Tech Narratives is a subscription website, which offers expert commentary on the day's top tech news from Jan Dawson, along with various other features, for $10/month. If you are not a subscriber (or are not signed in), some of the content below will be unavailable. If you would like to learn more or subscribe, you can do so on this page. Older comments are available for free.
Snapchat Is Justifying Its $20 Billion Valuation by Emphasizing User Engagement – Bloomberg (Jan 20, 2017)
This is in some ways part of the same story we heard a while back about Snap positioning itself as another Facebook rather than another Twitter. Facebook is all about engagement, providing numbers on not just monthly but daily active users, and talking up time spent as well. Though Twitter briefly dabbled with metrics like timeline views as an indicator of engagement, it quickly abandoned that metric and has steadfastly refused to provide daily active user numbers, focusing instead on MAUs and directional rather than absolute measures of engagement. Snap clearly wants to avoid those associations with Twitter and so has provided to investors data on engagement across several dimensions, which will hopefully be made available in its S-1 filing when that’s made public too. A key part of the Snapchat value proposition is that its users spend a lot of time on the service, so proving that to investors will be critical.
Proposed state law would require emissions-free autonomous vehicles, and tax them by the mile – The Boston Globe (Jan 20, 2017)
As regulators and governments seek to provide a legal framework for autonomous driving, we’ll see something of a dichotomy between those who try to be as welcoming as possible to experiments and development of the technology, and those who see this as yet another opportunity to drive tax revenue or other separate goals. These lawmakers in Massachusetts seem to be taking the second approach, proposing that autonomous cars pay a per-mile road tax and produce zero emissions. Contrast this with Arizona’s governor, who has been very open to testing of autonomous vehicles and famously invited Uber’s self-driving cars to his state when they were banned from San Francisco. San Francisco, of course, is somewhere in the middle, largely open to testing of the technology, but with reasonable limits. Just as there will be fierce competition between tech companies around autonomous driving, there will be competition between states and cities around the technology too. Policies such as those being advocated in Massachusetts are likely to do little to endear the state to would-be autonomous driving companies.
AT&T’s Streaming Service DirecTV Now Peaking At 35,000 Simultaneous Users – StreamingMediaBlog.com (Jan 20, 2017)
Update: AT&T has now released official numbers, with over 200k paying users. So it appears Dan’s estimates were a little short. Though given that AT&T offered a free Apple TV for those who committed to three months of service, it’s possible some of those users aren’t active and will churn shortly.
Dan’s very good at what he does, so I have no reason to doubt that he’s in the right ballpark here, and these numbers are interesting in their own right. What’s even more interesting is how poorly this service has performed, and how unapologetic AT&T has been about it. I met with Enrique Rodriguez, the CTO for AT&T’s Entertainment Group, at CES, and although he acknowledged there were issues, he downplayed them. I have had better luck than some with the service once the first few days were over, but many people are still clearly having lots of issues, which is just baffling for something AT&T talked up so much ahead of time. Moreover, the platform AT&T is using for DirecTV Now is the same one it plans to use for Sunday Ticket online, its TV Everywhere services, and more going forward. I’d hope things start to change quickly here, because the way things are going right now this doesn’t look pretty.
This is NHTSA’s report on the Tesla Autopilot crash in May 2016, which was investigating whether the Autopilot system was at fault. The headline from Tesla’s perspective is that the Autopilot system wasn’t at fault, because it (a) operated as expected, and (b) wasn’t intended to be able to avoid such cross-traffic collisions. That’s good for Tesla, because it exonerates its system, and also because NHTSA determined that its Autosteer system increases safety by 40%. Incidentally, the report also classifies Autopilot as a Level 2 system, whereas I’ve seen some people incorrectly refer to its as Level 3. The key here is that Level 3 systems allow the driver to stop paying attention, whereas Level 2 systems require full driver attention at all times. The problem in this crash was that the driver treated the system as a Level 3 system (which the term Autopilot somewhat implies), and paid insufficient attention to notice the truck crossing the car’s path. Tesla’s system may not have been at fault, but there’s a reasonable argument to be made that it’s not doing enough to train drivers not to treat its Level 2 system as something more – though NHTSA didn’t address that point in its report.
