推特 (TWTR) 2014 Q4 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Twitter fourth-quarter earnings conference call.

(Operator Instructions)

As a reminder, this conference call is being recorded. I would now like to turn the call over to Krista Bessinger, Senior Director, Investor Relations.

Thanks, Nicolas, and good afternoon. Welcome to our Q4 earnings call, and thanks for joining us. We have with us today our CEO, Dick Costolo; and CFO, Anthony Noto.

We'll begin with approximately 20 minutes of prepared remarks, followed by Q&A. During the Q&A, we'll take questions submitted via Twitter, in addition to questions from conference call participants. Questions submitted via Twitter should be directed to @TwitterIR using the #TWTRearnings.

We'd like to remind everyone that we'll be making forward-looking statements on this call, such as our outlook for Q1 in 2015, 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 take a look at our filings with the SEC for a discussion of the factors that could cause our results to differ materially. These 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.

During this call, we will discuss certain non-GAAP financial measures. Reconciliations to the most directly comparable GAAP financial measures are provided in our earnings release. These non-GAAP measures are not intended to be a substitute for our GAAP results. 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 the call over to our CEO, Dick Costolo.

Hi, everyone. Thanks for joining us. I'll be giving you an overview of our financial results, and then I'd like to spend some time following up on the product initiatives we discussed at our Analyst Day in November regarding all the work we're doing to grow our audience and our business.

Before we jump into the financial results, I want to point out that our pace of execution across the Company, which I noted as an area of focus in our Q3 call, is already improved. And I'm delighted about that and about the results I'm seeing from the team. You already see evidence of this improvement in our recent cadence of product launches that are making Twitter better for current users, people who are new to Twitter, developers and marketers.

Financially, we had another great quarter, with strong revenue growth and very strong profit. Total revenue was $479 million, up 97% year over year. Adjusted EBITDA was $141 million, more than doubling since last quarter, when it was $68 million, and up 216% year over year. Our adjusted EBITDA reached 30% margins, also up significantly from 18% margins in Q4 of 2013.

That quarter closed out a very strong year of financial results for us. In 2014, we had $1.4 billion in revenue, up 111% year over year. Our revenue growth accelerated in 2014 compared to 2013, which was itself a year of significant growth for us. So I want to publicly congratulate everybody involved in building the business engine at Twitter. It's a remarkable collection of people really doing tremendous work.

Okay, moving on to operating metrics. We ended the quarter with 288 million monthly active users. We added 4 million users this quarter, and 47 million across 2014. There are quarter-specific factors that impacted our net adds in Q4, which include seasonality and a couple of issues related to the launch of iOS 8. We will discuss that in more detail later in this call.

Importantly, I want to highlight that the user numbers we saw in January of this year indicate that our MAU trend has already turned around, and our Q1 trend is likely to be back in the range of absolute net adds that we saw during the first three quarters of 2014. We have a number of projects underway to grow our user base and provide a compelling valuable experience to anyone in the world, whether they have a Twitter account or not. I'd like to give you an update on a number of those improvements we've made that serve the three objectives I discussed with you at Analyst Day, and also on our Q3 call.

Those reject those are one, strengthen the core. Two, remove barriers to consumption. And three, build new applications and services. When I talk about strengthening the core, what I mean is making Twitter more engaging, valuable and easy for logged in users.

On our Q3 earnings call, and again at our Analyst Day, we outlined three specific things we would do in this area. One, introduce better media creation and consumption. Two, enhance private messaging between users and groups. And three, improve the new user experience and make it easier for people to immediately get value out of Twitter the moment they sign up. We have launched or have begun testing products in each of these areas.

Let me talk first about native mobile video. We're clearly at the beginning of mobile video sharing, and we'll see with video what we've seen with photos -- an abundance of creation and consumption happening from the device we have with us all the time. We started our foray into video with Vine, which continues to do quite well and is seeing significant growth and consumption. They're now seeing more than 1.5 billion Vine loops a day.

And we've also been bringing video to Twitter through publishers and advertisers over the last several months via both our Amplify program and are roll-out of video. Twitter.com tools to professional publishers. The data all tell us that people love watching video on Twitter. And marketers are seeing great engagement, along with tons of earned media.

