Samsung today announced its preliminary results for Q3, but also announced that its CEO, Oh-Hyun Kwon, will step down in March 2018. The preliminary results are very much in keeping with results for the last several quarters, with record revenues and profits for any quarter, likely again driven by very strong memory sales and higher memory prices, though the smartphone business also likely did well again. Samsung’s management structure is a little different from most western tech companies: it technically has three CEOs, who each run a big chunk of the business, with Kwon overseeing the Device Solutions segment while also being chairman of the board of directors and therefore the closest thing to an overall CEO the company has. That Device Solutions business, interestingly, is the home of memory and other component products at Samsung, which have been the source of its great financial strength, so Kwon’s resignation certainly isn’t a response to poor performance. Rather, it seems likely to be a response to the broader scandals that have engulfed the Samsung group of companies, including Jae-Yong Lee, grandson of the company’s founder, who is currently in jail (and was also technically a vice chairman at Samsung Electronics). His resignation later certainly makes it sound like he is resigning as a matter of honor and to allow the company to move on under fresh leadership, rather than an indication that he’s tainted at all by association with the scandals.
AT&T has pre-announced some figures for its third quarter results in an SEC filing, including nearly 300,000 DirecTV Now streaming subscriber additions in the quarter, but around 90k traditional pay TV losses. Assuming that latter number is reported on the same basis as in the past and therefore excludes the DirecTV Now customers, it would represent a significant improvement, as the company lost over 300k subs in Q3 2016, and over 200k subs in Q2 this year. But losses are losses, and although the company through hurricanes into the mix as a driver, it’s clear that the underlying drivers that caused previous declines are still big factors too, and those include competition (and have in the past included the challenge AT&T faces of not being able to provide broadband-TV bundles in big chunks of the US).
Two wireless items in the filing are also worth noting. Firstly, the company said it saw 900k fewer postpaid phone upgrades in the quarter, a continuation of a long-standing trend at AT&T of lower upgrades over time, which has seen it fall to by far the lowest upgrade rates among the big four US carriers. Secondly, it’s breaking out certain prepaid IoT connections – notably those associated with connected cars – in its reporting for the first time, and it sounds like it has just over half a million of those as of the end of the quarter. That’s a tiny fraction of its overall connected car connections, which stood at a cumulative 13 million connections as of the end of Q2, the vast majority of which are low-revenue telematics connections sold to car manufacturers rather than directly to end users.
Tesla today released its customary quarterly update on car production and deliveries, for Q3 2017. The overall number of cars produced was 26 thousand, just 5% up on last year’s total for the quarter, with just 220 Model 3 cars produced relative to the 1500 the company had projected. The company delivered to customers slightly fewer cars, including 260 Model 3s, indicating that it’s still a very long way from the mass production of these cheaper cars which it’s been forecasting. Tesla’s statement on the lower than expected Model 3 production is worth quoting in full: “Model 3 production was less than anticipated due to production bottlenecks. Although the vast majority of manufacturing subsystems at both our California car plant and our Nevada Gigafactory are able to operate at high rate, a handful have taken longer to activate than expected. It is important to emphasize that there are no fundamental issues with the Model 3 production or supply chain. We understand what needs to be fixed and we are confident of addressing the manufacturing bottleneck issues in the near-term.” All of this is classic Tesla – in effect: we fell short of our targets, unexpectedly, but we’ll still be able to meet our previously stated targets in the end. Past experience shows Tesla does generally recover from such setbacks, but not usually enough to deliver on original goals – something that’s been pointed out repeatedly by me and by others, but which still seems to engender remarkably little skepticism about its public pronouncements by many investors.
Roku today made public its S-1 filing with the SEC as the first step towards a long-awaited IPO. I’ve been tweeting charts and nuggets from the filing for the last couple of hours in this thread, but I’ll provide a brief summary here. The long and the short of it is that Roku is growing at a decent clip, is currently unprofitable with little sign of that changing, and is in the midst of a big shift in its business model. Whereas for most of its history selling its streaming boxes has been its core revenue stream, it’s recently added a platform licensing business, but that’s not actually where its new revenue streams are coming from. Rather, it licenses its platform very cheaply and monetizes usage by taking a cut of certain subscriptions sold through its platform and serving up ads. It’s the latter which is a surprisingly important part of its business model (though there have been signs of this shift) and which is a major focus of much of the text in the S-1 filing. Last year, this advertising and subscription revenue share was nearly $50 million out of its $400 million in total revenue, and half of its platform revenue, and that accounted for essentially all of its growth in 2016. In that sense, though Roku on paper looks like principally a hardware company, it’s in some ways more like a Facebook or a Google – a company that collects millions of data points on its customers (18TB of uncompressed data per day) and will use that to target advertising. In that sense, Roku is an unusual player in the streaming space, given how many modern streaming services eschew advertising, but sees itself as a key beneficiary of the move of TV advertising dollars from traditional channels to streaming. This is going to be a fascinating IPO to watch and I’ll have plenty more analysis on Roku in the next few days.
