Alphabet Inc - Class A (GOOGL) — Q3 2016 Earnings Call Transcript

Thank you. Good afternoon, everyone, and welcome to Alphabet's third quarter 2016 earnings conference call. With us today are Ruth Porat and Sundar Pichai. While you've been waiting for the call to start, you've been listening to Dua Lipa, a rising new pop star from London, whose most recent single on YouTube has found fans all over the world and cracked the top 40 in the U.S. ahead of her debut album release early next year. Now I'll quickly cover the Safe Harbor. Some of the statements that we make today may be considered forward-looking, including statements regarding our future investments, our long-term growth and innovation, the expected performance of our businesses, and our expected level of capital expenditures. These statements involve a number of risks and uncertainties that could cause actual results to differ materially. For more information, please refer to the risk factors discussed in our Form 10-K for 2015 filed with the SEC. Any forward-looking statements that we make are based on assumptions as of today, and we undertake no obligation to update them. During this call, we will present both GAAP and non-GAAP financial measures. A reconciliation of GAAP to non-GAAP measures is included in today's earnings press release. As you know, we distribute our earnings release through our Investor Relations website located at abc.xyz/investor. This call is also being webcast from our IR website, where a replay of the call will be available later today. And now I'll turn the call over to Ruth.

Thank you, Ellen. Our revenue of $22.5 billion in the third quarter highlights the excellent performance of our global businesses. For this quarter, our consolidated revenue increased by 23% in constant currency compared to last year, despite a tough year-on-year comparison. Once again, Mobile Search was the main contributor, supported by strong results from YouTube, along with key contributions from programmatic advertising and Play. I'll present the information in the following order: first, a consolidated overview for Alphabet; second, the results for Google and Other Bets individually; and finally, our outlook. Sundar will then share our business and product highlights for the quarter, followed by a Q&A session. To start with the summary of Alphabet's consolidated financial performance, total revenue reached $22.5 billion, reflecting a 20% year-over-year increase and a 4% sequential increase. We faced a negative impact of $196 million on our revenues year-over-year due to currency fluctuations, which was $91 million after accounting for our hedging program. Holding currency steady against previous periods, our total revenue grew 23% year-over-year and increased 5% sequentially. The geographic breakdown of Alphabet's revenues demonstrates the robustness of our businesses worldwide. U.S. revenue rose by 22% year-over-year to $10.6 billion. UK revenue increased by 5% year-over-year to $1.9 billion, impacted significantly by the decline in the British pound compared to last year. In fixed currency terms, the UK revenue grew by 18% year-over-year. Revenue from the rest of the world was up 22% from last year to $9.9 billion and increased by 25% year-over-year in fixed currency terms. GAAP other cost of revenues was $4.5 billion, up 30% year-over-year. Non-GAAP other cost of revenues was $4.2 billion, an increase of 29% year-over-year, primarily due to Google-related expenses, including costs for operating our data centers and content acquisition costs mainly for YouTube. GAAP operating expenses were $8 billion for the quarter, up 15% year-over-year, while non-GAAP operating expenses were $6.5 billion, up 13% year-over-year. On a GAAP basis, operating income was $5.8 billion, a 22% increase compared to last year, resulting in an operating margin of 26%. Non-GAAP operating income amounted to $7.6 billion, a 24% rise year-over-year, with an operating margin of 34%. Stock-based compensation totaled $1.9 billion, increasing by 30% year-over-year and 24% sequentially, mainly reflecting the annual equity refresh for employees at the start of Q3. At the end of the quarter, our headcount was 69,953, having grown by 3,378 people since last quarter, which is typical for the third quarter as new graduates join. Most new hires were engineers and product managers supporting growth in critical areas such as cloud. Other income and expenses were $278 million, and we provide further details in our earnings press release. Our effective tax rate stood at 16%. Net income was $5.1 billion on a GAAP basis and $6.3 billion on a non-GAAP basis. Earnings per diluted share were $7.25 on a GAAP basis and $9.06 on a non-GAAP basis. Regarding CapEx and operating cash flow, CapEx in the quarter was $2.6 billion, primarily supporting the Google segment. Operating cash flow was $9.8 billion, with free cash flow reaching $7.3 billion. We concluded the quarter with cash and marketable securities totaling $83.1 billion, of which around $50 billion or 60% is held overseas. Now, turning to our segment financial results, let's start with the Google segment. Revenue was $22.3 billion, reflecting a 20% increase year-over-year, including the impact of foreign exchange. Breaking down the revenue details, Google Sites revenue was $16.1 billion, which was up 23% year-over-year and up 4% sequentially. Year-on-year growth was driven by Moblie Search. We also see solid growth from desktop and tablet search. YouTube revenue continues to grow significantly, mainly due to video advertising across TrueView, with growing contributions from purchases on DoubleClick Bid Manager. Network revenue was $3.7 billion, up 1% year-over-year and stable sequentially, reflecting the strong growth of programmatic and AdMob, offset by traditional network businesses. Other revenue for Google was $2.4 billion, a 39% increase year-over-year and a 12% sequential increase. This growth was a result of contributions from Play and Cloud. We provide monetization metrics to give insight into the pricing and volume dynamics of our advertising businesses, which can be affected by currency fluctuations. Total traffic acquisition costs were $4.2 billion, or 21% of total