downwards while ad revenue share for other apps will rise.

This is a bold suggestion and I would certainly like to see some more information related to this. The current study is still too crude to draw a definitive conclusion, some of the reasons for which I will outline below;

  1. Of the 18% Time Spent on Google, 14% is actually the total of time spent inside a browser. Flurry is attributing 100% of the time spent in a browser (including Safari on iOS) to time spent on Google. This is an understandable approximation because a large proportion of ads on the web are from Google, but still quite extreme.
  2. Of the 65% Time Spent on Other Apps, a large proportion will be showing apps from AdMob, Google’s mobile advertising network which they acquired in 2009. Hence much of this 65% could be attributed to Time Spent on Google.
  3. Although Google has very strong market positions in both search ads (AdWords) and display ads (AdSense), search ad revenue is growing much faster. It is also the majority of their revenue. It is very possible that search-related ads can earn a disproportionate amount of money that cannot be inferred from time-spent alone.

On the desktop, Google worked hard to provide free solutions for the activities that people did on their PCs. They created Gmail and Google Docs because email and productivity applications were the activities that people did. They acquired Blogger because a lot of the content that people were reading on the web were blogs. They created their own RSS reader because that was how many people read content.

Google was not content with simply supplying ads to websites and webapps; they created or helped create the websites/webapps themselves, or they tried to provide tools to read the content more effectively. They tried to be much more involved than a simple advertising agency.

On mobile, it does seem that Google is struggling to do this. They have failed to create an engaging social network, and they aren’t involved in the creation of games. YouTube, although significant at 4% of total time, is still quite small compared to games and Facebook. They don’t have a direct presence inside the dominant mobile activities.

It will be interesting to see whether Google will push harder to be more than an advertising agency on mobile, or whether it will be content with pushing ads through search and AdMob.

Will High-End Android be a Samsung Exclusive?

I have been very wary of Google’s efforts to build in the tools that Android requires to compete with Apple in the high-end smartphone market. My feeling has been that Google is no longer interested in the high-end smartphone market and is satisfied to let Apple have it. Instead, Google is focusing on the low-end market and on bringing low-cost smartphones to the emerging markets.

Recent announcements reinforce my thinking.

On Nov 13, 2014, Samsung and BlackBerry announced a partnership to build and market a tightly integrated, end-to-end secure solution aimed for enterprises. It is well known that Apple dominates corporate market share in smartphones and tablets, and security is one of the reasons why Android is struggling. The interesting thing is that Google seems very uninterested in developing a solution of their own for this quite lucrative market.

At the keynote of Samsung Developer Conference 2014, Samsung introduced Samsung Flow, which is basically their version of Apple’s Continuity features. The important thing to note is that Continuity is a feature that is only valuable for customers who can afford multiple devices. Again, why did Samsung and not Google develop this? Is it because Google is relatively uninterested in the customers who are wealthy enough?

Samsung recently announced its answer to Apple’s iBeacon feature, Proximity. This is a technology to enhance customer experience at stores, mall, stadiums, museums, etc. and also push you special offers and stuff. As stores embrace iBeacons, there was the possibility that some of the offers would end up being exclusive for iPhone users. Now it will be exclusive to iPhone and Samsung users. Why wasn’t Google interested?

If this continues, Samsung might create a rather formidable barrier-to-entry for the high-end Android market, blocking HTC, LG, Lenovo and others from competing.

Although I do suspect that the overall market for high-end Android devices might shrink, I do not doubt that Samsung will continue to dominate for the foreseeable future.

How Apple Has Actually Introduced New Category Products Every Few Years

There are several people who seem to have had difficulty applying Clayton Christensen’s Disruption Theory to the tech world, and have proposed that this theory itself does not apply to either the tech market or consumer goods. However Horace Dediu, who is now working for the Clayton Christensen Institute specifically for the task of applying Disruption Theory-based analysis to the tech market, has argued in a podcast with Ben Bajarin, that there is no reason why it cannot be done. Horace argues that the challenge lies in defining what constitutes the “jobs-to-be-done”, and a failure to do this successfully is why Disruption Theory sometimes seems to fail.