★ Lenovo Reports Q2 Loss, Flat Smartphone Sales (Aug 18, 2017)
Snap Inc reported earnings for Q2 2017 this afternoon, and it missed analyst consensus estimates pretty much across the board, with lower user growth, ARPU, and therefore revenue, as well as EPS, than expected. Snap’s user growth sequentially was 7 million, a little above its Q4 2016 number but below every other quarter’s growth for the last two years. Interestingly, in a reversal of the recent trend from Facebook, it saw better growth in North America, where its ARPU potential is much higher (currently 5x Europe and 7x Rest of World) but costs to serve are more or less the same than in its two other regions. But if user growth is going to remain slow, ARPU really has to grow rapidly, and it’s not yet seeing the kind of ramp it needs in that trajectory to justify rising expectations of its future performance. Though it saw 60% of its recent ad impressions generated through its self-serve and automated (API) platforms, the increased inventory made available through those platforms has generated lower prices per ad, so even though impressions went up quite a bit thanks to increased time spent and modest user growth, that offset it somewhat.
Management commentary on the call was mostly focused on the new creative tools being made available to users, which have historically driven increased engagement, as well as the evolving ad platform, which is the other major component to driving ARPU up. But the evidence from the Q2 reported numbers suggests neither of these is dramatically changing the trajectory from prior quarters. There is still tons of headroom in ARPU – Facebook’s global ad ARPU is over four times Snap’s, while its North American ARPU is roughly ten times as high – but Snap is a long way from reaching that level yet, with relatively modest increases in North America in the quarter and better growth overseas. My basic thesis on Snap remains the same: it’s a platform with a slow-growing and smallish (relative to the ad giants) base of highly-engaged users heavily skewed toward particular demographics. That will continue to be attractive to those looking to reach those demographics, but will continue to fall short of the appeal of the much bigger audiences and more sophisticated ad tools available at Google and Facebook. I don’t see any of that changing soon, which means Snap is best seen as its own animal, at similar scale to Twitter (though that scale is measured differently) and with some of the same problems. In other Snap news today, Mashable’s Kerry Flynn reports that it has apparently acquired selfie drone company Zero Zero Robotics. There was no mention of this on the call, but if I heard correctly management did mention roughly $200 million in acquisition costs in the quarter, which would gel with this reporting.
Square Reports Strong Growth and Shrinking Losses in Q2 (Aug 2, 2017)
Apple reported its fiscal third quarter / calendar second quarter results today, and they came in at the high end of its guidance and beat analyst estimates. One of the biggest surprises was strong iPad unit growth year on year after four years of declines, and just the second quarter of revenue growth for iPads during that period, thanks largely to sales of the lower-priced $329 iPad introduced earlier this year. But Apple said all its product categories saw year on year revenue and unit growth, with Apple Watch reportedly growing 50% year on year, and Mac and iPhone unit growth up modestly, while the Services business continued on its recent tear, driven largely by the App Store, but also to an extent by Apple Music and iCloud storage plans. iPhone ASPs were up modestly year on year driven by stronger sales of the latest Plus models, and would have been up more if not for the fact that the company sold down its inventory significantly, with almost all the reduction being made up of more expensive phones.
Perhaps more significantly for the longer term outlook, the company provided guidance for the September quarter which essentially guarantees new iPhone hardware in September. I would guess that at the very least Apple will have the successors to the current phones on sale in the usual timeframe and in the usual volumes, while my hunch is that the new higher-end model will also go on sale at the same time but be even more heavily supply-constrained than new iPhones usually are.
Apple continued to talk up performance in mainland China as distinct from the Greater China region it reports, where sales were down 10% year on year, the best result in nearly two years, but still a drag on overall results with other regions all growing, all but Japan at double digit rates. Tim Cook also addressed the issue of VPNs in China which I wrote about yesterday, and defended Apple’s stance, which is a combination of following the law in each country where it operates, and believing that it’s better to engage and stay in a country than leave, even where it disagrees with policy (my notes on this portion can be seen here).
Overall, Apple’s management on the call seemed as bullish as they have for some time, clearly looking forward to what they expect to be a strong finish to the year in both product and financial terms. Tim Cook wasn’t drawn the slightest bit on new iPhones, but did hint at new products this fall, talked about the role of autonomy beyond vehicles and Apple’s big project in this area, raved about ARKit and the potential of AR, among other things. There’s clearly a good mix of products coming to market in the near term and investment for the long term which Apple’s management is also happy about. That’s no guarantee of a strong performance in the September or more importantly the December quarter, but I continue to be pretty bullish on what’s coming over the next few months from Apple.