advertising revenue, up 17% year-over-year and 5% sequentially. The rise in both Sites TAC as a percentage of Sites revenue and network TAC as a percentage of network revenue indicates that our strongest growth areas, Mobile Search and programmatic, have higher TAC. Total TAC as a portion of total advertising revenues saw a slight sequential increase due to higher TAC for Mobile Search, offsetting the benefits from the shift in revenue mix from network to Sites. Operating income excluding stock-based compensation was $8.4 billion, representing a 19% increase from last year, resulting in an operating margin of 38%. With stock-based compensation included, operating income was $6.8 billion, a 17% rise year-over-year, with an operating margin of 30%. For the quarter, Google’s CapEx was $2.4 billion, focusing on investments in production equipment, facilities, and data center construction. Turning to Other Bets, as previously stated, it is most useful to consider the financials for Other Bets over the long term, given the variability in quarterly revenues and expenses. These fluctuations are primarily due to the early-stage nature of Other Bets, the diversity of businesses involved, and potential one-time events like partnership arrangements. In the third quarter, Other Bets generated revenue of $197 million, mainly from Nest, Fiber, and Verily. The operating loss excluding stock-based compensation was $665 million for the quarter. Including stock-based compensation, the operating loss was $865 million. Capital expenditures for Other Bets were $324 million in Q3, primarily for ongoing investments in our Fiber initiative. I'd like to conclude with some observations on our progress since the formation of Alphabet just over a year ago and a review of our key themes. As we've consistently mentioned, our transition to Alphabet stems from our belief that groundbreaking ideas will drive our next growth areas. Achieving long-term success necessitates a commitment to making bold bets, placing the right talent and resources behind those initiatives, and remaining adaptable as we pursue them. We believe our structure offers the transparency and oversight required to make informed decisions about our investment opportunities within Google and across Other Bets. As we strive for significant breakthroughs that will have a lasting impact, some course corrections are inevitable and not all efforts will succeed. Over the past year, you’ve seen us make advancements in certain areas while revisiting or pausing efforts in others. We are strategically laying the groundwork for a stronger future. Looking forward, our revenue growth reflects our ongoing investment in innovation, primarily within Google. This unyielding focus has led to advancements in our advertising platforms that continue to drive substantial growth on an already large scale, while we are also building new businesses that will serve as future revenue sources. Notably, Google Cloud is seeing substantial revenue growth, underlining the ongoing momentum in this business and the vast opportunity within this sector. Earlier this month, we introduced a new line of hardware devices that, for the first time, brings together both hardware and software developed by Google for consumers. As discussed previously, since most of our Other Bets are not yet generating revenue, the Other Bets revenue line offers only a partial view of our progress, which we will complement with insights regarding product advancements. For instance, at Nest, product enhancements, including a new outdoor version of the Nest Cam, are leading to increased consumer interest in its product suite for the home. Our self-driving car team is making impressive strides in transforming mobility, having recently surpassed 2 million miles of autonomous driving with our test cars. We are currently testing in four cities to experience varied weather conditions and driving scenarios. In terms of expenses, as I mentioned last quarter, several factors are driving higher TAC in both our Sites and network businesses, which continued into the third quarter. The shift to mobile in Q3 was the primary factor for the rise in Sites TAC as a share of revenue. We anticipate that Sites TAC will persist in increasing as a share of Sites revenue. The growth in network TAC during Q3 was attributed to the rising adoption of programmatic platforms by advertisers, which incur a higher TAC rate, a trend we expect to continue. Additionally, regarding Google's operating expenses, we remain committed to investing in the valuable opportunities we have identified. As for Other Bets, we are systematically and thoughtfully expanding these businesses, investing in accordance with the opportunities we perceive. Before I conclude my remarks on expenses, I want to remind everyone that marketing costs in the fourth quarter tend to be higher due to the holiday season. Compared to last year, we have an expanded range of hardware products, and we anticipate marketing costs will rise in the fourth quarter to support this lineup. Regarding CapEx, the Google team continues to drive significant efficiencies in planning and operations for our technical infrastructure. For Other Bets, our investments in Fiber remain the primary driver. Finally, looking at our balance sheet, it remains a powerful asset reflecting the strength of our cash flow, allowing us to invest aggressively in support of our long-term growth. Our main focus is on investing across the breadth of opportunities within Alphabet. As mentioned earlier, our capital allocation strategy begins with our outlook for the businesses, alongside sensitivity analyses regarding potential CapEx and M&A, as well as considerations for working capital and maintaining a prudent liquidity buffer. This strategy also takes into account complementary uses, such as share repurchase. As announced today, our board has authorized the repurchase of up to $7,019,340,976.83 of our Class C capital stock. In summary, this third quarter saw us deliver robust revenue growth while expanding our product and service portfolio. We extend our gratitude to all our colleagues worldwide for their creativity and commitment to pushing the limits. I will now hand the call over to Sundar.