This point is very important and is worth reiterating. The reason why Disruption Theory occasionally fails to explain a certain situation is not because the theory itself is limited in its scope; it is because identifying the jobs-to-be-done is extremely difficult. In fact in the typical example of the jobs-to-be-done, the milkshake example, the jobs-to-be-done is so unintuitive that its unlikely that even an industry expert would have accurately predicted it. It is no wonder that Clayton Christensen himself failed to predict how Disruption Theory would affect Apple.

More often then not, the people who attempt to expand, supplant or even discredit Disruption Theory have simply neglected to carefully analyze the jobs-to-be-done.

To further complicate things, if you look at the brief history of personal computing, which has only been with us for four decades at most, you can also observe that the jobs-to-be-done has shifted extremely rapidly. At most, a certain jobs-to-be-done will be mainstream for only five years.

For example, the PC started out in the Apple I era as a hardware hobbyist’s kit. With the Apple II, the PC was now a platform for a hobbyist software programmer. Then with the advent of packaged software like VisiCalc, the PC became a business tool for performing large numbers of calculations. With the arrival of the Macintosh and the Laser Writer, the PC now became a tool for creative professionals, and then with the Internet, it became a tool for communication and collaboration. After the year 2000, the PC became a tool to manage digital photos, music and video.

With each shift in the jobs-to-be-done, the required hardware specifications to sufficiently perform the task increased. At the same time, the customer base continuously expanded to less tech-savvy users which required the user interface to improve. This meant that the PC rarely reached the good-enough threshold because the bar was constantly being raised.

The exact same thing can be said for smartphones. The original iPhone started out as a phone, an iPod and an Internet communicator. In a short amount of time, it quickly became your main camera, your gaming console, your map, your photo album, your fitness tracker, your newsreader, the pacifier for your kids, your TV and so much more. And now with Touch ID and Apple Pay, Apple is making your iPhone your ID and credit card. From its initial humble jobs-to-be-done, the smartphone is now the center of a huge portion of your life. The jobs-to-be-done of smartphones has exploded.

And as with PCs, each shift in the jobs-to-be-done has required new and better hardware. Being your main camera has demanded better optical and image processing hardware and software. Being your gaming console has demanded better 3D graphics performance which technologies like Metal and better embedded GPUs can provide. Being you fitness tracker has resulted in technologies like the M7 motion co-processor which can constantly track your movements with minimal battery drain. And of course Touch ID and Apple Pay required new biometric hardware.

By now it should be plainly obvious why Apple has avoided being disrupted; Apple has consistently been at the forefront of the shifts in the jobs-to-be-done in personal computing (except for the years when Apple was run by John Sculley and R&D was run by Jean-Louis Gassée). That is why new Apple hardware has constantly been in high demand and can still command a high premium.

On the other hand, the reason why Samsung is being disrupted at the low-end is because Android is not expanding the frontiers of smartphone jobs-to-be-done. Other than UI tweaks that work equally well on less capable devices, Android has recently failed to introduce compelling features that require new or better hardware. In fact, this might have been intentional on Google’s part as an initiative to reduce fragmentation. As a result, Android phones have become as good-enough as they can be, even on the hardware that can be bought for $200-300. The Android OS is holding Samsung back.

This also means that Apple products will be disrupted if they start failing to create new jobs-to-be-done. It also means that the resurgence of the Mac could be attributed to a new jobs-to-be-done that the Mac can uniquely satisfy. Strong integration with iPhones could obviously be one of these jobs-to-be-done which is not available on Windows PCs.

Actually, if you look at computing from a jobs-to-be-done standpoint, the idea of a “new category” device starts to look rather ridiculous. It becomes clear that the emphasis should be on whether or not a new jobs-to-be-done has emerged. Sometimes this might require a new device, but oftentimes, it simply requires new hardware on top of a preexisting device. This in itself is sufficient to transform the previous hardware into a new category device. For example, Touch ID has transformed the iPhone into a digital ID and wallet with unprecedented security and convenience, something that was hitherto impossible with a smartphone. Focussing solely on new category devices is completely missing the point.

