推特 (TWTR) 2017 Q3 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Twitter Third Quarter 2017 Earnings Conference Call. (Operator Instructions) I would now like to turn the call over to your host, Krista Bessinger, Senior Director, Investor Relations. Please go ahead.

Hi, everyone, and thanks for joining our Q3 earnings conference call. We have with us today our CEO, Jack Dorsey; COO, Anthony Noto; and CFO, Ned Segal. We hope you've had a chance to read our shareholder letter published on our Investor Relations website this morning. Like last quarter, we'll begin with just a few prepared remarks before opening the call directly to your questions. During the Q&A, we'll take questions asked via Twitter, in addition to questions from conference call participants. Questions submitted via Twitter should be directed to @TwitterIR using the #TWTR.

We would also like to remind everyone that we'll be making forward-looking statements on this call, such as our outlook for Q4 and the full year of 2017, and our operational plans and strategies. Actual results could differ materially from those contemplated by our forward-looking statements and reported results should not be considered as an indication of future performance.

Please also take a look at our filings with the SEC for discussion of the factors that could cause our results to differ materially.

The forward-looking statements on this call are based on information available to us as of today's date, and we disclaim any obligation to update any forward-looking statements except as required by law.

Also, during this call, we will discuss certain non-GAAP financial measures. Reconciliations to the most directly comparable GAAP financial measures are included in our shareholder letter. These non-GAAP measures are not intended to be a substitute for our GAAP results.

And finally, this call in its entirety is being webcast from our Investor Relations website, and an audio replay of this call will also be available via Twitter and on our website in a few hours.

And with that, I would like to turn it over to Jack.

Good morning, everyone, and thank you for joining us today.

Before we take questions, I want to highlight a few things. We had a good quarter and made progress in 3 key areas of our business: We grew our audience and engagement, made progress on a return to revenue growth and achieved record profitability. I'll outline these in a bit more detail for you.

First, our audience and engagement are growing. Our monthly and daily active usage grew year-over-year as well as sequentially. Importantly, this growth was broad-based, both in the United States and internationally, with growth in the majority of our top 10 markets. And as we said last quarter, product improvements continue to drive audience and engagement growth. Our work to increase relevance and make Twitter easier is making an impact.

Second, this quarter, we made progress on a return to revenue growth. Growth was driven by improved execution from our sales team, strength in video and direct response ad formats and continued strength in our data business, where we saw our third consecutive quarter of accelerating growth.

And third, we achieved record profitability in Q3. We continued to improve our GAAP net margin and delivered our highest adjusted EBITDA margin to date. This profitability reflects our improved prioritization and disciplined execution across all of our strategic priorities.

In addition to the progress we made in audience and engagement growth, revenue growth and profitability, we're more focused than ever on making Twitter a safe place for everyone. We made this a priority in 2016, updating our policies and increasing the size of our teams. And in 2017, we rolled out a number of product updates with an increased sense of urgency. We know this isn't enough, and we're taking a more aggressive stance in our abuse rules and how we enforce them. We're addressing this from a policy, enforcement and product perspective. And just last week, published a calendar of the upcoming safety work we have planned through January 2018. This is the first time we've shared this level of visibility into our work, and we hope it builds trust along the way.

Finally, along with our COO, Anthony, we're also joined by our CFO, Ned, for his first earnings call with us, and we're excited to have him. I'll turn it over to him to say a few words.

Thanks, Jack, and good morning, everyone. I'm happy to be here to discuss the quarter with you.

As Jack said, this quarter, our audience and engagement grew, we're making progress on a return to revenue growth and we achieved record profitability. We've got a great opportunity ahead of us to put Twitter in as many people's hands as possible and to make the service more relevant to more people every day.

And with that, Jack, Anthony and I look forward to taking your questions.

(Operator Instructions)

Great. And our first question comes from Heath Terry at Goldman Sachs.

Curious if you could give us a little bit more sense about sort of your long-term margins goals. Given the upside margins we saw this quarter, the guidance for next quarter, is there a point where you feel comfortable with the level of profitability the business is seeing to the extent that you're going to want to reinvest more, whether it's in technology, marketing, people? And then also, just on revenue growth, is there also a point that we're getting to, with some of the restructuring that you've been doing in your advertising products, where we start to see advertising growth aligned more with growth in advertising impressions or at least growth in DAUs?