Just last week we brought native mobile phone video to all Twitter users. Now anyone can capture, edit and share videos right within the Twitter app. In expanding video to more people, we want to develop features that make video fast and easy. The simple video editor now in the Tweet Composer makes it super easy to capture moments as they happen, with video, on Twitter. We have a lot more coming on the mobile video front, and I'm personally investing a good deal of time in this area.

Next I want to talk about privacy messaging, which we have long made available through direct messages on Twitter, or DMs for short. You've heard me talk before about taking public conversations private, and we've launched a number of new features and improvements to make that even easier.

Just last week we introduced Group DMs. Now people can have private conversations with a group of people on Twitter, in addition to having those one-on-one conversations. When you couple this release with the ability to share tweets via DM, which we shipped in Q4, you can see some of the work we're doing to more easily move between public and private conversations on the platform. Importantly, direct messages give people a way to reach and talk to people they're connected to -- in some cases, only connected to on Twitter.

Finally, for the new users who sign up for Twitter every day, we're experimenting with Instant Timelines as a new way to help people get started more quickly. We discussed the Instant Timeline capability at our Analyst Day, and I'm excited that we've already launched Instant Timeline experiments in our mobile app.

The idea behind this capability is to remove the friction of finding a large set of accounts to follow when you first sign up. So with Instant Timeline, a new user gets a rich timeline the moment they sign up. And by immediately dropping users into a high-quality timeline of great content, we believe we will be better able to convert those new sign-ups into healthy long-term users.

The second objective I discussed with you is to remove barriers to consumption. From our most active users, to someone visiting Twitter for the first time, to making it really simple for publishers who want to embed Twitter content in their websites or apps, we want people to immediately get value out of Twitter and always see content that is meaningful and relevant to them.

There are more than 6,000 tweets a minute every day. What one person sees on Twitter is just a very narrow slice of all the amazing content we have on the platform. And we have an obligation to better organize content and deliver it in compelling ways to our total audience.

As we've discussed with you before, every month, more than 0.5 billion unique visitors come to Twitter but don't log in. I want to now give you three examples of what I mean by removing barriers to consumption, two of which are already out to users, and another that's coming in the next couple of weeks.

First, we've been working hard in a compelling product for logged-out users, and just this week, we launched an initial experiment of a logged-out home page. We're going to learn a lot from this initial experiment, and I expect us to iterate quickly, based on those lessons learns. The content people see will be based on algorithms, as well as curation.

Secondly, it has to be just as easy for logged-in users who follow a large number of accounts to quickly catch up. A couple of weeks ago, we introduced a new feature called -- While You Were Away -- the name is fairly self-explanatory. It shows you a few tweets you might be interested in, that you might have missed if you haven't been logged into Twitter for a few hours. We've been testing this for some time internally, and I think it's an elegant way of delivering compelling and engaging tweets to users the moment they come back to the platform.

Finally, in a couple of weeks, for the Cricket World Cup, we are providing global cricket fans with an immersive experience, similar to what we provided for the FIFA World Cup last summer. The difference will be that this time, anyone with a feature phone or smartphone will just as easily be able to see and enjoy these experiences, even people without Twitter accounts. You may recall we recently announced our acquisition of ZipDial. We'll be working with them to bring key moments in commentary around the Cricket World Cup to a much larger audience on Twitter.

Our third objective is to build an ecosystem of new applications and services. We closed out 2014 with the introduction of Fabric, our new mobile software development platform that makes it easy for developers to build great apps. I believe Fabric can become the infrastructure software layer of the mobile application ecosystem.

This is a significant opportunity, because Fabric helps us build the largest audience in the world, and creates new monetization opportunities for us. Fabric makes it easy for applications to integrate tweets and Twitter Timelines, with just one line of code. And as more and more mobile apps integrate Twitter content, the Twitter audience grows.

Furthermore, with Fabric, we have the technology in place that makes it easy for Twitter to be the monetization engine for mobile apps. So as I've highlighted on this call, we've launched or have public experiments out for nearly all of the new features we talked about at our Analyst Day in November.

Across our three objectives -- strengthen the core, remove barriers to consumption, and build an ecosystem of new apps and services -- we'll leverage the power of our monetization engine. And I've got some exciting business updates to discuss with you today on that front, as well. Because our primary ad unit is a tweet, the ads can go wherever tweets go, and our monetization engine follows our content on sites across the web or on other mobile apps.

