Company / division: Google
Google Turns Image Search into an E-Commerce Funnel (Apr 13, 2017)
Google’s search advertising business is increasingly under threat from other sites pre-empting Google searches with their own search functions in specific areas, among them Amazon in e-commerce and Pinterest in fashion and other categories. As such, Google recently beefed up its image search function to serve up related results from its Shopping feature, and now also shows related images which show fashion products in use alongside other clothing or accessories. All of this is algorithmically generated without human curation, and leans on Google’s AI and machine learning technology. Google is going to have to get better and better at serving up results in these various categories if it’s to fend off the threat from the specialists, but if starting elsewhere has already become a habit for some users, they’ll never even see these Google advances.
via TechCrunch
Burger King TV Ad Attempts to Trigger Google Assistant in Homes (Apr 12, 2017)
This content requires a subscription to Tech Narratives. Subscribe now by clicking on this link, or read more about subscriptions here.
LG Confirms Interest in its Display Business, Doesn’t Mention Google – Android Authority (Apr 12, 2017)
This content requires a subscription to Tech Narratives. Subscribe now by clicking on this link, or read more about subscriptions here.
Google Details Processes but Not Data in Response to Department of Labor Accusations Over Equal Pay (Apr 11, 2017)
Last week, the Department of Labor accused Google of violating equal pay laws, in the context of an investigation into compliance as a result of Google’s work as a federal contractor. Google pushed back hard against those allegations immediately, but has now released a more detailed statement through its blog. That statement outlines the processes Google has in place to ensure fair pay, through the use of salary setting processes in which analysts calculating compensation packages don’t know the gender of the person for whom they are setting the salary, and other mechanisms to ensure fairness. What Google doesn’t do in this post is say what the current ratio of male to female pay is at the company, or share any other numbers to back up its claims, which is a bit surprising. The DoL claims to have found massive disparities in pay and systemic bias against women, so one would have thought the simplest way to rebut those accusations would be sharing some data, which Google hasn’t done publicly (though presumably feels it has done as part of the investigation). The DoL, meanwhile, continues to seek more data which Google refuses to provide, hence the lawsuit. As I said last week, given the issues over diversity and equal pay in the tech industry generally, it wouldn’t be enormously surprising to find that Google exhibited some of the same problems, but if evidence of significant issues does emerge, it would be more damaging to a company of its size than a smaller one with less of a reputation to maintain. So far, though, neither side is releasing data that would allow independent observers to draw their own conclusions.
via Google
Google Home is rolling out support for multiple users – The Next Web (Apr 10, 2017)
This shouldn’t be a surprise to anybody – Google actually showed off what appeared to be multi-user support in its demo of Google Home at I/O last year, but then it turned out the finished product didn’t support it when released in the fall. A little while back, rumors began surfacing that it would add the feature soon (and Amazon too), but that hasn’t materialized yet. The screenshot shared here suggests it’s imminent at this point. This is important, because assistants have to be personal if they’re to be really useful, and most people live in homes with other people, whether family members or roommates, and so things like calendar, email, to-do lists aren’t much use to individuals unless they can be recognized and served up different results. That’s not easy to do, especially because these speakers tend to process voices before recognition takes place, which actually makes it harder to recognize the speaker, but the companies were bound to figure it out eventually. If Google does end up launching this before Amazon, this will be yet another performance advantage, even though its distribution disadvantage remains enormous.
via The Next Web
Google accused of ‘extreme’ gender pay discrimination by US labor department – The Guardian (Apr 7, 2017)
The Department of Labor is suing Google over an alleged failure to adequately disclose its compliance with equal opportunity laws as a federal contractor. During the course of the court case, the DoL has accused Google of having a significant gender pay disparity, something Google strongly denies. Were the allegations to be true, it would be extremely damaging for Google’s reputation as an employer, but given that Google hasn’t given the DoL all the documents it’s asking for, you have to ask whether the Department has a full picture of Google’s pay practices. Google, in turn, has argued that the DoL has gone too far in its request for documents and that it has already adequately complied with the applicable regulations. Recent surveys have shown significant gender and race pay disparities within the industry, so it wouldn’t be surprising if those patterns held at Google too, but it claims its own data shows no such disparity. The court case will presumably eventually come down one way or another both on this question and whether Google has adequately complied with regulations, so it’s worth keeping an eye on how this develops.