Thanks, Heath. I'll take the first question and Anthony will take the second one. On margins, there's no change to our long-term thinking, and we made great progress this quarter towards our goals. We are at the point, though, where after a lot of hard decisions made over the last couple of years and an expense base has come down quite a bit where the expense base will selectively grow as we invest against our priorities. So to the extent you see margin improvement, it's more likely to come from revenue growth than it is from cutting the expense base. As you said, we want to make sure that we're investing appropriately against the opportunities that we see. Let me turn it to Anthony to talk about the revenue side and ads.

Thanks, Ned, and good morning, Heath. So I want to address your question about revenue growth and will that -- will we start to see a trend towards our audience growth. And before doing so, I figured it will be good to give you just an update on how we feel about the progress that we're making in a return to revenue growth overall and what's driving it. First message I'd like to send is that we feel really good about the progress we've made. We're ahead of where I would have guessed back in January today. And the progress that we're making falls into 3 areas. First is value for our advertisers, value for our content owners and value for our enterprise partners, strong execution and diversification of revenue. I'm going to focus specifically on value. We have a much larger audience. It's growing. It's growing consistently. We have much lower prices, and we have better products that are performing better. All of that equals better ROI. Let me take each one of those one by one. First, audience growth. We had 6 quarters of solid growth in daily active users, with the last 4 quarters up double digits, and that's against an increasingly tough comparison. Having audience growth is really important for our partners, both on the advertising and on the content side, so they are confident investing in a platform that they know is going to continue to grow. And importantly, as Jack mentioned, that growth is driven by product changes, in addition to other factors. Second, our prices. Our prices are down meaningfully measured by cost per engagement. They're down more than 40% in a year-over-year basis in aggregate, driven by mix shift, but importantly, significant year-over-year declines on a like-for-like basis. Bigger audience, lower prices is better value. Third, we have better performing products. We went through the painstaking process of eliminating some of our products in our portfolio that we didn't think we could win with, that we didn't think were good for advertisers. Our team's also invested a lot in our platform. The combination of those 2 things allow our products to perform better for our partners with higher click-through rates year-over-year on a like-for-like basis. Bigger audience, lower prices, better performing ads, more value. And then last, we have a really differentiated portfolio of products. Right now, our team is focused on communicating to our advertisers the strength we have in our video products. We have 4 really differentiated products. One of which we launched in the fourth quarter, which is a video card view that was used by Apple in its most recent campaign. Those differentiated products, combined with a bigger audience, lower prices and better performing ads, has really helped. And so we feel good about the direction that we're going and the progress that we've made. And a year ago, we had to talk about the fact that our audience was growing, and it would take a while for revenue growth to start to mirror that, and we feel like we're making progress to that, and we're on the precipice of seeing real improvement.

And our next question comes from Doug Anmuth at JP Morgan.

I wanted to follow up there with 2. First, Anthony, can you talk -- and Ned, can you just talk about whether you think you can grow ad revenue in 4Q, excluding the headwinds from TellApart and the NFL? And if you can talk a little bit qualitatively about how you think that sets you up into '18. And then, Anthony, you talked about CPE. It's basically been down, I think, between 40% and 60% roughly in each of the last 9 quarters. And conversely, you've gotten a triple-digit engagement growth or impression growth throughout the same period. A lot happening there in terms of video, in DR, in other mix shifts, also the like-for-like you mentioned, but when does that start to kind of stabilize more going forward?

Doug, it's Ned. I'll start, then Anthony will chime in. I'll talk a little bit about the fourth quarter first. So I'll just remind you and everyone that we don't give revenue guidance. We give EBITDA guidance and guidance for EBITDA margins, but we do recognize that people will do math to back end other parts of the P&L. Our EBITDA guidance is meant to reflect what we believe are the most likely ranges of outcomes. And if you were to do the math and compare it to prior periods, you'd recognize that when you -- let me just go back to a point that Anthony made in July, where he mentioned $75 million of headwinds in the fourth -- in the second half of the year. I'll break it out for you. It was $20 million of TellApart in Q3, $40 million of TellApart in Q4 and $15 million of tough comps in live in Q4. If you remove TellApart and you do some math around the last couple of quarters, just to ground you a little more, Q3 saw a 1% decline in revenue and Q2 saw a 2% decline in revenue. The ranges of outcomes that you could back into for the fourth quarter would include growth ex TellApart. Having said that, it's too early for us to call where we'll end up with much of the quarter to play out. In November and December, we'll get most of our revenue for the fourth quarter. And as you can imagine, there are lots of puts and takes from one year to the next. Along with the TellApart, there's also the presidential election and live Thursday Night Football, which don't repeat themselves this quarter and then there are things that will happen this quarter that didn't happen in the fourth quarter of last year. So too early to predict when we'll return to revenue growth, but you better believe we're doing everything -- that we're working hard to make it happen.