via NHTSA Finds No Fault in Tesla Autopilot With Regard to May 2016 Fatal Crash – NHTSA (PDF) – see also news coverage of the report on Techmeme
Twitter users are diverse but not its staff – USA Today (Jan 19, 2017)
Lots of the coverage of Twitter’s new diversity report (and accompanying blog post) today was positive – the company has made real progress over the past year in increasing its diversity and achieving its own goals, though it also fell short in some areas. However, Jessica Guynn at USA Today has been one of the most active reporters on the diversity beat and doesn’t think the gains are good enough – for one thing, Twitter didn’t share last year’s equivalent report, so it’s impossible to know exactly what gains have been made over the past year alone (the last report released publicly was the 2014 one). But the overall numbers for some minority groups are still very small relative to their share of the US population overall. Kudos to Twitter for increasing its diversity, but kudos to Guynn too, for holding not just Twitter’s but all tech companies’ feet to the fire on this issue.
Errors in Facebook ad metrics could lead to more independent audits – Silicon Valley Business Journal (Jan 19, 2017)
This is the fallout from Facebook’s series of admissions towards the end of 2016 about its metrics relating to both content and ad performance: major advertisers are now going to be calling for more third party auditing of ad performance on Facebook. To the extent that Facebook is already said to be working with some outside groups on this, that effort needs to accelerate and come to a rapid conclusion to satisfy advertisers. On the other hand, it’s also clear from the same survey that Facebook is far from the only company whose ad metrics are mistrusted by advertisers – only Google has the confidence of over 50% of buyers, while AOL has the confidence of just 26%. But having said all that, advertisers don’t seem to feel they have alternatives to the big two, on which they plan to continue spending more money this year than last.
Your real-world purchases will soon determine what ads you see on Snapchat – Mashable (Jan 19, 2017)
Here’s further evidence that Snap is evolving Snapchat’s advertising targeting capabilities, something it badly needs to do to ramp up ad spending ahead of a potential IPO. But that also means going back on some of the commitments Evan Spiegel has made in the past to avoid “creepy” targeting. The reality is that Snapchat has captured a nice little share of ad spending purely on the basis of having a great target market for a certain generation at a general level, but if it wants to capture more targeted advertising, it needs to provide the tools that Facebook, Google, and others already provide. That means buying in data from Oracle (as in this deal, and further to a previous deal with Oracle for measuring ROI) or other data collection houses (as Facebook already does) in order both to target advertising and to capture information about subsequent purchases to prove an ROI. Though Snapchat’s target market is generally more open to ad-based business models and the attendant privacy implications, there’s a point at which even millennials will balk, and Snap has to be careful not to cross that line.
Uber has often been willing to flout regulation in order to stake a foothold in a market, at which point it typically turns its customers into advocates and makes arguments about the contribution it’s making to the local economy in a bid to win formal approval from local authorities. This case brought by the FTC alleged that Uber had exaggerated those benefits significantly – it claimed NYC Uber drivers earned a median income of over $90,000, but the FTC found that under 10% of drivers earned that much, for example. Because Uber settled the case without admitting formal wrongdoing, there is no legal confirmation here that Uber lied, but that almost doesn’t matter. To the extent that Uber gets a reputation (accurate or otherwise) for lying about its economic benefits, its whole “better to ask for forgiveness than permission” strategy starts to break down.
This is the latest in a string of occasions when Facebook has blocked specific content or an entire account on the basis of a supposed violation of its terms, only to reverse itself. But in this case, it’s a bit different – RT is a highly controversial Russian state-funded news outlet at a time when Russian interference in the US electoral process is a hot topic. The account’s privileges were quickly reinstated in this case, but there now appears to have been no legitimate reason to withdraw them in the first place, raising questions about who at Facebook made the decision to suspend the account and why. At a time when Facebook is trying to be more responsible about policing fake news and also working more closely with news organizations, this kind of thing won’t inspire a lot of confidence either among news organizations or among those inclined to belief Facebook’s fake news clampdown has a partisan bent.
