Home vs. Echo – Battle of the Home pt. III.

Reply to this post






Google has everything to do.

  • Google is great at getting third parties to build hardware that uses its software, but needs to work on developers of smart home devices if it wants to trounce Amazon.
  • Following the general availability of the Google Assistant SDK that allows anyone to embed Google Assistant into almost anything, Google has also announced a series of third party devices which will be launched at IFA next month.
  • These include the Anker Zolo Mojo, the Panasonic GA10 and the TicHome Mini all of which will go on sale during Q4 17.
  • Amazon has followed suit but as of yet, there appears to be less traction with hardware makers.
  • Alexa is likely to power the next generation of Sonos speakers and may make an appearance in some VW cars but it looks like Google has more momentum when it comes to hardware.
  • This is likely to ensure a race to the bottom in terms of voice enabled smart speakers from which I think Google will be the only likely winner (just like Android).
  • It badly needs to close the gap on Amazon which has around 70% of the home speaker market and having a much wider selection of attractively priced products will be of great help.
  • What will further help Google is the fact that Google Assistant is a vastly superior product compared to Amazon Alexa.
  • This is because the AI that sits behind Google Assistant is the best available, meaning that Alexa answers fewer questions correctly and gets stuck much more often.
  • However, where Google comes completely unstuck is in the smart home.
  • Amazon has aggressively pursued developers and showered them with love and support meaning that almost every developer of anything that has a Bluetooth or WiFi radio can be controlled with Alexa.
  • The same cannot be said for Google Assistant which I think has been caused by Google’s surprising lack of support for developers of this type (see here).
  • I think that part of the reason for this is that Google Assistant has been brought to life by one part of Google (hardware) but was created and managed by another.
  • Google is addressing this by encouraging developers to write directly to the assistant meaning that any device be it a smartphone, speaker or thermostat can run the smart home but progress to date has been slow.
  • Amazon Alexa has over 15,000 skills which don’t work very well but importantly, there are there and do work with a little effort.
  • Google Assistant is hopeless by comparison and it is here that it is at real risk of suffering a Betamax-like defeat.
  • I think that Google needs to bring all of these devices together such that “OK Google, I am going to bed” results in the whole house shutting down rather than a long series of carefully constructed instructions to each device individually to go into night mode.
  • For many of Alexa’s skills, it is simply easier and quicker to perform the operation manually than to ask Alexa to do it.
  • Unfortunately, so far there is no sign of smart integration from Google meaning that the advantage remains with Amazon.
  • The market remains very lowly penetrated meaning that everything is still to play for but this won’t last forever.
  • Valuation keeps me from liking Alphabet and Amazon leaving Microsoft, Tencent and Baidu as my top choices.


Google vs. HERE – Perfect parking

Reply to this post






HERE has more potential but needs to execute.

  • Google is upping the ante in the race to provide value-added services on top of digital maps, but with the right execution, HERE should be able to provide a much better service.
  • In January Google launched its parking difficulty icon on Android Maps which give the user an idea of how difficult it will be to park at the user’s destination.
  • This was initially launched in 25 US cities but this has been: 1) extended to another 25 locations (Canada, Europe and Brazil), and 2) expanded to offer parking suggestions in the area.
  • This service is based on historical parking data as well as data gathered from smartphones using Google services within a certain location to ascertain how busy that location is.
  • This is similar to the popular times and visit duration data that Google provides for businesses and will give the user an idea of how long he should expect to spend looking for a place to park.
  • While this will be a useful addition to Google Maps, I think that HERE should be able to offer a service that is vastly superior.
  • This is for two reasons:
    • First, data quality: While Google’s service is based on estimates and AI, HERE’s service should be based on much more specific data.
    • This is because HERE has access to automotive sensor data while Google does not.
    • For example, when HERE’s location platform detects an ignition start, it can be almost certain that the space occupied by that vehicle is about to be vacated.
    • It will also know from ignition switch-off which spaces are occupied and which are not.
    • This gives it a highly accurate, real time picture of the parking environment meaning that it’s HERE ON-Street Parking service should be much more accurate than Google.
    • Second, positioning. Vehicle positioning is often much more accurate than that offered by mobile phones as the antennas are larger and are almost always open to the sky.
    • This means that HERE should have a more accurate real-time picture of exactly where the devices connected to its platform are.
    • Combining this with the highly granular data it gets directly from the vehicle, should allow HERE to provide its users with a more accurate and relevant parking service than Google Maps.
  • This is exactly the kind of differentiation that HERE needs to win the attention of users but there are caveats.
  • Google is present on almost every smartphone in the market (except China) meaning that although its data set is much less accurate, it has a much fuller picture of the environment.
  • HERE by comparison is at a very early stage in getting devices connected to its location platform meaning that its lacks the visibility of the environment to make its service work really well.
  • This allows Google to offer a workable service today, while HERE is still at the stage of building out its network of data collecting devices.
  • Furthermore, should Google manage to get access to the sensor data generated by vehicles (Android Auto offers no access), then HERE’s key advantage will be lost.
  • However, most automakers have recognised that Google represents a meaningful long-term threat and are keen to keep their sensor data to themselves.
  • Google has done a deal with Volvo and Audi but whether it has managed to gain access to sensor data is still unclear.
  • The net result is that HERE has an opportunity to roll-out a much better service and win over users, but it needs to quickly achieve scale or risk being swamped by Google should it gain access to sensor data.