via The Guardian
Google Home Vs. Amazon’s Alexa: 54 Questions, 1 Clear Winner – Forbes (Apr 7, 2017)
This is a fun little comparison done by a user in the UK of the ability to the two major home smart speaker units to answer 54 questions. Google Home wins in the end, with 32.5 answered correctly, to 19.5 for Echo/Alexa. The questions were a mix of simple and challenging, and the user was in the UK and asked quite a few UK-specific questions, taking advantage of the fact that both devices recently launched there. But it’s a great illustration of both how Google has the existing skillset to do really well in this category, and also the fact that all these assistants have some way still to go to answer all the questions users might reasonably expect them to deal with.
via Forbes
Google expands fact-checking effort to all searches worldwide – Search Engine Land (Apr 7, 2017)
This is the second fake news-combatting announcement this week, after Facebook’s announcement about teaching users how to spot fake news yesterday. This is one of the broadest and most direct steps Google has taken in this area, and will specifically flag particular news articles or other sites with an additional link to a fact checking site such as Snopes or PolitiFact with a brief summary of who is making a claim and whether those sites consider it to be true. This is somewhat similar to Facebook’s effort to flag fake news, but the big difference is that it will be done algorithmically through special markup those sites will use, which will be picked by Google’s crawlers. That should mean that at least in some cases Google will flag something as false long before Facebook will, and I’d hope that Facebook would move to do something similar over time too.
via Search Engine Land (Google’s blog post here)
Google Develops Federated Machine Learning Method Which Keeps Personal Data on Devices (Apr 6, 2017)
This is an interesting new development from Google, which says it has created a new method for machine learning which combines cloud and local elements in a way which keeps personal data on devices but feeds back the things it learns from training to the cloud, such that many devices operating independently can collectively improve the techniques they’re all working on. This would be better for user privacy as well as efficiency and speed, which would be great for users, and importantly Google is already testing this approach on a commercial product, its Gboard Android keyboard. It’s unusual to see Google focusing on a device-level approach to machine learning, as it’s typically majored on cloud-based approaches, whereas it’s been Apple which has been more focused on device-based techniques. Interestingly, some have suggested that Apple’s approach limits its effectiveness in AI and machine learning, whereas this new technique from Google suggests a sort of best of both worlds is possible. That’s not to say Apple will adopt the same approach, and indeed it has favored differential privacy as a solution to using data from individual devices without attributing it to specific users. But this is both a counterpoint to the usual narrative about Google sacrificing privacy to data gathering and AI capabilities and to the narrative about device-based AI approaches being inherently inferior.
via Google
Gear VR Most-Used Headset, Daydream Barely Registers at VR Porn Site – Variety (Apr 6, 2017)
It’s unfortunate that we have to rely on stats from a porn site to measure VR market share, but beggars can’t be choosers. Obviously, there may be reasons why the usage this site sees isn’t representative of the market as a whole, but the numbers here are far from surprising: Gear VR is by far the largest chunk of usage, which absolutely aligns with the numbers we’re seen in terms of devices sold / in use. Google’s Daydream, meanwhile, has a tiny fraction of the market, which is also unsurprising given its relative newness and the limited distribution of headsets and compatible phones. Gear VR has become the de facto standard for Android VR and mobile VR more broadly, and Daydream VR will only do well if essentially every other Android vendor supports it in their handsets and pushes it aggressively to consumers. So far, that hasn’t happened, with predictable results.
via Variety
WPP Drops as Major Client Unilever Slashes Advertising Costs – Bloomberg (Apr 6, 2017)
I’m tagging this against Alphabet/Google even though there’s no explicit connection here, simply because that company is by far the biggest beneficiary of online advertising spend, and this story suggests that we may see lower spend going forward as some of the world’s largest advertisers start to cut their marketing budgets. As of right now, there are only some early warning signs, with Unilever signaling cuts, but this could easily be the beginning of a much bigger pullback, which could affect not just Google but Facebook and other online ad players. The article cites some other softness in the overall advertising space, and we’ll have to see if this turns into a meaningful drop in spending across the board or just ends up being a minor dip (and whether either of those affects digital advertising specifically).