So Doug, to your question about how we're setting up for 2018 and the fourth quarter, then I'll talk about CPE question. Our goal is to overcome every headwind that we have and we're working every day to do that. We're going to the fourth quarter with -- very well positioned. It's really just about execution. We have been able to really increase the portfolio of live product that we have in content. We have had over 850 events in this most recent quarter, which is a significant increase versus the prior quarters, and the selection is now broad-based. It's international, it's not just domestic. We hope to leverage that in a number of product initiatives as it relates to in-stream video and pre-roll video as we go into the fourth quarter, in addition to just the continued progress that we're making on product improvement. So we're going to do everything we can to overcome those headwinds, and we're focused on getting back to growth. What will it take to get back to growth as we go into 2018? I think it's all about execution and continued diversification of our revenue. We like to focus on people, processes and technology. I went through the value that we're starting to really communicate well to our partners. We have the right products. We have the right value. It's about making sure each one of our partners understands that, and we are making great progress in that regard, and I'll give you some numbers to support it. But before doing so, it's people on the ground, calling on the right clients, calling on them frequently enough, having the right conversations and showing them that our product works, and that's what our team is doing. In fact, our top 100 advertisers globally in the quarter were up 23% year-over-year. So as you peel the onion a few layers, you start to see growth by the next layer or two down that will start to surface itself. So top 100 advertisers globally up 23%. Top 100 advertisers in the U.S., up 7%. As it relates to communicating the value and proving it, this quarter, we conducted 43% more custom-advertising studies in Q3 2017 than we had in the same time period in 2016. And it was a combination of both third-party measurement studies as well as our team's investment in self-serve and we'll do more of that. The second factor, which we haven't talked about, is diversification of our revenue. Our Data and Enterprise Solutions business accelerated for the third consecutive quarter on a year-over-year basis. We still are very optimistic about the growth opportunities we have in that product category. And then our self-serve advertising business is a relatively large business. It's a high-margin business because there's not a lot of people given that it's self-serve through our online platform. That business returned to growth this quarter as well. So making really strong progress in a number of areas and it really comes down to execution given we have the right value equation for advertisers. As it relates to 2018 and as it relates to CPE, what I'd say is this. We think about CPE as a combination. We think about CPE in the context of yield, so effective CPM. And effective CPM is nothing more than the product of CTR, click-through rate, x CPE, cost per engagement. Our effective CPM has been relatively stable the last 3 quarters, and we think it will remain relatively stable, all else equal, in the competitive environment and that's how we think about CPE in the context of yield to the platform, and we're glad that it has reached the stability point because now it will allow us to just drive incremental demand against the existing (inaudible) and that's going to back to revenue growth.

Great. And our next question comes from Colin Sebastian at R.W. Baird.

I have a couple of questions. First, I just wanted to mention on a personal note that during the wildfires here, Twitter was a very useful source of communication and real-time information for my family, so I appreciated that. In terms of the questions, first off, wondering what the 1 or 2 most important factors were in driving the increase in DAU growth and engagement? You said a couple of factors in the release. And then, more specifically, on the Explore tab, I'm wondering how usage and time spent compares with the core time line, in particular, among new or more casual users?

Thanks for the question, Colin. Our audience growth in Q3 benefited from all 3 levers that we talked about: organic growth, product and marketing. They're all meaningful and their contributions will vary in any given period. Our product changes, though, should continue to drive improvement in audience and engagement, and they should compound over time. Let me turn it to Jack to talk about the specific product changes that we're making that are driving growth.