Thanks, Ruth. I don't recall a time in Google's history as busy as the past few months. We introduced Google Assistant, utilizing our advanced machine learning technology. We presented a stunning range of new hardware, including Google Home and the Pixel phone, which has received excellent reviews. We launched new apps, Allo for messaging and Duo for video chats. It has been an amazing quarter. Throughout this time, we've remained focused on the priorities I shared earlier this year, and I will briefly review our progress over the past quarter. We are well-positioned as we enter a new era of computing, where people will engage with technology more naturally in their daily lives, supported by intelligent assistants and the cloud. This transition is as significant as the shift from desktops to mobile devices over the past decade. As we have introduced products like Google Assistant, Pixel, Google Home, and Allo, we've glimpsed this future. With our expanding cloud business, we are also helping our enterprise clients capitalize on this new era of computing. Today, I want to discuss our progress in key focus areas and the success of our growing advertising business. First, our goal has always been to make knowledge and information accessible to everyone, helping people find what they need at the right moment, sometimes even before they think to ask. We reached a significant milestone this quarter with the launch of Google Assistant, which enables users to interact with Google naturally, whether typing or speaking, across different devices. We first introduced it in our smart messaging app, Allo, allowing users to chat with the Assistant, even including it in group conversations to help decide on dinner plans. Early adoption of Allo and Duo has surpassed our expectations. The Assistant is integral to both the Pixel phone and Google Home. With Google Home, you can request the Assistant to play music or switch on lights, while with Pixel, pressing the home button lets you ask the Assistant to send texts or access photos. I encourage everyone to try them out soon. The Assistant is user-friendly and just the beginning; its capabilities will continue to grow, and we're excited to introduce new features across more of your contacts shortly. The Assistant's development is a result of our extensive investments in computer science and machine learning. Our knowledge graph now comprehends over 70 billion facts about people, places, and things. Just last month, we unveiled our latest advancements in neural networks, significantly enhancing translation quality, which will provide more accurate translations globally. Before moving on, I want to highlight our team's efforts to keep American voters informed about the upcoming election, where we launched improved search tools for voter registration and key voting deadlines in both English and Spanish. I checked Google Translate, and I want to remind everyone, don't forget to vote. Next, regarding our initiatives to broaden content accessibility, we've been helping users find excellent content across various platforms such as the web, YouTube, and Google Play. Recently, one of our major initiatives to enhance mobile web speed, the open-source Project AMP, celebrated its first anniversary, with a broad adoption from publishers worldwide, covering over 700,000 domains. For instance, tens of millions of WordPress sites now feature AMP pages, improving page load times significantly. We're also introducing AMP for Ads, focusing on creating and delivering fast ad experiences. On YouTube, we are committed to providing engaging content and worked with several news organizations to livestream all three U.S. presidential debates, resulting in record-breaking engagement. The three debates were the most viewed political live streams ever on YouTube, amassing over 8.5 million hours of live viewing, a fivefold increase from the 2012 debates. Election-related searches on YouTube are at an all-time high, almost 550% higher compared to the last election period. YouTube originals like the newly released Single by 30 are highly popular and continue to drive subscriptions. In Google Play, we recently launched Family Library, allowing up to six family members to share purchases. Third, regarding powerful platforms and hardware, we've been dedicated to creating robust platforms like Android, Chrome, and now Daydream for virtual reality, ensuring high-quality experiences across all devices. We believe an open, horizontal platform like Android fosters innovation, offers users more choices, and creates better business models for partners. With over 4,000 distinct Android devices, we're proud of our partners' successes on the platform. This quarter, we rolled out Android Nougat and announced the integration of Android apps on Chromebooks. We're excited about the amazing experiences developers will create on Daydream, our latest platform for top-notch VR experiences. We also see potential in deeply integrating software and