In fact you could even argue that every few years, Apple has introduced a new category product in the guise of the iPhone or a new iOS version; a product that enables new jobs-to-be-done to emerge.

Who Is To Blame For Samsung’s Bad Fortune?

As the profits plunged on Samsung’s smartphone business, the web has been awash with reasons.

Ben Bajarin has shown very nicely that the largest problem that Samsung faces is the decline of the high-end business, which is also mentioned by a Samsung executive in the Guardian article.

The high-end of the business has been dominated by Samsung and Apple and still is. This means that there are two possibilities.

  1. Apple took away Samsung’s sales in the high-end. That is to say, users of high-end Android phones (who were mostly using Samsung devices) switched to the iPhone.
  2. The high-end market for Android smartphones saw a sudden shrinking. That is to say, mid-range smartphones were perceived as good enough and hence there was no need for customers to purchase high-end Galaxy devices anymore.

I suspect that both of these happened but I want to analyze them in isolation because it makes the situation easier to understand. Although these two look similar, they are actually very different. The first means that Apple was able to steal market share away from Samsung. The second means that vendors of mid-range smartphones (including Samsung of course) captured the customers who previously bought high-end phones. We will look at each separately.

Apple is stealing away the high-end

This is obviously happening. All reports point to Apple selling huge numbers of iPhones and it has been suggested that a lot of these are switchers who have abandoned Android phones.

The important thing is why. Of course the triggering event is the increased screen size of the iPhone 6. However, what is more important is why couldn’t Samsung match the iPhone 6 before Apple threw down the gauntlet. Why was Samsung left clinging to screen size as the only feature that could keep it competitive in the high-end.

Although design and/or Apple’s brand could well be a factor, it is also as likely that iOS and its app ecosystem could have been perceived to be superior than Android. If this was the case, then the blame would have to be put onto Google. Google failed to create an operating system and ecosystem that was competitive against iOS. The only reason that the high-end Android market existed at all was because Samsung had large screens while Apple did not.

If it was design or branding, then it would be harder to place the blame on either Samsung and Google simply because Apple is so good at these. Either way though, the result is that the high-end Android market cannot exist anymore.

The high-end Android market is shrinking

This is a completely different dynamic. If this were the case, then we should be seeing customers who previously owned the flagship Galaxy devices either downgrade to mid-range Android devices or to extend their replacement cycle. I have not yet seen a statistic that suggests that this is happening, but it is plausible.

This can only happen if Android smartphone hardware is starting to be considered as good enough, even by previous high-end purchasers. This also has to happen while at the same time, on the Apple side of the fence, Apple customers are not considering iOS hardware to be good enough. There must be something very different happening to Android customers and iOS customers.

The good enough of hardware is determined by software. If the software can take advantage of new hardware and create a true benefit for the customer, then old hardware will not be good enough. On the other hand, if the software does not have any compelling features that require new hardware, then old hardware will be good enough. No matter how much the hardware improves, whether customers will demand it depends on software.

In the case of iOS, the OS made full use of the 64-bit hardware to enable much faster processing of photos and movies. The OS made use of the TouchID sensor, which is also now being used by the Apple Pay service. Apple has given each piece of new hardware a significant reason for existing, and that is why customers want new devices.

On the Android side, that has not been the case. Google has not moved quickly to 64-bit, it has not worked hard on corporate level security, and it has not introduced software support for biometric sensor technology. Instead, Google has introduced a lot of software technologies that enable low-powered devices to smoothly run the latest operating system. Instead of adding new features that would take advantage of new high-end hardware, they focused on making sure that the mid-range and low-end hardware would be able to run the latest operating system and to take advantage of all of its features. In summary, Google actively designed their new operating system so that Samsung would have a hard time differentiating itself.