And thank you for the mention of the -- of using Twitter during the wildfires. We saw the same sort of utility. I'll touch a little bit on, more broadly, what we're interested in and it relates to the Explore tab. Explore is a way for us to organize around topics and interest. You'll find search there. You'll find moments. You'll find topic modules, which we just launched, which is what we infer you might be interested in, and some of the best tweets and accounts that you see there, and of course, trends. We're playing a lot with better matching people with their interest and with topics that they care about within this. This is an area of experimentation. We're experimenting very, very quickly to figure out the right approach. And then as we learn more, bring that more into the time line, and you'll see some of those experiments going forward. One of the things we're excited about is, we just launched a WhatsHappening now module, which is focused today on sports. So if you're interested in baseball, for instance, you would have seen a callout at the top of your time line to see the discussion and the conversation around the game last night. And as this gets better and better and we get more personalized, we'll continue to see results not only in daily active, but across the board. We do believe that as we personalize Twitter more and as we continue to add more selection and continue to show off the conversation, it has a direct impact on our results, so that's where we're focused. All this comes back to the job we're serving and making sure that we are showing people what's going on, what's happening right now and keeping them informed, both when they're in the app, through the time line and Explore, and also when they're out, through notifications.

Great. And our next question comes from Brian Wieser at Pivotal Research.

Going into the data licensing and other segment, I was wondering if you could talk about the growth trend between MoPub versus your data licensing. And then, separately, you mentioned the new customer success team providing support to top 5% of advertisers in the letter, driving 40% revenue lift. I'm wondering if you could talk about how broad you think you can possibly extend that sort of extended customer support and just sort of roll out plans for that and if you think that's going to continue to add some incremental growth.

So I'll start on data licensing, then you'll hear from Anthony. On the data licensing, we don't break out data licensing relative to MoPub, but it's important to point out, around data licensing, that we saw our third straight quarter of accelerating growth. We're really pleased with the progress there, and we feel like there's a lot of room in front of us, both as we continue to expand the product that we offer, that we bring it to more customers, and we expand our relationship with our existing customers, that continues to be an area of focus for us, one that adds visibility and growth to our opportunity. Anthony, do you want to talk about the top 5 advertisers?

Yes. And Brian, could you restate your question? I'm not sure I understood it clearly, your second question, as it relates to top 5 advertisers.

Sure. There's a note in the letter that says that you've launched a new customer success team that provided support to your top 5% of advertisers. You mentioned already that 500 advertisers increased their spending level that was certainly well in excess of the rest of the company. But to the extent of providing and customer success team contributed to extra growth, I'm just wondering if you plan to broaden this customer success team or if this is sort of -- just further thoughts around that, that would be useful to hear.

Sure. And just to clarify your point. The comment in our letter about the top 5 advertisers and customer success is for our self-service advertising business, and none of our self-serving advertising partners fall into our top 100, so it's a different group and a different channel. So they're both showing positive trends, but they're not -- they're different. Let me talk about the self-serve business for a second. The business did grow in the third quarter. It has benefited from a really strong leadership from our team, both on the business side, and importantly, on the product side. And the combination of those 2 working together, I think, bodes really well for incremental growth. As you know, we do not have millions of advertisers like our competitive peers. And a big reason why we do not is we haven't had a "turn it on, forget it" type of advertising product and "who is on" advertising product. We think the self-serve channel has the potential to be that, and so we're investing in that area to see if we can find a great product market fit. So in addition to the customer success team, we've also made investments on onboarding, and we've made investments in a new subscription product, that for $99, you don't have to pick what ad format you want to use on Twitter, we do all that for you. We get you account follows, our technology machine learning gets you impressions and you get a value back that we've delivered 100% for you, so it what really is a subscription product like you'd use in other formats. It's early. It's being tested, but the initial signs are promising, and that's as a result of the innovation of our product team and our leadership team. The 5% of our top advertising customer success, I would say, we'll continue to invest in that area, but incrementally and very cautiously. We don't want to add costs to that business relative to its profitability and the contribution it makes relative to the overall business unless it's warranted. But it is one of the things that we've done to drive an improvement. And then just the last thing, I mentioned onboarding. We shift to a new version of our campaign setup experience, which made it much easier for new advertisers to get started and for existing advertisers to launch better campaigns. That drove a higher conversion rate for new advertisers and higher average revenue per advertiser from both new and returning markets. So more product innovation driving more adoption and with a better product market fit.

And our next question comes from Twitter. It will come from the Twitter account of TMFJMo and he asks, have you received any feedback from video partners that discovery of live video is not optimal? And do you have any thoughts on how to improve it?