hardware to enhance user experiences, which will contribute to future innovation. Earlier this month, we introduced a new range of hardware products, including Google Pixel, Google Home, the Daydream View headset, Chromecast Ultra, and Google Wifi. This array of products exemplifies our ability to blend top-tier hardware and software for the best Google experience. Fourth, our momentum in the cloud sector continues to grow. Our cloud focus has resulted in heightened customer engagement. At our recent Horizon event, we launched Google Cloud, a unique suite of products and services that empower customers to thrive in a digital environment. Google Cloud encompasses the Google Cloud platform, G Suite productivity applications, our data analytics, machine learning tools, and enterprise-grade Android devices for cloud access. Our Cloud team is actively enhancing our offerings, with plans to expand to eight new cloud regions in 2017. We've upgraded Google Hangouts for enhanced media experiences and improved Google BigQuery, our data analytics service, while also making Google Cloud Machine Learning available in beta to assist businesses in training quality models. We're forming partnerships with leading companies, including OCTA, Box, and Accenture. Our machine learning investments present a clear advantage for Google Cloud, enabling clients to implement ML effectively. For example, Ocado, the largest online-only grocery retailer, utilizes machine learning to manage and prioritize customer emails efficiently. As we officially enter the Google Cloud era, our goal remains unchanged: to create the most open cloud for all businesses, facilitating the development of top-quality software. Our team is committed to delivering the best cloud products in the industry for our clients. Finally, let's discuss our flourishing advertising business in a mobile landscape. Our mobile marketing strategy is straightforward and resonates well. Our services like Search, YouTube, Maps, and Google Play are where people turn when actively interested in something. They use our offerings because they want to engage with content they're passionate about or because they seek information or products. They are highly attentive, similar to how audiences would watch television during prime time. Our platforms are the prime time for mobile, making it an optimal moment for brands to run their ads. For instance, hotel chain La Quinta now derives a third of its website traffic from mobile and leverages our hotel ads to connect with travelers at the perfect time. They've discovered that conversion rates are double that of regular mobile traffic, capitalizing on those moments when potential customers are ready to book. YouTube continues to excel, with over 1 billion monthly users viewing hundreds of millions of hours daily. YouTube is the preferred platform for major brands with a highly engaged audience, and recent studies reveal that nearly half of U.S. adults aged 18-54 say that YouTube influences their purchasing decisions at least once a month. One successful ad format is bumper ads, short 6-second videos that help brands expand their reach. Brands like Universal Pictures are seeing fantastic results by pairing these shorter ads with standard-length ads on YouTube, especially on mobile devices. We've also enhanced our capabilities to allow marketers to understand how TV and YouTube campaigns increase brand searches. Early tests reveal that YouTube generates nearly double the searches per impression compared to TV. Measurement is a key investment for our advertising business, as we aim to provide marketers with the best tools to connect digital and traditional marketing strategies. For example, IKEA and their agency iProspect UK utilized our store visit measurement tools to assess the impact of their digital campaigns on store traffic and found that over 10% of those clicking on their search ads visited physical stores. They realized their ROI from online ads was five times higher than they had previously estimated. We introduced more measurement solutions during Advertising Week in New York, including new tools like location extensions for the Google Display Network. We have numerous partners achieving exceptional results with our products. I am pleased with the collaboration among our business teams, advertising partners, agencies, marketers, and publishers. Their success is ours, and we have intensified our efforts to forge deeper partnerships and long-term, mutually beneficial relationships. This is a significant focus for us. To conclude, the past few months have been incredibly exciting, and we've had a very successful quarter. We feel confident as we transition into a new computing era and are eager to collaborate with our partners on this journey. I want to express my gratitude to all the Googlers worldwide who work diligently to turn ideas into reality for our users and partners. It is a privilege to work alongside you all. Now, I'll turn it back over to Ruth.