Research Publication – Automotive ecosystem – Sitting ducks

Reply to this post

August 9th 2017:  Radio Free Mobile deepens its coverage of digital ecosystems with the publication of: Automotive Ecosystem – Sitting ducks.

RFM research subscribers will receive their copy directly by email. 

Click here for a sample and here for purchase options









Transport is the next industry to be digitised and almost all vehicle makers are unprepared. Electric Vehicles (EVs) and autonomy could reduce overall USA transport spending by 65% which needs to be supplemented with digital services by players wishing to survive. RFM thinks that sensor data is the one area where vehicle makers have an edge and hence, is critical to their future. They must control this asset or face becoming sitting ducks for those that would reduce them to handsets on wheels.

  • Digital differentiation. Transport is ripe for disruption. Furthermore, there is a real possibility that demand for vehicle shipments falls substantially over the next 10-15 years. RFM thinks that embracing digital, controlling sensor data combined with a completely new way of thinking is required by vehicle makers wishing to survive for the long-term.
  • Sensor data will be the new vehicular currency. RFM thinks that Digital Life services from smartphones will become ubiquitous and unlikely to offer value for vehicle makers. However, sensor data is unique, required for autonomy and critically, they still have a lock on access to it. RFM sees sensor data as the opportunity for vehicle makers to avoid severe disruption.
  • The infotainment unit could become the most important part of the vehicle as it is where all the sensor data can be accessed in one place. Furthermore, it is the main digital interface with the user meaning that the digital user experience will be defined here.
  • The gatekeepers. Despite the threat, RFM believes that the fact that OEMs are the gatekeepers to sensor data will give them a seat at the table as well as the opportunity to differentiate. How they execute on this is likely to define who survives and who does not.
  • Monetisation. RFM calculates that Digital Life (smartphone only) in the vehicle could be worth $112 per user per year in USA or $32.1bn in revenues. The use of sensor data could drive that figure higher. Potential substantial falls in both the radio advertising ($17.7bn) market and transportation ($2.6tn) market provide a plentiful source for spending on new digital services.
  • EVs and autonomy have the capacity to cause substantial declines in both vehicle shipments and transport spending as a whole. RFM calculates that manual EVs could reduce cost per mile to $0.40 per mile from $0.88 where it is today. Autonomy promises to reduce this still further to $0.29 per mile. There is huge economic incentive for consumers to switch to EVs which together with autonomy, could cause a 44% reduction in USA vehicle shipments.
  • Sitting ducks. While vehicle makers are aware of the threat, many are in denial and few have any real idea how to address it. Most are easy targets for those that would reduce them to handsets on wheels.

Wearables – No show

Reply to this post






Adding LTE to Apple Watch is pointless.

  • The problem with wearables that I described nearly 4 years ago (see here) of a solution looking for a problem continues to plague wearables with Fitbit struggling along, sector consolidation and even Apple seems out of ideas.
  • If the rumour machine citing endless anonymous sources is to believed Apple’s next version of the Apple Watch will feature a LTE modem.
  • This will give the device independence from the iPhone, meaning that the user won’t have to have the iPhone in immediate proximity for the device to work.
  • I think that putting a LTE modem is pointless and could even harm what little appeal the product has.
  • This is for two reasons:
    • First: In the US (I suspect Europe is similar) users are now glued to their smartphones for an average of 300 minutes per day (Flurry).
    • This essentially means that users keep their smartphones in their immediate vicinity at all times and will go to great inconvenience to ensure that that remains the case.
    • As a result, there is only a tiny period (if any) of time when the smartphone is out of Bluetooth range of the user and hence any wearable that he has on him.
    • Therefore, the inclusion of a cellular modem will be able to improve the functionality of the Apple Watch for only a tiny percentage of the user’s day.
    • This renders it effectively useless in my opinion.
    • Second: A LTE modem (even with a soft SIM) is going to cost money, take up space that could be something used for something else and will be a net drain on the battery.
    • Battery life is a major issue for all wearables (including the Apple Watch) and the addition of a modem will place a further drain on already very limited resources.
    • Hence, I think that a modem will cause deterioration of the user experience for no perceptible improvement.
  • What Apple should really be working on is a use case or function for this product that makes it a must have causing everyone to rush out and buy it.
  • This is the genius for which Apple has been known in the past but of which there has not been much sign over the last few years and certainly not with this product.
  • Consequently, I hope that like the large screen TV and the vehicle, this product never sees the light of day.
  • Instead, I am looking for a use case that can really kick start the wearable market as without this spark of genius, it is likely to continue bumbling along with little real interest or volume.