Our live business has just anniversaried its first year. It's a very different strategy than our competitors, and I think it's important to understand that strategy and the value that we provide for partners because it plays into your question. So to be very clear, our live strategy is about leveraging a strength that existed on Twitter for the last decade and that's the strength that we've had passionate audiences about specific topics, already talking about things that are happening in the world, even though the video was not on the platform, it was someplace else. We've been leveraging that strength to pick the right content, to meet the passion of each of those audiences and we combine their conversation that already exists on Twitter all over everyone's time line into one curated time line with the video there. The idea is to make our content being delivered faster. It's to drive more selection. It's to drive more discussion and more personalization. And in doing so, we think we can help content owners and advertisers extend the reach that they have in television to people that either aren't at home to watch or people that do not pay for television, and we've been able to prove to our content partners that we are extending their reach. It's an incremental audience they wouldn't otherwise get. It's an incremental reach for advertisers. So it's not a replacement strategy at all. It's not a strategy that's trying to disintermediate the content owners from the advertisers or from the audience, and that has worked really well for us, and that's why we've had such strong increase in selection on a year-over-year basis. As it relates specifically to your question about discovery, the real question is, are we delivering the audience scale that they want, and are we delivering the value that they want in advertising? And I think the best way to answer that question is just the significant growth that we've had across sports, news and entertainment. We had 830 events in the quarter. That was up versus 625 events in Q2. We saw growth in unique users of the products. And we saw growth internationally and domestically in terms of selection. And so really proud of what the team has done there. But individually, we deliver value for each content owner and that's how we measure it.

So our next question comes from Rich Greenfield at BTIG.

So 2. First, just on the DAU front. When you gave your last update, I think at the end of -- sorry, in Q2, you had mentioned that DAUs were tracking up 12% through the month of July, and you reported 14%. So that would indicate that you kind of exited the quarter with north of 15% growth in the back 2 months. Just wondering, is that correct that you did see a notable pickup in the back end or the second -- or the back 2/3 of the quarter? If there's anything that specifically drove that? And anything you can tell us about those trends as you go up against even tougher comps in Q4? And then just I want to come back to this video question because I'm less interested in live video discovery, I know you're working on that, but just video discovery in general, like, whether it's a live video or a piece of content that aired a few weeks ago, finding video, there's a lot of great video that, whether it was live or whether it happened this morning, you still care about, how are you going to surface that video? I mean, is there going to be a dedicated video experience? Or just -- there's so much content that has value beyond live that I'm curious how you're thinking about surfacing that.

Rich, I'll start on DAU, and then, Jack will talk about discovery broadly. So the comment on the Q2 earnings call was really meant to be directional. I wouldn't want you to take away that we were exactly at 12% or to try to break the months apart. Although we generally expect organic growth, product and marketing to continue to drive DAU higher in subsequent periods, I don't think it makes sense to break them out as they're all meaningful and their contributions were there in any given period, and they all contributed this past quarter. Go ahead, Jack.

In terms of video discovery, the way we're thinking about this is just adding more personalization across the board to Twitter. So our biggest efforts are really applying machine learning and deep learning to every single tweet. Some tweets carry texts, some tweets carry images, some tweets carry video. And we want to make sure that if we infer or you explicitly tell us you're interested in something, that we're delivering the right media format for you at the right time. Sometimes that is text, sometimes that is an image, sometimes that is video. And then when you go even deeper and really understand what (inaudible) a video as well, so we are thinking about this more from a broad-based perspective and being somewhat format agnostic so that we can make sure that we're delivering what is something that matters to people, something that is going to really feed what their interest is in particular. So looking at this, making this a lot easier, we're exploring a lot and experimenting a lot in the Explore tab. Also, we want to make sure that this blends into your time line so that people don't have to do any work to find what matters to them.

Meaning that the ultimate goal, Jack, is that you won't even need to use the Explore tab because you'll be good enough from an AI standpoint to show me what I want to see in the main tab?

Correct. Correct. Our goal is personalizing and getting this right so that we're delivering something that really matters to you and whatever format it comes through.

And where would you say the AI team is in terms of size, growth and kind of against that effort?

We've had a lot of strength in our AI discipline, machine learning and deep learning in particular. For the past 1.5 years, we've really been applying it to the core aspects of the product, so we feel really good about the size of the team and the depth of the team. But we're hiring, and we continue to look for leadership in all these areas, but I feel really confident in our application and our expertise.