Thank you, Sundar. We will now take your questions.

Thank you so much for taking the question. Sundar, maybe for you, on the enterprise, thanks for laying out all the vision for the medium to long term on the enterprise and cloud. I think one of the big questions we get from investors all the time are where are the key areas you need to invest going forward in both capabilities, go-to-market, and how those might evolve over the next couple years. And then maybe a second question for Ruth, which is more a housekeeping matter, Ruth, with highlighting the increased marketing expense behind the product launches this year, were you referencing quarter-over-quarter or comparing year-over-year when you also launched products a year ago? Thanks so much.

On cloud, let me talk about a few areas. We have stepped up our partnering considerably so that we can offer our customers everything they want, and we have more vehicles for going to market. So scaling up through partnerships is a big area of focus and investment for us. We are also establishing a large cloud machine learning group so that we can take advantages of working with our cloud customers and make machine learning more accessible to all of them. I would say other areas are hiring across sales, engineering, and marketing. And as we head into 2017, I expect cloud to be one of our largest areas of investment and head count growth.

And then on your second question, there were really two parts to the statement. One, the obvious, is that sales and marketing is seasonally higher whenever we go into the holiday season. And so year on year, I wanted just to accentuate the point that we expect this trend to be pronounced this year because we're expanding. We're launching an even more expanded suite of hardware products. So it's really emphasizing the fact that the fourth quarter is higher and even more so this year with the expanded suite.

Thank you. I also wanted to ask a follow-up on the cloud. I guess I was wondering, Sundar, if you could share with us. What milestones should we be watching for in this business? And then also, if you had to characterize – you mentioned a couple different ones, but I'm wondering. What type of workloads are you having the most success with in GCP, and what's on your customers' wish list? What's the top one or two things that your enterprise customers for cloud might be asking for? Thanks.

Thanks, Heather. In general, the way we see it is that customers don't want to be locked in and they want to make sure their workloads can work in and be managed by containers that run on any platform. So our virtual machines and containers, including excellent open-sourced container management Kubernetes, which was developed by Google, offers that solution. So Kubernetes manages the provisioning, reliability, and auto-scaling of workloads. So we want to pursue a hybrid strategy, including on-prem and in all public clouds. So from a customer standpoint, they will want to use an open source workload management product, and Kubernetes can work with multiple customers. So that's broadly how we think about it.

Thanks, guys, two questions. Sundar, first on the Alphabet structure high level, so there's been some incredible innovation that's come out of Alphabet and the various businesses. If we look at something like Otto, which was acquired by Uber, or Niantic, which was the old Google Maps team, those were teams that were inside of Google that left Alphabet to start a new company and could have been potentially wholly-owned businesses of Google in a parallel universe. So do you think this structure is ideal for entire companies, or is it just new products and new technologies that you see as the vision here for Alphabet long term? And then, Ruth, the second question is just on – core operating margins for Google have declined a little bit versus the prior quarter trend of increasing. Can you give us some color on how much of that was due to mix towards cloud and other things that might carry lower margin versus margin compression within the advertising business? Thank you.

Ross, on the first question, I would say overall when I look across Google and Alphabet, the number of areas where we have been able to build world-class products and achieve scale and success, we today have over seven products which serve 1 billion users each, so I think our track record speaks there. And we generally want to encourage a culture of innovation and that's what we focus on, and I think it is fine that some of them happen outside. So we don't view it as a zero-sum game, and we're very comfortable with how we approach it.

And then in terms of the margin trends, so overall operating margins year on year are up modestly on both a GAAP and non-GAAP basis, and that primarily reflects trends in Other Bets. I think your question was probably more within Google. The operating margin on a non-GAAP basis is down modestly year on year. That's primarily to other cost of sales as distinct from TAC. The Google GAAP operating margins year on year and quarter on quarter do reflect the impact of the equity refresh that we called out last quarter that I talked about. I think one of the core things to your question is just how we're looking at it, and our view continues to be that given the breadth of opportunities and our commitment to long-term revenue growth, as we've talked about, quarter after quarter we do remain committed to investing in this growing set of opportunities. And we've spent a lot of time trying to manage that revenue growth, and as I said, manage expense growth with the utmost respect for the resources deployed and getting the best return on those resources. So there are a number of different factors in here. I think I mentioned one of the things as you're looking forward was the sales and marketing. And just to make another point, which I said in my opening comments, we do expect ongoing gross margin pressure from higher TAC associated with Mobile Search and with programmatic. That does still result in more revenue and gross profit dollars but at a lower margin, and that's the other really important point with our starting point. We are continuing to invest, and we would point you to the revenue and gross profit dollars that come from that.