Samsung Bixby– Lightweight

Reply to this post






Bixby is hopelessly outclassed.

  • Bixby voice only shines in the areas where Samsung has given it special access to hardware that competitors do not have.
  • Outside of this area, Bixby is a third-rate experience that is unlikely to generate much traction especially as the vastly superior Google Assistant is just a button press away.
  • The voice piece of the Bixby digital assistant has finally launched but despite months of feverish activity in trying to teach Bixby to speak English, it is still not very good at it.
  • Bixby has been granted exclusive hardware access such that it can work when the screen is off or the device is locked.
  • This is something that Google Assistant cannot do but it also comes with the reality that Bixby is always listening.
  • I think that this will make some users very uncomfortable as a microphone in one’s pocket is far more intrusive than a microphone listening in the kitchen.
  • Voice recognition works best when there is an element of training involved as users often say things in very different ways.
  • Unfortunately, it appears that for some users, Bixby is unable to recognise the training sentences implying that this part of the system still needs work.
  • In effect Samsung has programmed Bixby with a series of standard functions that can be used to operate the smartphone as well as basic functions in the apps.
  • Outside of that area, the user is pretty much out of luck.
  • Unfortunately, these only really work for Samsung apps which outside of the messaging app for SMS, I think no one really uses.
  • Furthermore, for navigation and search, Bixby uses Google but without some of the bells and whistles that make Google so good.
  • For these functions, it makes no sense to use Bixby when one can go straight to Google.
  • Bixby does support third party apps through the “Bixby Labs” program but unfortunately it doesn’t seem to work properly.
  • Bixby can open things like Google Maps, YouTube and so on but does not seem to be able to get past the main screen of those apps.
  • The problem with Bixby is simply that its creator, Samsung, has no artificial intelligence expertise to speak of and digital assistants are only about AI.
  • Google Assistant is the best not because Google knows how to make an assistant but because the AI that runs it is the best in the world.
  • This contrast is so stark, that Samsung has had to resort to hobbling Google Assistant in certain areas (hot word) just to give Bixby a chance.
  • I think that this will encourage users to try Bixby once or twice but when they realise how poor it is, they will go back to Google Assistant.
  • Google will not be losing any sleep over Bixby even though it could end up on a very high percentage of Google ecosystem devices.
  • Samsung is now the No. 1 semiconductor manufacturer in the world, but I still rank it almost dead last when it comes to AI.
  • I think its investments in this space would be better accruing to shareholders in the form of higher profits rather than being invested in functions that are likely to damage Samsung’s reputation rather than improve it.
  • Samsung’s recent rally has removed the valuation argument for Samsung which leaves me preferring Tencent, Baidu and Microsoft.

GrubHub – Big appetite

Reply to this post






Eats24 is an essential acquisition.