We recently announced that we started syndicating ads, starting with Flipboard, which launched Promoted Tweets in its Twitter section on February 1. And just last week, we signed an agreement with Yahoo! Japan to monetize our syndicated Twitter Timeline on their web properties. By syndicating ads the same way we syndicate tweets, the audience that our marketers can reach obviously extends far beyond our owned and operated mobile apps and websites.

These ads leverage the same rich data and targeting that marketers have access to through Twitter today, and ultimately, the same measurement. Only now, we are bringing this advertising technology to the rest of the mobile ecosystem. That we are beginning to realize the potential for ad product by showing relevant, timely ads wherever there is Twitter content, is a huge accomplishment.

So to wrap up, our product updates introduced improvements to Twitter that make Twitter better for existing users. They make it easier for new users to get started and get immediate value out of Twitter. They provide important tools to developers and make it easy for them to build great apps. And they give marketers more channels through which they can take their messages to consumers. I hope to see the impact of these releases, from both a growth and revenue perspective, over the coming quarters.

To close out, let me just take a minute to acknowledge that 2014 brought with it a lot of change for the Company, particularly at the executive level. Those changes have brought us to where we are today. We have a great team. Our pace and quality of execution is the best it's been.

I'm proud of the way the Company has embraced and been fueled by these changes. And I couldn't be more excited about where we're going, and the profound impact Twitter will have along the way. With that, I'll turn it over to Anthony to go deeper into the financials.

Thank you, Dick, and good afternoon, everyone. I will discuss our financial and operating performance for Q4, and provide guidance for Q1 and FY15. In terms of revenue, Q4 was another very strong quarter for Twitter. Total revenue reached $479 million, an increase of 97% year over year.

Ad revenue reached $432 million, also up 97% year over year. Once again, ad revenue was driven by strong growth in both the number of advertisers and average revenue-per-advertiser in each channel and geography. Our direct sales channel remains the largest contributor, accounting for nearly half of the year-over-year growth, as we saw particular strength in the US, Japan and Canada.

Our expansion efforts in the mid-market and small and medium business channels continues to pay off, as growth rates in these channels once again significantly outperformed the overall growth rate of our ad business. In particular, year-over-year growth in SMB revenue accelerated for the second quarter in a row, benefiting from new markets, growth in advertisers and an increase in the average revenue-per-advertiser.

Looking across our products, the vast majority of our year-over-year revenue growth was again driven by Promoted Tweets, up 113% year over year. Website cards, mobile app downloads and Promoted Video were the primary drivers of Promoted Tweet growth. I would point out that two of these three products were introduced in the back half of year, and Promoted Video currently remains in beta.

In terms of geographies, in Q4, US ad revenue grew 70%, while international ad revenue grew 141% year over year, and accounted for 36% of total ad revenue. International growth was driven by strong growth in both APAC and EMEA. Major movers year over year in EMEA included the United Kingdom and Ireland. In APAC, Japan once again performed well, and was the primary driver of results in the region.

We continue to expand our sales force globally, entering 13 new markets in Q4. And in fact, we now have a sales presence in 73 countries around the world, and see significant room for continued international revenue growth, as we continue to sign up new advertisers and further expand and grow each channel around the globe.

Data licensing and other revenue contributed $47 million in the quarter, an increase of 105% year over year. We signed an important new partnership with IBM in Q4, which allows enterprises to incorporate Twitter data into their decision-making. IBM will be a key channel partner for us in that regard. Our mobile ad exchange business continues to see very strong growth in the exchange, with revenue more than doubling versus the prior year.

Moving on to cost and EBITDA, in Q4, total non-GAAP expenses were $390 million, a 73% increase year over year. The year-over-year increase was primarily driven by headcount and related overhead costs, as well as infrastructure investment. We continue to invest in our workforce, to scale our business and drive continued product innovation.

Adjusted EBITDA totaled $141 million, a 216% year-over-year increase. Adjusted EBITDA margin for Q4 was 30% versus 18% in the prior-year period. Adjusted EBITDA significantly outperformed our expectations, due to the out-performance of revenue and the challenge of re-investing revenue upside on such a real time basis.