Thanks. With Google Home and the Pixel phones, are you rethinking at all your go-to-market strategy in terms of distribution and marketing? Google over the years has put out different hardware products. Have there been learnings from that that make you approach those two particular areas in Google Home and the Pixel phone lineup differently? Thank you.

Thanks, Mark. As you point out, we have done hardware products over the years, but we saw an opportunity to bring all our display hardware efforts together in a thoughtful structure. I'm glad Rick is here, and we've been very focused on how we approach everything end to end so that we can bring together software and hardware for a great user experience. So along those lines, both Google Home in Google Pixel are important new efforts for us, and I think we will thoughtfully evolve our go-to-market strategy as well. These are important areas. And you already see with Google Pixel we have a deep partnership at Verizon with which we are going to market in the U.S. And so we are constantly thinking about how to do this well, and you will see new approaches as we go through it more.

Thanks for taking the question. Sundar, you talked about a lot of products across advertising and hardware. I think one thing we did not hear about was Maps. And I think you've talked about it in the past as 1 billion-plus users and pretty massive engagement. So can you talk a little bit more about the monetization strategy here and how you can really sell more advertising here but also preserve the user experience at the same time? And maybe talk about a timeframe that we could see more here. Thanks.

That's a good question. Today, a big part of what you're seeing with our transition to mobile that's working really well. The reason Mobile Search and mobile monetization works really well for us is because a lot of it is inherently local by nature. And the thing which helps us deliver a great local search experience is Maps. And so that's the direction we have always pointed in. We want it to be a great experience for consumers as Google Maps, but in terms of also really enhancing the local search experience. And I think you'll continue to see us perceive it that way. And over time to the extent there are opportunities to create value within the application itself, we'll pursue that as well.

Good afternoon, thanks for taking the question. I wanted to ask one about Voice. Sundar, you talked about the success you've had in improving your natural language recognition, and we know that voice queries are growing quickly. I'm wondering if you could share your thoughts on how the increase in voice queries may or may not impact monetization going forward. Is it a risk, or is the growth of voice queries much more skewed to less commercial activity? Just any thoughts there, thanks so much.

Thanks, Peter. From my standpoint, I look at it as – and if you look back in time, all this is – computing is becoming more and more important to people, and so they are engaging with it more and more. So as we went from desktops to mobile, it's not like one replaced the other. The sum total of all of this, it expanded the pie. I approach this the same way. I think as I see people using Voice, et cetera, they are interacting more with computing and with Google too. So we view this as providing users more access across many different surfaces, many different contexts, being there for them when they need it. So in that view, I think it will all be a big positive for us going forward.

Thanks for taking my question. I have two. Just to go back to the Voice Search question, Sundar, could you just talk a little bit about if we do continue to migrate toward a world of Voice Search, just talk about infrastructure, potential build you think you need to put in place to continue to monetize search in a voice world as well as you do in a phone or a desktop world. And then secondly, could you just talk a little bit about where you are on the rollout of extended text ads? I know you talked about a 20% bump to click-through rates. I was just curious for where you are now, a push-out, and is there any update on the click-throughs? Thanks.

On the first question, I think I briefly answered it before. In terms of – we are thinking about the voice experience deeply end to end. A lot of it is going – because of how we are approaching our core investments both from a software and hardware standpoint, so we leverage it all to make it better. So for example, even things like TP usage we talked about at Google I/O play a role into something like Voice Search you're talking about. And as we evolve the Google Assistant, I think voice is going to play a major role that way as well. So we are in very early days of all of this in relation to our overall volumes we see, and so I think we will be thoughtful about it. For example, the Assistant team talked about conversational actions as a way by which we can integrate third parties into the Voice experience. So it's early days, and I think we will evolve it a lot in the coming years. In terms of your second question around expanded text ads, I would say again it's early. It's being adopted by advertisers across the board. We see both large and small advertisers using it. For us, we find advertisers who actually spend their time being thoughtful about the ad creatives and extensively testing and optimizing for this new format, they find strong performance. So we are pleased with the progress so far, but the transition is going to take some time as advertisers get comfortable with it, but we are excited about it.