Although I’m not sure whether Google did this intentionally, it has made it very difficult for high-end Android smartphones to compete with mid-range ones. This is not only a challenge for Samsung, but it will also be a challenge for any OEM that plans to move upmarket. It will mean that companies like Huawei, Lenovo and Xiaomi will not be able to move up-market unless they gain significant control of the OS.

So what should we blame?

I think that Google was targeting the low-end from the start, but Andy Rubin was not. I genuinely think that Andy Rubin was much more focused on the high-end and he didn’t seriously consider making Android work better on low-end devices. I think he wanted to make Android as good as or even better than iOS. The fact that his reign coincided with when Samsung was strongest is no coincidence.

When Andy Rubin was removed and Sundar Pichai took over, it became rather clear that instead of fighting with iOS, Android would focus on the low-end. In fact, most products that Google creates (many of which were under the supervision of Sundar) aim at the very low-end where prices are normally zero. Google Docs is a prime example of this, as is Chrome OS. Google’s strategy is to commoditize all markets except for search and advertising, by providing a good enough product for free.

Samsung could have tried harder to take control of Android so that they could create software that took advantage of high-end software. In fact, they tried. Considering that Samsung was mainly a hardware company, I don’t think that they ever misunderstood that they needed good software; it was just that they didn’t have the resources or the culture to create great software. It’s hard to blame their strategic thinking for this.

Google could have tried harder to preserve the high-end. However, it’s priorities were clearly in the low-end. It’s hard to focus on both.

I would say that the only strategy that we could actually blame was Samsung’s decision to team up with Android. Samsung should have seen that Google would ultimately aim to commoditize their own OS and all hardware vendors using their platform. Samsung should not have helped Android to gain market share, and instead waited for a contender whose priorities aligned better with Samsung’s goals. Of course, that is what Nokia did.

iPad Installed Base

A lot of pundits are bemoaning over the flat sales of the iPad, declining sales of tablets overall, and the absence of radical improvements to the iPad lineup that would spark replacement sales.

Flat sales would be a large issue if it was caused by less users using and enjoying the product. It would however not be an issue if people were holding on to their old product longer, but still actively using and enjoying it. That is to say, what matters is not the number of units sold per quarter. What matters is how many people are using it.

One way to understand how many people are using iPads is to look at web usage data. Although web usage data is only a crude proxy, I believe that it can give us a good idea of the trends if we are take sufficient care. Here, I have taken web usage data from StatCounter.

USA
StatCounter os US monthly 201102 201409

United Kingdom
StatCounter os GB monthly 201102 201409

Europe
StatCounter os eu monthly 201102 201409

What we see is that iPad web usage (we are only looking at Desktop & Tablet devices so iOS is equal to iPad) has steadily risen since its introduction in 2011. In the USA data, we see usage stagnating in 2013. However this has recovered in 2014. We also do not see a similar slowdown in neither UK nor Europe data, so we don’t think that the USA stagnation is something to seriously contemplate.

This strongly suggests that despite a slowdown in sales since 2013, iPad usage has risen steadily since introduction. The longevity of old devices like the iPad2 has definitely contributed to this. In fact, the iPad 2 is still the most used iPad model.

Of course, this is not without its problems. If old devices stick around too long, it can slow down progress in the ecosystem since developers will have to support older and less capable devices. Although this is an issue that Apple definitely has to address, this is a totally different issue from the tablet category losing appeal.

So to summarize, despite a slowdown in sales, more and more people are actively using iPads and enjoying them. Although it would be ideal if sales were rising, it is more important for Apple to ensure that more and more people start using iPads and continue to love them. Apple seems to be doing this quite successfully. However, Apple will have an issue in the future with fragmentation because old devices will stay in use for longer.

Misguided Expectations for Replacements Cycles

Many people have blamed the slowdown of iPad sales on the fact that the replacement cycle of iPads is quite slow. In fact, we don’t really know what the replacement cycle is yet because the device is still very new (even the first replacement cycle hasn’t yet kicked in) and the second generation device, the iPad 2 (introduced March, 2011) is still used quite a lot.