This piece is an interesting counterpart to a couple of others I’ve recently linked to – another quoting Wave7 estimates for Pixel sales, and this WSJ analysis from earlier today on how hard Alphabet has been pushing sales of its hardware on Google search, given what this piece says about heavy TV advertising by both Google and Verizon around the Pixel. It’s also worth reading this Verge piece, which takes a much harsher stance on what these sales numbers and the supply shortages mean, though it focuses almost exclusively on the 128GB model. The point is, Pixels are in short supply, and there’s an estimate in this Bloomberg piece of around half a million sales, so this is a very different supply shortage from Apple struggling to meet demand for over 70 million phones per quarter – in other words, this isn’t about hitting up against theoretical maximum capacity for building phones, but about very conservative planning on Google’s part. Half a million isn’t bad, but it’s fairly clear sales could have been a lot higher with better supply. Presumably Google will learn from this experience as it looks to update the Pixel, possibly later this year.
For all the recent brouhaha about Apple’s commitment to pro users, and allegations that the company has somehow lost its knack for creating innovative, compelling products that “just work”, it’s always good to see something of a reality check. The survey covered in this opinion piece (results in PDF here and a somewhat vague explanation of the survey here) has Apple placing top in every category in which it competes: smartphones, laptops, tablets, and online music services. This is the kind of stuff Tim Cook loves to cite on earnings calls, because it’s the best possible validation that Apple is doing its job well. For all the criticisms from specific groups, whether that’s creative professionals or their would-be spokespeople in the media, it’s this kind of broad-based success among regular people that Apple has always been committed to, and which I suspect is its north star today. This sort of brand engagement and loyalty is the key driver of sales, profits, and growth, and I would expect Apple to continue to try to create products that generate this kind of broad-based brand value and loyalty.
Amazon launches Pivot a new training program to help employees in danger of being fired – Business Insider (Jan 19, 2017)
Amazon has gained a reputation over time for treating its employees poorly – the New York Times famously did an in-depth investigate piece on this topic as it relates to Amazon’s white collar employees, and it talked about Amazon’s Performance Improvement Plan for underperforming employees. This article talks about a new program intended to benefit those placed on a PIP by helping them develop their skills, and can be seen as an effort by Amazon to help those with poor evaluations rather than merely taking the first steps towards an eventual dismissal for cause. The PIP process was far from the only element of working in a white collar job at Amazon that the New York Times wrote about, and of course Amazon pushed back against some of the other allegations in the report. And then there are the working conditions in blue collar jobs at Amazon’s warehouses and fulfillment centers. So this is part of changing the narrative, but only really addresses one small piece of it.
Google Uses Its Search Engine to Hawk Its Products – WSJ (Jan 19, 2017)
This is a really fantastic bit of a analysis commissioned by the Wall Street Journal. It found that for 91% of searches relating to top consumer electronics categories, a Google or Nest product was in the top ad slot above the results, and in 43% it had the top two slots. This is Google competing with its own advertisers, and Google apparently was so taken aback by the analysis that it tweaked its strategy when the WSJ spoke to Google about it, and the numbers are now much lower. Google’s hardware push therefore not only puts it in conflict with its OEMs, but also with its biggest customers – advertisers. I’m intrigued to know how other big consumer electronics brands feel about this, but the challenge of course is that they have few alternatives – Google dominates search, and so it also has a dominant position in pitching its own products. There’s a close analogy here to Amazon’s hawking of its hardware on Amazon.com, but competitors know what they’re getting into there to a greater extent.
Minecraft Pocket Edition will no longer receive updates for Windows mobile devices | Windows Central (Jan 19, 2017)
It’s becoming ever clearer that Microsoft is going to go through yet another revamp of its smartphone OS strategy, on top of Windows Mobile and Windows Phone (you could argue Windows 10 Mobile as it currently stands is already a third, but it was very much an update to Windows Phone in reality). Other developers have been abandoning their Windows Phone apps for some time now, but for Microsoft itself to do it is the clearest sign yet that there’s no meaningful future for Windows on consumer mobile devices as it currently stands. The article holds out hope for the full Windows on ARM strategy Microsoft is working on, but it still seems very likely that this will end up being a marginal and enterprise-centric play rather than something that gets Microsoft back into the consumer smartphone market.