Good afternoon and thanks for taking my questions. First one for Ruth, heading into this quarter's results, you had called out the ad format change in the third quarter of last year, which had driven a step up in the year-ago growth rate. I wonder, is there anything specific on the revenue side that you'd like to call out in terms of comparisons versus the fourth quarter of last year as we try to calibrate our revenue expectations for the fourth quarter, either in terms of Mobile Search, YouTube, or ad tech programmatic? And then my second question is either for Ruth or Sundar. Can you just talk about the decision to pause efforts for Google Fiber? I'm wondering, was that decision more about financial discipline, as you I think explained, Ruth, in your prepared remarks for the Other Bets, or does it more have to do with the shift to wireless Internet technology or point-to-point wireless, which ultimately can take the place of a Fiber or facilities-based infrastructure over time? Thank you.

Okay, great, two good questions. So in terms of Sites revenue, we are very pleased by the strength. It was broad-based. It had two real drivers, first, our ongoing focus on improving the experience for advertisers and users. And we introduced a number of enhancements, including format changes in ad tools. I think the most important point is that no one enhancement came close to the magnitude of the changes that we made in the third quarter of last year. That's why we kept calling out the 2015 change. And what's been gratifying is in the aggregate the results reflect the benefit of ongoing innovation. But very importantly, it wasn't one particular item. The second contributor is the secular shift to mobile, and we continue to benefit from that, be an important part of enhancing opportunities for engagement. As we look forward to the fourth quarter, I think that was part of your question as well. I think the only thing to point out is that in looking at growth rates, we're obviously at a higher revenue base versus last year, so that to us is an important point. It may be obvious, but important to note there. In terms of Fiber, the impetus for it was really about the opportunities that we see to focus on innovation, and what does that mean if the objective with Other Bets is really these 10X opportunities. And when you go back to the initial impetus for creating the business, it was the founders' view that there's a sizable opportunity given the need for abundant connectivity on networks that are always fast and always open, and we do continue to be committed to that vision. The team had some important breakthroughs in new technologies. You noted the most important in our view, all that we're doing with wireless, but also technologies that are key to implementation. And we believe that both of those, a number of things they're doing enhance both our effectiveness and efficiency. And so we wanted to focus on the potential with these efforts before we reaccelerate deployment. And it was about ensuring that we can take advantage of those before again pushing forth. We were very active in a lot of cities in the third quarter alone. We rolled out four new cities, so that brings us to 12 cities across the U.S. where we're deployed, in construction, or in development, and we're making great progress in those cities. We remain very committed to growth across those cities. And then we also have a presence in six cities with our wireless acquisition, Webpass. So we're pausing for now our work in eight cities where we've been in exploratory discussions. But very much to your question, it's to better integrate some of the technology work we've been developing, and there's more detail on the cities on the Fiber side to the extent you want to go into those.

Okay, thank you very much. So, Sundar, it seems like from the outside looking in, the pace of product development and release seems to have accelerated, while at the same time it seems like you're gaining efficiencies with the assets you're deploying to run your business and as your CapEx growth is moderating. So I'm wondering what concrete steps you may have taken to increase productivity or focus at the company and what you're doing now to continue to drive those gains going forward. Then separately, as you called out earlier, you are shipping an unprecedented amount of hardware devices. So as you think about what your product portfolio might look like over the next five years or even the next decade, does the Pixel phone market change in direction for Google to become maybe more of an integrated software and hardware company? Thanks.

Thanks, Stephen. On the first thing, I would say we are very focused on our core mission. And we see a huge opportunity to do that in a unique way, thanks to what we view as a point of inflection with machine learning. So refocusing the company on a set of initiatives, recognizing that point of inflection is what has helped us really focus on things. And things like the Google Assistant are a manifestation of that, and you will see us continue to stay focused and innovate that way. In terms of hardware, I think in our vision, computing is becoming more and more integral. It's going to be there for users in many different contexts. And so to really think and evolve it, you need to think about software and hardware together. That's where a lot of innovations happen. And so for us to push the paradigm, push the boundary, we are very committed to doing that. So it's a thoughtful effort from us. But overall, as I said in my remarks, we deeply remain committed to building an open ecosystem because at the end of the day we want Google to be there for every user everywhere. And to do that well, we want to work with partners and build a great ecosystem to make it happen.