via Bloomberg
YouTube Cracks Down on Fake Channels by Setting 10,000-View Minimum Before Serving Ads – Variety (Apr 6, 2017)
As I’ve said essentially from the beginning of the advertiser boycotts of YouTube, one easy way for the company to resolve at least some of the issues would be to raise thresholds on the channels and videos that could carry ads, and it looks like YouTube is now taking baby steps in this direction, albeit in the apparent context of impersonation rather than other content issues. Channels will now have to earn 10,000 views of their videos before they can fully join the YouTube Partner Program and begin serving ads, which should help weed out some troublesome channels before they get to the point of monetization. As of right now, that should have only a tiny effect on the ability of creators in general to monetize their YouTube activity – 10,000 views generate such a minimal amount of ad revenue that this isn’t going to hurt anyone’s ability to make money. As long as the threshold stays at this low level, then, this might be a relatively painless way to introduce at least a low bar to monetization on YouTube.
via Variety
Apple fans, Android world scramble to patch Broadcom’s nasty drive-by Wi-Fi security hole – The Register (Apr 6, 2017)
There are two interesting things here, both worth discussing briefly. Firstly, Broadcom, which provides chips for many popular smartphones including the iPhone, has a vulnerability in its WiFi element which can be hacked, allowing a way into the device. Apple issued a patch this week to deal with the issue, and Android vendors will be working to close the vulnerability too, though there’s no specific timeframe yet, highlighting yet another challenge with Android’s fragmented ecosystem. The second thing that’s interesting here is that the vulnerability was discovered by Google’s Project Zero team, which is set up to discover and fix vulnerabilities like this, and has been doing great work lately doing just that, including on non-Google devices like the iPhone. Vulnerabilities like this are always worrying, and it’s great to have Google out there with what seems like a strong team detecting these and notifying vulnerable vendors so they can patch the issues.
via The Register
Google Shares Performance Characteristics for its Machine Learning Chip (Apr 5, 2017)
It’s time to roll out that old Alan Kay maxim again: “those who are serious about software should make their own hardware”. Google started working on its own machine learning chip, which it calls a Tensor Processing Unit or TPU, a few years back, and has now shared some performance characteristics, suggesting that it’s more efficient and faster than CPUs and GPUs on the market today for machine learning tasks. While Nvidia and others have done very well out of selling GPU lines originally designed for computer graphics to companies doing machine learning work, Google is doing impressive work here too, and open sourcing the software framework it uses for machine learning. As I’ve said before, it’s extremely hard to definitively answer the question of who’s ahead in AI and machine learning, but Google consistently churns out evidence that it’s moving fast and doing very interesting things in the space.
via Google Cloud Platform Blog
YouTube TV Reviews Suggest Nice User Interface, Features, Limited Content (Apr 5, 2017)
It looks like the embargo on reviews of YouTube TV lifted this morning and a slew of those reviews have now emerged. On balance, the reviews seem positive about the user interface and features (notably the DVR), but note limitations across content, including the complete absence of Turner and Viacom channels, geographies (the service is only available in five cities to start, due to local broadcast rights), and devices (only Chromecast and Google Cast-enabled TVs, no Apple TV or Roku). I’ll link back here to my first take on YouTube TV following the launch event, because my overall sense of this service hasn’t changed. But it does seem as though the app does better on some key concepts than the other offerings already out there, notably that DVR feature and the way it taps into favorite shows and sports teams. As you’d expect from a brand that’s always stood for ease of use and discovery in video consumption, it’s good at those things. But those limitations are going to mean it’s not a viable competitor in the vast majority of the country for now. The limitations are, in fact, probably a good thing in that they’ll allow YouTube TV to avoid some of the scaling issues suffered by DirecTV Now when it launched last fall, but it’s going to need to find a way to go national sooner rather than later, which means working out those complicated local affiliate rights which have bedeviled everyone else in this business too.
via Business Insider (more reviews on Techmeme)
Google Plots New Hardware to Take on Echo — The Information (Apr 5, 2017)
What do you do if you have two separate hardware products for the home which are selling modestly but not fantastically and have some common elements? You combine them, of course, and so Google is apparently considering a future device which would bring the features of its Home and WiFi devices together in a single unit. That would lower the combined cost and depending on the price potentially also increase the attractiveness relative to either the standalone Home or WiFi devices as they exist today. Given that a single unit of either item today costs $129, it’s entirely feasible that Google could combine the two in a new unit that would still be price competitive with the Amazon Echo while offering a lot more functionality, so this is an interesting angle. But Google Home’s main challenges continue to be less about price and more about name recognition and distribution – the Echo captured the early interest in this space and quickly became the market, heavily leveraging Amazon’s retail distribution channel, while Google continues to struggle to get adoption for its version. Though this move may help spur sales, I don’t think it’s going to lead to the kind of step change Google needs to be a more meaningful competitor.