NewImage

My question is, is the replacement cycle too long and should we be blaming Apple (as quite a few analysts are) for the lack of reasons to upgrade? Should we blame Apple for not introducing compelling improvements to the iPad that would drive users to buy new devices? Should we blame Apple for mismanaging the App Store to the effect that not enough exciting titles are being released for iPad?

This hinges on what the natural replacement cycle for a tablet device should look like. If the natural cycle should be something like two years, then yes we can blame Apple. If it is however something like 4 years, then we cannot conclude that Apple is doing a bad job.

Therefore, I think we should give some thought on to what the natural replacement cycle for a tablet device should actually look like.

Smartphones

The replacement cycle for the phone market varies from less than 2 years to over 10 years (interestingly, Android phones seem to have a much faster cycle).

Screen Shot 2014 10 18 23 08 44

Recon Analytics sums up the reason for differences in replacement cycles as follows;

Based on the data and analysis outlined in the report, it is conclusive that over the last four years, handset subsidization is the dominant factor influencing the handset replacement cycle. The percentage of subscribers on postpaid and prepaid plans, as well as the relative income level in the countries, had a negligible impact on the handset replacement cycle.

Considering that the majority of iPads are WiFi-only and that these are not subsidized, we can expect iPad replacement cycles to be significantly longer that phones. There is very little reason to expect iPad replacements every two years.

PCs

The replacement cycle for business PCs in the US was a bit longer than 3 years. Why do they replace them so often?

  1. Increased productivity: If the old PC is much slower than the most recent models, then a new one would increase productivity.
  2. Escalating support costs: If the old PC tends to break down a lot, then buying a new computer may become cheaper than the maintenance costs.
  3. Software requirements: If the old PC cannot run new software, then it’s time to upgrade to a new PC.

Now how much of this would apply to tablets?

The amazing thing about the iPad, even the original model, is how fast it was on the limited hardware. Apple went to great lengths to achieve this, even sacrificing features that have been found on PCs since 2000 like multitasking in the background. Apple has kept third party software under strict restrictions, and this has helped keep software from bogging down the system. Apple itself has worked hard not to make iOS bloated.

As a result, the iPad 2 from 2011 still has enough performance to run the most recent iOS (iOS 8) with OK speed. Hence “increased productivity” does not apply very much to iPads and neither do “software requirements”. We also have to understand that iPads are mostly used by consumers, and so less emphasis is places on “increased productivity”.

Another amazing thing about the iPad is how durable it is. Without almost any moving parts, not even a keyboard, there is very little that can break. The build quality of the device was also superb from the onset. Also, unlike phones which you carry about you all day, you are much less likely to drop and shatter an iPad on concrete. Simply put, the cost of maintenance for an iPad is remarkable low.

Since none of the reasons for a 3 year PC replacement cycle apply to iPads, there is no justification for expecting similar cycles for iPads. It is very possible that the replacement cycle for an iPad is much longer than 3 years.

The one thing to note is that the iOS 8 is bearable on iPad 2, but stutters quite a bit. This is probably due to the fact that it only has 512 Mbytes of RAM and I think that it is unlikely that the next iOS version will support it. If so, then “software requirements” will demand a replacement next year.

Other consumer electronics devices

For most consumer electronics devices, we generally only replace them if they break down or our family gets larger (and we need a larger refrigerator or washing machine). Unless you buy them from a manufacturer that is seriously skimping on important components, they should last at least 5 years.

Summary

As we can see, the 2 year replacement cycle that many analysts were initially expecting for tablets was completely misguided, and hence we cannot blame Apple for a cycle that may be 3 years or longer.

We could even argue that having compelling new features is only rarely a reason why people ever upgrade their devices. This is for the most part irrelevant to the upgrade cycle. In fact, the main pain points cited for upgrading PCs are mitigated by stricter control of third-partly applications, better hardware build quality and simpler hardware design on iPads.