Normally I’d link to a company’s own report on its earnings, but since Netflix’s earnings material is all in non-web file formats like PDFs and Excel spreadsheets, I’m linking instead to the Techmeme cluster of articles on the earnings report. Broadly speaking, this is a great set of results for Netflix – subscriber growth both domestically and internationally was higher than it forecast, with domestic growth bouncing back nicely now after a couple of tough quarters in which price increases were a drag on net adds. The international business is nearing profitability, though Netflix will invest to keep it just in the red in 2017, and margins expanded nicely domestically thanks to those price increases. With short-term growth concerns somewhat alleviated, the main focus returns to Netflix’s content spending and whether it’s sustainable. It had a non-GAAP free cash flow loss of $639m in Q4 and $1.7bn in 2016 as a whole, both massively up from the year before as it invests in original content, which has to be paid for upfront. Over time, that much higher investment will flow through into the P&L too, and continued strong growth is critical for staying ahead of those costs.
You may also be interested in the Netflix Q4 2016 deck in the Jackdaw Research Quarterly Decks Service.
A lot of the coverage of the lack of diversity in the tech industry focuses on employment and the lack of opportunities and barriers to entry for minorities and women. In other words, the focus is on the negative impact on those who would like to work in the industry. But this article highlights one of what I’d argue are many practical reasons why this lack of diversity is also bad from a product perspective – less diverse teams produce products which are poorer at meeting the needs of a diverse base of users. In this case, the specific issue is face recognition software and its inability to effectively recognize darker faces, in part because it tends to be trained on data sets of largely white faces and tested by mostly white engineers.
One of the big challenges for Snap ahead of its IPO is ramping up advertiser spending from experimental to more serious levels, and WPP is one of the biggest funnels through which those ad dollars flow. $90 million is three times what WPP expected its clients to spend on Snapchat at the beginning of the year, but of course it’s a tiny fraction of what those clients spend on Facebook and Google. Getting advertisers to spend more means providing more of the tools and analytics other large online ad platforms provide, and Snap has been somewhat resistant to that approach, although it appears to recognize the need to evolve on this point going forward.
Twitter canceled its annual developer conference, Flight, back in October, and I posited at the time that this would send a strong message to developers working with Twitter, though I got pushback from some people at Twitter. Now, Twitter is selling its developer tools (collectively known as Fabric) to Google, rather validating my initial take on the Flight cancelation (no pun intended). This is certainly a result of Twitter’s narrower focus going forward on user-facing, live content including video, but it reinforces the sense that Twitter has really messed around with developers over its history. Developers will still be able to create and run apps for Twitter, but Fabric was a big part of Twitter’s developer toolset and a major focus of those Flight conferences in the past. Lots of those tools, though, had little to do with the core Twitter product, so there’s definitely some logic in selling it a company – Google – which is committed to providing a broad set of generic tools to developers.
One of the criticisms of Apple which has become loudest lately is that it is increasingly ignoring the professional creatives who use Macs to do their work, and I’ve seen this not just in relation to Apple’s Mac lineup but also a supposed neglect of Apple’s pro apps. However, at the MacBook Pro launch event a couple of months ago, Apple provided a big update to Final Cut Pro, which I’m told by video pros is a big deal, and now we’re seeing a big update to another of Apple’s big creative apps, Logic Pro. While I think some of the Mac criticism is reasonable (though I still think we’ll see an update on the desktops soon), this stuff about the pro apps clearly isn’t true – Apple is still investing in a big way here.
Apple iPhone 8 rumors: Features may include facial recognition, laser sensor – Business Insider (Jan 18, 2017)
Cowen doesn’t have the same track record in predicting future iPhones as KGI, which has by far the best, so we should take all this with a pinch of salt. But it’s in keeping with the broad sense that Apple is very interested in augmented reality, and would need to put more sensors and other technology into its products to enable AR functions. I’m still intrigued by the idea of further splitting the iPhone line – there are already three sizes, and this research note posits a fourth, larger one, with exclusive access to an OLED screen and embedded fingerprint sensor. There’s some logic to that, because all the supply chain chatter suggests Apple would have a very hard time finding enough OLED technology to power all of the next generation of iPhones, so making it exclusive to the highest end device would limit demand to a smaller number. Even so, that device is likely to be in high demand, as was the 7 Plus with Jet Black finish, another phone with supply constraints.