via The Information
Google Launches Android Patent Licensing Alliance PAX with LG, Samsung, Others (Apr 3, 2017)
Google has today announced a patent licensing alliance which is intended to provide cover to member companies using each other’s patents. The idea is that any member can use any other member’s patents without fear of being sued, something that’s actually been quite common between members of the broader ecosystem over the last few years. The alliance has only nine members to start with, about half of which are smaller smartphone brands, but the members do include Samsung, LG, and of course Google itself, as well as Foxconn. Those members alone apparently have 230,000 patents between them which will now be freely available to other members within the context of Android devices. This is a fascinating move, and it’s impressive that Google was able to get Samsung and LG in particular on board without also having some of the other big Android vendors. Of course, none of this will stop these companies from suing those outside the Android ecosystem (or this alliance), but it might help temper some of the animosity that has sometimes characterized competition between Android OEMs.
via Google
Android overtakes Windows as the internet’s most used operating system – TechCrunch (Apr 3, 2017)
This is an interesting counterpart to last week’s item about revenue from Android apps surpassing revenue from iOS apps in 2017. That news had been a long time coming, because Android has long since been way out in front of iOS in terms of user numbers, but revenue for developers has lagged anyway. This week, the news from Statcounter, which measures online traffic, is that Android has surpassed Windows as the source of the greatest share of online traffic by operating system. That, too, is likely a lagging indicator for the number of people using Android versus the number of people using Windows, but for a different reason – much of online usage on mobile is in apps, whereas on a PC it’s almost all web-based, so PCs will always over-index on web usage relative to mobile devices. There’s a good chance that Android has hundreds of millions more users than Windows already.
via TechCrunch
Google says its YouTube ad problem is “very very very small” but it’s getting better at fixing it anyway – Recode (Apr 3, 2017)
There are one or two interesting data points in here, with the most interesting probably being that videos big advertisers “had flagged received less than 1/1000th of a percent of the advertisers’ total impressions” – in other words, the problematic videos were around one in 100,000 or less of the videos where ads appeared (it’s worth noting that this number only relates to the videos flagged by brands – there may have been quite a few more they didn’t find). The other interesting thing here is the suggestion Google exec Schindler makes several times that there’s some unnamed person behind the recent attention this issue is getting: there was a report recently that someone who developed detection technology for videos was pushing the story as a way to get attention for that technology, and I wonder if Schindler is hinting at that without being explicit and thereby drawing more attention to the issue. The last thing worth noting is that Google is now allowing outside firms like DoubleVerify and comScore to start auditing ad placement, which is something that third parties have been wanting. The issue is definitely fading from the headlines as the stream of advertisers announcing boycotts dries up, but it certainly hasn’t been dealt with definitively, and as I’ve argued, as an issue it goes well beyond just YouTube and affects programmatic buying more broadly as well.
via Recode
Chase Had Ads on 400,000 Sites. Then on Just 5,000. Same Results – The New York Times (Mar 30, 2017)
This is a great illustration of the broader fallout from the recent YouTube / Google backlash from advertisers: in this case, JP Morgan is dramatically narrowing the number and range of sites on which it will advertise, and it’s about to do the same thing with YouTube as well. In both cases, it’s taking that action itself to whitelist sites and channels rather than relying on any Google tools to do so. In other words, even if Google doesn’t choose to solve the problem by limiting ads to a smaller number of sites or YouTube channels, advertisers may go ahead and do so anyway. That’s going to be bad for the long tail of creators and sites, who may see both overall ad volumes and prices drop significantly. This isn’t nearly over, and even if Google and YouTube address some advertiser concerns, I suspect we’ll see ongoing fallout and a narrowing of where big brands advertise going forward. See also this post from music video service Vevo, which is making a pitch that its YouTube channel and videos are a safe space within the platform.
via New York Times