And just to add a bit more to Sundar's first answer or first response to your question about efficiency as it's expressed through CapEx spend, I think it will probably be helpful to add. The pace of spend reflects the ongoing success of the team driving meaningful efficiencies in planning and operations for our technical infrastructure. And that's enabled us to support growing demand but at a stable investment level. We've talked about that on prior calls as well, but I think we're proud of what they've been doing there. Some examples of efficiencies include improvements in server utilization and the use of machine learning that Sundar has talked about, and the deployment of innovations like our Tensor Processing Units that he has commented. So the main thing is we're building greater productivity with existing machines. And what's important to note is that's not only good for Google products generally, but it's also valuable to our cloud offering for our enterprise customers.

Thank you. Sundar, you mentioned a point of inflection within machine learning. Can you talk a bit more about the trade-offs in the productization and sale of that inflection through Google Cloud versus leveraging that innovation yourself through Google Assistant? And maybe for Ruth, just any thoughts on potential future disclosures around the cloud business, that would be great. Thank you.

On the first one, look, I think we are – it's a big platform shift, and it has to be available for everyone. So we've always, just like we have done with things like Android, when we see platform shifts, we provide Android to everyone. So that's the way we think about Google Cloud. We want to make sure that all these new capabilities for machine learning and AI are available through Google Cloud to all our partners. We don't see it as a zero-sum game. I think internally, core to our mission, we see areas where we can execute and we will continue to do that, but we want to do both and we can do it thoughtfully well.

On your second question, we constantly look to assess if or when additional data makes sense given specific performances. And when we went through the third quarter results, that has been our intent with all of the color commentary on cloud. I guess the only other thing to add is that the largest percentage growth year on year in our other revenue line, actually even across all of our revenue lines, was in our Google Cloud platform, and that reflects significant momentum in compute and storage.

Thank you. Sundar, I apologize if I missed it. But could you talk about machine learning in search and how much it's making a difference over the past few quarters and how much you still have to go, how important that is to revenues? And then, Ruth, a couple questions, any thoughts on whether hardware sales could make a difference to margins going forward, and also on the stock-based comp, how Google thinks about that expense internally? Thank you.

On machine learning, in areas like search and even ads, on search we've had an effort called RankBrain, which is bringing Google Brain in the context of search. And we've made great progress on it, but I would still characterize it as very early stages in terms of the long-term impact we can have. Generally and similarly, we are in very, very early stages of incorporating machine learning, the newer machine learning systems in ads. And again, that's the beginning of a long journey as well. Overall, I think all of these systems are incredibly complex systems, and they are handcrafted systems over many, many years. And so over time, I think machine learning will surface newer approaches and newer insights. And so we see it as a huge area of opportunity, but it will play out over a period of time.

And then in terms of the hardware family, I think Sundar has given a lot of color on that. It's still early days with the rollout of some of these newer lines that we're super-excited about. And so early to make a call on that, but certainly investing meaningfully in the line given the importance we see in this. And then in terms of stock-based compensation, we've always said we're going to remain focused on long-term revenue growth, and that does require investing in talent. And we do believe equity ownership is a really valuable part of our overall compensation, consistent with alignment of interests. It's something we keep an eye on and we're mindful of the full cost of equity-based compensation and certainly look at that as part of the overall costs that we're investing in, in and across the businesses.

Hey, guys. Good afternoon, thanks for taking the question. Sundar, last month Google was part of the founding group for the Partnership for AI. And I was just curious to hear your thoughts on what your goals are for the group and how you may take lessons back to your leadership at Google. Thanks.

Look, I think I'm very, very glad to see the group come together. We are in extraordinarily early days for AI and it's super-important, but we approach it thoughtfully as an industry. I am encouraged to see the commitment across these companies. Our goal is to promote open collaboration, help the public understand AI, and establish best practices in R&D. Without something like this, I think ideas would develop in silos, and so I think it's good to do this to promote an informed dialogue on AI. And so I'm pretty excited all along, just like with TensorFlow, where we are doing this in an open way and all the other things we have done at Google. I think it's important as we work on new technology to contribute and to give back. And so in that context, I think all of this is personally very meaningful to me.

Thanks to everyone for joining us today. We look forward to speaking with you again on our fourth quarter 2016 call. Thank you and have a good day.