Great. And our next question comes from Lloyd Walmsley at Deutsche Bank.

I guess one for Anthony. You talked about growing audience, better product and lower pricing makes the product work better for advertisers. So wondering, as you see advertisers who are coming back to Twitter perhaps after a period away, how are you seeing spend levels trend from these advertisers coming back to better product and lower prices?

Yes. What I'd say is it's different by markets, so I don't want to give you a one answer because the averages kind of mask what's happening in each individual market. I will pick the Japanese market as an example. It's our second strongest market. We do disclose the revenue there, it's greater than 10%. Our Japanese revenue accelerated this quarter on a year-over-year basis. It accelerated because our team is focused on a by product approach and by advertiser approach. And as we see the growth in spend for advertisers slowing and the growth in number of advertisers slowing, we look at what opportunities we have to sell more of a broader array of our products to meet the needs of those advertisers. And so one of the key things that we did in Japan in the most recent quarter is we increased the strength of other products that are really complementary. And in fact, one of those is the video product. The Japanese market has historically been a DR market, which is in contrast to the rest of our business, which is largely a brand business. And now we're starting to see the uptick of our video products in the Japanese market by branded advertisers because the product is strong, the prices are really competitive, and our targeting is better, and we have a really big audience there. And so each market is going to be different, but I'll just tell you in the Japanese market, it was about broadening the products that the advertisers buy, and those products are broadening because they are valuable and they're valuable across the 3 dimensions that I mentioned. So each market we're focused on I'd like to say SKU level, the ad products that are available to purchase, what are they buying? Are they the right products for that advertiser? Are they the right price? And can we deliver the right ROI? And then that's the approach we're taking.

Great. And our next question comes from John Blackledge at Cowen.

Just a couple of questions. I think the direct response format was called out as a driver of topline. Could you discuss perhaps in more detail, the drivers of improvement in the DR channel? And did DR advertising grow on a year-over-year basis? And then maybe if you could just explain the change in the MAU reporting and why the change now?

Sure. On the direct response question, we did have a strong quarter in direct response and mobile application download. We also had a strength in video. In [MAP], in particular, we saw a strength this quarter from an increased launch of new games and apps. This is one area of strength that we've had in mobile application download products. Historically, we haven't had as much success scaling the mobile application download revenue stream as we'd like and so we're allocating our resources accordingly, but the MAP business did have a good quarter on the back of several large partners with new entrants into the market in the app space, and we can uniquely help them drive penetration given the significant scale that we have, the fact that we're mobile, and the fact that we're global. And that was the reference in the commentary of the letter.

And John, I'll go into MAU for you. During the quarter, we realized that we had included 1 million to 2 million users over the past year of a service for third-party apps in our MAU measurement that should not have been included. It's important to note that since it's a small absolute number on a base of over 300 million MAU, that historical growth rates and absolute numbers really don't change very much, so -- and on an adjusted or unadjusted basis, we grew 4 million MAU this quarter, from 326 million to 330 million, with strength in the U.S. and internationally contributing to that 4 million. You will remember the DAU wouldn't be affected by this, as DAU does not include users of any third-party services in the calculation. More on MAU, I'll just point out that although we don't give guidance for MAU or DAU, I just want to note that Q4 is typically seasonally weak for MAU. And in addition, there is a change to the Safari interface that will affect some third-party apps, including Twitter. It impacts our third-party auto-polling MAU who aren't serve ads, so don't have revenue tied to them. DAU, again, wouldn't be affected by a change like this.

Great. And our next question comes from Ross Sandler at Barclays.

Yes, 2 questions. So one, on the programmatic, you mentioned in the letter that you're in outflow with the new programmatic ad channel. So this is something, I think, you worked on in the past with different partners, et cetera. Can you just talk about this new alpha, what the strategy is, and what the success that you're seeing early on? Any color there? And then the second question, on data licensing. Can you just talk about, within the Gnip business, the difference between price increases versus new account growth in terms of driving the overall acceleration there? And do you expect this growth to kind of consistently stay at this level into '18 and '19? Any color there on data licensing would be helpful.