  • GrubHub is making exactly the right moves to give it the best chance to beat its larger and much better financed rivals in the brutal food delivery business.
  • Alongside results that broadly met expectations, GrubHub announced the acquisition of Eat24 for $288m well over double what Yelp paid for it in February 2015.
  • This may seem to a huge premium but when one looks at what Eat24 can bring to GrubHub, it is not difficult to make the case that this could be the most important move GrubHub has made in its history.
  • GrubHub is an online marketplace where diners come to order amd have delivered food from participating restaurants.
  • As an online marketplace (network business) it is subject to exactly the same dynamics as ride hailing, classifieds and so on.
  • 20 months ago I proposed a rule of thumb that states: A company that relies on the network must have at least 60% market share or be at least double the size of its nearest rivals to begin really making profit (see here).
  • In effect, by hitting one of these two criteria the marketplace becomes to the go-to place to transact meaning that buyers become somewhat less price sensitive and sellers will pay more to sell their goods there.
  • It is this that allows the marketplace to make proper money but before this level is reached all players will almost certainly be under excruciating pressure.
  • GrubHub is no different in that since the advent of UberEats and Amazon’s entrance into this space, there has been relentless pressure on margins.
  • In the last 12 months EBIT margins have fallen to 13.9% in Q2 17A from 18.7% in Q2 2016 despite a 32% increase in revenues.
  • GrubHub is the market leader with 34% but Uber is not that far behind with 20%, Eats24 with 16% and Amazon on 11% (Cowen &Co).
  • Regulatory scrutiny has been high on GrubHub’s previous acquisitions but the fact that this deal is likely to be passed with barely a ripple is an indication of how much more competitive the market is now considered to be.
  • There is some overlap between GrubHub and Eats 24 but importantly once combined the platform will have 75,000 unique restaurants on its books and 48% market share of transactions.
  • Assuming that the acquisition and integration proceeds flawlessly, then GrubHub will be more than double the size of its nearest rival (Uber) and able to at least stabilise its margins.
  • Furthermore, as long as it can hold onto this advantage, it should be able to withstand the pressure from its rivals despite the fact that they have very big brothers backing them up.
  • Consequently, I think that GrubHub had to make this acquisition otherwise it faced being ground down by its better financed rivals until it was forced to sell itself to one of them.
  • GrubHub has made the right strategic move to ensure its longevity but now it comes down to execution to determine its future.

Spotify – Free foundation

Reply to this post






Apple Music still making no dent.

  • Spotify has crossed 60m paid users and while its absolute level of growth is slowing due to the law of large numbers, it is still adding paying subscribers at the very healthy rate of 2m per month.
  • In September 2016 Spotify hit 40m, passed 50m in March 2017 and hit 60m at the end of July 2017.
  • For the last 18 months, Spotify has been steadily adding subscribers at around 2m per month which is showing no signs of slowing down.
  • This has held steady for the last 5 months indicating that Apple Music is having very little impact on Spotify despite the substantial advantage Apple has in owning the App Store and having complete control over the iPhone.
  • I continue to believe that this is for two reasons
    • First:  Spotify remains fundamentally a better service.
    • This is driven by the fact that the music is now incidental in that anyone can create a service with 40m tracks and a search box.
    • Where Spotify is different is that it uses the data that it collects from all of its users in order to make its service better.
    • Apple also does this but Spotify’s AI in music continues to meaningfully outperform Apple’s.
    • By understanding the characteristics of the music offered by its service and the preferences of its listeners, it can accurately match the two together.
    • This also allows it to come up with innovative services that keeps its service fresh and one step ahead of the competition.
    • Second: Spotify has a large and engaged free tier of users that serve as the funnel for conversion into paid users.
    • Free users get to spend time with the service without paying for it, making it much easier to make these users understand why the service is better than anything else available.
    • This meaningfully offsets the disadvantage that Spotify has compared to Apple when it comes to marketing.
    • These free users generate data which Spotify can use to train its algorithms which can in turn be used to make the service better.
    • Apple also has a lot of data but has not been nearly as good at turning raw data into actionable intelligence with which it can improve its service.
  • The net result is that Spotify’s position is strengthening with every new user that it adds and between them, Apple and Spotify account for almost all of the growth in the recorded music industry.
  • Consequently, I remain unconcerned that Apple will be able to put real pressure on Spotify and think that its path to better profitability remains clear as the labels increasingly need Spotify more than Spotify needs the labels.
  • This position is becoming clearer as Spotify was able to strike a better deal with Universal (see here) and the outlook is that the rest of the industry will be signed on similar terms.
  • I still think that the key issue for Spotify going forward is to maintain momentum of growth of its free users.
  • It is the free user pool that has been the foundation of its outperformance of Apple, meaning that the free tier will be critical to keep the paid tier (where the real money will be made) increasing at this very healthy rate.
  • I continue to think that there is enough space in this market for 2 big players and with those spots filled, it is the fortunes of Pandora, Tidal, Deezer and so on that trouble me now.

Apple FQ3 17A – No pause here.

Reply to this post






Solid base for new product launches.