Sure. So let me take your programmatic real-time bidding question. I want to emphasize it's very, very early days for us. It's a test. We've talked about previously from a priority standpoint, as we think about revenue opportunities and prioritization. We've talked about tapping into new channels of demand. One new channel of demand is video and tapping into those budgets outside of digital and social. The second opportunity for incremental demand is third-party demand. We think there's a real opportunity to tap into premium display budgets. We do not really tap into those budgets today. We think our TV product could be very competitive against premium display budgets, and that's the opportunity that we see in front of us from a demand perspective. On a supply perspective, we have some inventory that's not being sold. We're in a demand-constrained environment, not supply-constrained environment. So we think we can tap into this third-party demand through programmatic RTB, sell more of our existing inventory and owned and operated, potentially, I want to emphasize the word in quotes potentially, tap into monetizing our logged out inventory and syndicated inventory. And so that's the rationale behind the initiative. It's very early. We're testing the pipes and technology. Right now, we really haven't even rolled out a real alpha. We're negotiating a number of partnerships with DSPs, and I'm only giving you that detail so that you don't expect any real impact from this for the foreseeable future because that's how early it is. But we do think it's a real opportunity and one that's worth investing in. As it relates to your second question, our Data and Enterprise Solutions business was one that was acquired over 4 years ago. It had a specific focus on one use case. We think there's an opportunity to expand outside of that use case. And sell data not just for the ability to understand sentiment analysis and to have brand watch types of use cases, but to actually use the data to do other things as it relates to delivering the right content to the right user at the right time, as it relates to buying media, as it relates to direct marketing. We haven't started to enter into that type of opportunity yet, but that is the biggest opportunity to drive incremental growth of data partners. The bulk of our growth that you're currently seeing is from our existing customers. And in fact, we're going to do business with fewer customers instead of more. And the reason why is because we are focused on a channel strategy and a partnership strategy that results in higher prices for the value that we're creating and are more products focused to land and expand. And so it's not about driving penetration of more data partners in the current execution today, it's about driving fewer partners and more spend per partners, so they invest more in the opportunities in front of them. We still think there's a significant opportunity in that strategy, but there's also an opportunity to expand our data partners through broader usage, and we're just at the very early stages of that and that hasn't shown up in the results yet.

Okay. And our next question comes from Justin Post at Bank of America.

A couple of questions. First, obviously, the DAU ratio to MAU is growing. Can you talk about engagement with new content? Are you seeing time spent increase at a similar ratio per user, and how you're thinking about that? And then just thinking about growing advertising index here, I think you already discussed ad loads. You have more room there. But how do you feel about where you are on ad loads? And when do you think like-for-like pricing could maybe bottom out?

So first, on DAU to MAU, I think you're referring back to a comment that we made after the earnings call in July that DAU was less than 50% of MAU. That continues to be the case. And growing DAU inside our existing base of MAU is definitely one of the ways that we can grow DAU, but there's another way as well, which is to grow MAU and turn them into DAU. There is one way that I like to think about that, that really helps it come alive, which is that every day 2 million new users come to Twitter -- I'm sorry, 2 million or more new users come to Twitter. Of them, 2/3 are users who have not been to Twitter for the last 30 days, and so there's an opportunity to effectively resurrect them and bring them back to Twitter. And then the other 1/3 are users who have never been to Twitter before. Both are opportunities for us to create MAU and ultimately to create DAU, and that's one example that I like to point back to that helps bring the opportunity alive. On time spent, the first way that we like to look at engagement is to look at DAU. If we were to look at one metric, that would be it. As you can tell, we're pleased by the progress that we're making. There definitely are other ways to look at it, and we're pleased with the progress that we're making there too, but focusing on any one of them would probably give it too much importance relative to the others. Anthony, do you want to talk about ad load?

Sure. To put the answer upfront, we do have more room in ad load, I would say, globally and then also in the United States. But I'd also point out that we're really also focused on diversification, earlier I said our return to growth was dependent on value, execution and diversification. And diversification is not just into data and our self-serve advertising business. It's also into areas that do not rely on ad load. So if you think about our video initiatives, many of those are driving consumption that are not necessarily dependent on ad load given the way that product works. But we have been in a demand-constrained environment, not a supply-constrained environment. So even without that diversification, we have an opportunity to sell a lot more supply, which is one of the reasons why we're executing against the opportunity of programmatic real-time bidding. In terms of pricing, I had said earlier in the call, we don't look -- we don't like to look at pricing by itself as a driver of stability. We like to look at yield to the platform measured by effective CPM, and effective CPM for the last 3 quarters has been relatively stable, and we think the combination of CPE going down and click-through rate going up will allow it to continue to be stable. I will tell you that's a really powerful equation for advertisers to be able to pay lower prices on a cost per engagement basis on a like-for-like basis and get better relevancy measured by click-through rate is really the combination we want to see.

Great. And our next question comes from Twitter. It comes from the account of Francis (inaudible) who asks besides live video and data licensing, can you talk more about other innovations in the pipeline that are designed to boost revenue?

Sure. There's a number of opportunities that we are looking at that involve consumption on the platform that are currently not monetized. And I just want to emphasize looking at, these are not around the corner, but they're being tested and considered. A couple of quarters ago, it was in the media that we were testing the concept of a subscription product for TweetDeck. TweetDeck is a really valuable asset for us. It's really valuable for a distinct type of user. And the question that we're asking ourselves, is there a job to be done here that we can fulfill and create a business around it after fulfilling it? Because we have to do that first. We have to deliver on what they want to hire us for first. And then once we do that, we can create a business around it. And so TweetDeck is clearly being used and hired by communications experts, by the media, and increasingly, by marketing experts. And as Twitter has had a greater impact in the world, its relevancy and its use case can expand to beyond just that -- the users that I mentioned. And in doing so, if we add more value, is that a way to charge from a subscription standpoint? It's very early. We've only done concept tests. We haven't really moved beyond that phase. I just give you that one example. The rest of the stuff, I don't want to really highlight because it's really in ideation stage, and we don't want to get ahead of ourselves. But I know that was publicly talked about, so I feel comfortable telling you where we are there.

Great. So I think we have time for just one last question, and that question comes from Youssef Squali at SunTrust.

A couple of questions. On the OpEx side, both sales and marketing and R&D were down sequentially. Can you speak to, basically, the level of spend that you need to drive the growth in the business, particularly for 2018? Is this level of spend that we're seeing or we just saw in Q3 enough to put you to where you need to get on the top line? And then on the -- and then maybe a quick question for Jack. Any early read into the longer character limits on users' engagement that you've been playing with over the last few months?

So, Youssef, there's been a couple things on the spend side. So we after -- and I alluded to this earlier in the call, after a lot of hard decisions over the last couple of years and the expense base having come down a bunch, we're at the point where we will be selectively reinvesting, and you'll see expenses selectively go up from here. Expenses ought to actually go up in the fourth quarter, and you probably noticed that headcount, or you will notice that headcount went up in Q3. The timing of the hires is such that we didn't get a lot of expense from the new hires, but they obviously impact the P&L in Q4. We had better attrition in Q3 as well and are seeing good trends both in sales and marketing and R&D and improved attrition and hiring some great people onto the team. So, in absolute dollars, those are going to move around one period to another, especially sales and marketing because there is some spend that can get pushed a few days and can move from one quarter to the next, but directionally, those are both areas where we want to continue to invest because of the opportunity that we see in front of us, and there are 2 other drivers that will cause OpEx to increase on an absolute basis in Q4, as we selectively invest in the business.

And we're always looking for opportunities to make Twitter easier for everyone. And as we shared last month, we saw a bunch of patterns where people were abandoning tweets in certain languages, German, for instance, because they could not fit their thoughts in the 140-character constraint. So we want to make sure that we were rigorous about how we thought about expanding this and answering the question once and for all. One of the things we promised to do in that blockbuster month, a month ago, as we rolled out into our testing is to be very open and share our data and share what we find so that everyone understands what we're doing and why we're doing it. It's too early to tell right now. We're still watching and learning and observing how this impacts the service overall. We want to make sure that we are maintaining our sense of brevity, make sure that we are maintaining our sense of real time and showing what's happening in the world. But we want to make sure that we're also enabling more voices to speak and not be frustrated by the constraints, so they can actually get their thoughts out in real time and really share with the world what their thoughts and what's going on. So we look forward to sharing all of our data and what we found in the next few weeks.

Ladies and gentlemen, that concludes today's question-and-answer session. I would now like to turn the call back to management for closing remarks.

Great. Thank you. We made good progress this quarter in growing our audience and engagement, revenue growth and achieving record profitability. It's our job to make Twitter better for the people who use it, and we're focused on making our service faster, more relevant and safer for everyone. Thank you all for your time and your support, and we'll see you on Twitter.

Ladies and gentlemen, thanks for participating in today's program. This concludes the program. You may all disconnect. Have a good day, everyone.