  • Apple reported good results and guided strongly for the coming quarter stoking speculation with regards to the possible strength of the coming upgrade cycle with the new iPhones expected to be launched next month.
  • FQ3 17A revenues / adj-EPS were $45.4bn / $1.67 slightly ahead of consensus of $45.0bn / $1.57.
  • While iPhone held steady it was Services that really underpinned the results with YoY growth of 22% to $7.3bn.
    • iPhone shipments were 41m which included a 3.3m inventory reduction ahead of the new launches.
    • iPad shipments were 11.4m up 15% YoY driven mostly by the product refresh that saw a new iPad and the smaller version of the iPad Pro launch in March 2017.
    • Mac shipments were 4.3m units driven mostly by the new MacBook Pro.
  • Services was the star of the show where the Apple App Store is the main driver generating almost double the revenue of its nearest rival Google Play.
  • This was despite the fact that Android devices appear to have closed some of the monetisation gap on iPhone especially at the high end (see here).
  • Guidance was surprisingly strong with revenues / EBIT expected at $49bn – $52bn ($50.5bn) / $11.7bn – $13.0bn ($12.4bn) broadly in line with consensus at $50.4bn / $12.4bn.
  • The result was a relief rally as fears of a pause in performance ahead of the new product launches now looks unlikely to occur.
  • Consequently, all eyes are now on product launches where a major product refresh is hoped to trigger another replacement cycle.
  • Bezel-less devices are now all the rage and if Apple can replace fingerprint ID with an excellent facial recognition system, the iPhone 8 could end up triggering a good cycle of upgrades.
  • Samsung has a wealth of biometric ID systems but none of them work particularly well as the fingerprint sensor is on the back of the device and the facial recognition is not nearly as reliable as it should be.
  • I don’t think that Apple will see a cycle as strong as the iPhone 6 but there is potential for the iPhone 8 to meaningfully outperform the 6s and the 7.
  • That being said, I continue to think that much of this good news is already in the stock and the valuation argument for long term investors has long since evaporated.
  • I remain pretty indifferent to the shares.

Indian e-commerce – Road to ruin.

Reply to this post






The only likely winner in India is now Amazon.

  • Snapdeal has ended merger discussions with Flipkart in a move that, I think, snuffs out the one chance the local players had to keep Amazon at bay.
  • At the same time Softbank is now looking at committing $1.5bn – $2bn into Flipkart in a move that I think will solve nothing because in a network economy, two halves do not make a whole.
  • I think Softbank should not put any more money into Indian e-commerce as the most likely winner in this market is now Amazon in which Softbank has no stake.
  • Snapdeal’s strategy is now to become a niche player and is cutting costs and selling assets in order to raise the capital required to reach profitability in its niche.
  • In my opinion, this strategy demonstrates a fundamental misunderstanding of how Amazon works and what it is likely to do to win the Indian market.
  • Most companies have a strategy that involves trade-offs such as offering high quality or low prices.
  • This is the route that Snapdeal is taking by deciding to streamline and focus on by giving sellers the best experience in India.
  • This is not how Amazon functions as there is no either / or in its vocabulary.
  • Instead Amazon goes for dominance and offers high quality and low prices or in this case the best experience for both sellers and buyers
  • How Flipkart will alter its strategy following the failure of the merger remains to be seen, but without the scale that Snapdeal would have given it, its chances of seeing off Amazon are greatly reduced.
  • Flipkart, Snapdeal and Amazon are network businesses just like Uber, Alibaba, AirBnB, Craigslist and so on and consequently, they are bound by the same rules.
  • 20 months ago I proposed a rule of thumb that states: A company that relies on the network must have at least 60% market share or be at least double the size of its nearest rivals to begin really making profit (see here).
  • This, in a nutshell, is the problem faced by both Flipkart and Snapdeal in India.
  • Flipkart is bigger than Snapdeal and so it is in a slightly better position, but it is not double the size of its nearest rival.
  • Furthermore, both will now have to contend with Amazon which is absolutely determined not to make the same mess of India that it made in China when it went up against Alibaba and lost.
  • I estimate that Amazon pumped $400m of losses into the Indian market in Q1 17A and roughly the same amount again in Q2 17A and I don’t think it will be afraid to up the ante from here if needed.
  • Amazon is not the largest in India but it can lose far more money for far longer than either of the other two.
  • Flipkart has around 35% of monthly active users but it will need to reach at least 50% before it is double the size of Amazon (7Park Data).
  • This is why a Snapdeal merger made sense, because adding Snapdeal’s users to its own would have got it pretty close to achieving that milestone.
  • Consequently, Amazon will now be able to grind its two main rivals down to the point at which they either exit the market or agree to be acquired.
  • Both of these scenarios are likely to result in much lower valuations than were being discussed as part of the deal.
  • Hence, I think that Amazon is the only real winner from the failure of this merger which I think raises existential questions for local providers of e-commerce marketplaces in India.
  • In the short-term this is unlikely to help Amazon’s fundamentals much and so I remain unenthused with an investment in its shares.
  • I continue to prefer Tencent, Baidu and Microsoft.