Non-GAAP net income was $79 million in the fourth quarter, up from approximately $10 million in the same period a year ago. Our GAAP net loss in the fourth quarter was $125 million, which includes $177 million in stock-based compensation expense. Non-GAAP EPS was $0.12, while GAAP EPS was a loss of $0.20.

Before turning to metrics, I'll cover a few items related to cash and CapEx. We ended Q4 with roughly $3.6 billion of cash and marketable securities, unchanged from Q3. Cash flow from operations was $43 million. CapEx totaled $99 million, $30 million of which was financed through capital leases.

Now I would like to turn to our operating metrics. First on users. Average monthly active users reached 288 million for the quarter, reflecting year-over-year growth of 20%, or four million net additions on a quarter-over-quarter basis. There are a couple of specific factors that negatively impacted net additions in Q4.

First, we lost approximately four million net users due to the rollout of the iOS 8 integration, which primarily impacted our third-quarter pulling MAUs, but also, to a lesser extent, impacted Twitter owned and operated MAUs. Additionally, as we mentioned before, Q4 is our seasonally weakest period.

Importantly, we are pleased to see that are current Q1 trend in MAUs will likely result in our Q1 MAUs returning to the level of absolute net ads that we saw during the first three quarters of 2014. Although we don't expect the product launches and tests announced over the last two weeks to have a meaningful impact on Q1 user growth, we're hopeful that these product initiatives will contribute in subsequent quarters.

Timeline views increased to 180 billion, up 23% from the year-ago quarter. And timeline views-per-MAU totaled 631, up 3% year over year, and better than our outlook for Timeline-views-per-MAU to be flat versus Q4 2013. As we announced in November, we do not intend to disclose timeline views or any future periods.

In terms of monetization, ad revenue per 1,000 timeline views continues to show strong growth, reaching $2.37 in Q4, up 60% year over year, and 34% sequentially, and better than our outlook of up 28% to 30% sequentially. In Q4, ad revenue growth was primarily driven by an increase in ad engagements. Importantly, CP also contributed, as demand outweighed the supply we made available.

Ad engagements grew 70% year over year, driven by both an increase in ad load, as well as in total ad requests. Cost per ad engagement grew 10% year over year, due to both mix shift to higher-priced and higher-performing ad units, as well as an increase in same-format CPE. We are seeing strong demand for advertising our platform, and are encouraged by the continued increases in CPE, even with the increase in coverage.

Now I will turn to our guidance. Before I provide the detail on Q1 guidance, I want to quickly provide some color on the relative out-performance versus our guidance for Q4. Last quarter, we told you our guidance placed more weight on the opportunities in the quarter coming to fruition versus balancing the opportunities and risks as we had in the past. And our advice was to not to deviate meaningfully from our guidance range. Clearly our reported results exceeded guidance by a significant margin.

This significant out-performance was driven by greater-than-expected acceleration in December, as many brand advertisers ramped spending around the holiday shopping period, particularly in the US. In fact, when we gave revenue guidance on October 27, the high-end of our revenue guidance range implied 86% year-over-year growth. While our actual October 2014 revenue had increased 86% year over year. So our guidance at that time was very much in line with the growth rate we were actually seeing in our expectation for the remainder of the quarter.

That said, we are once again attempting to be more accurate in our revenue guidance relative to results, as we continue to place more weight on the opportunities in the quarter coming to fruition. For that reason, and based on current visibility, we do not recommend projections that deviate meaningfully from guidance.

In terms of guidance, for Q1, we expect revenue to be in the range of $440 million to $450 million, and EBITDA to be $89 million to $94 million. We expect stock-based compensation expense in the range of $160 million to $170 million. Finally, we expect the share count for Q1 to be approximately 645 million shares on a GAAP basis. And our fully diluted share count to be approximately 690 million shares on a non-GAAP basis.

For the full-year 2015, we expect revenue to be in the range of $2.3 billion to $2.35 billion, and adjusted EBITDA to be $550 million to $575 million. We expect stock-based compensation expense of $700 million to $750 million. And we expect CapEx to be between $500 million and $650 million.

Finally, I would like to note a few things on guidance before taking your questions. First, these guidance ranges are based on currency rates as of January 31, 2015. In fact, currency has negatively impacted our first-quarter and full-year 2015 outlook. Specifically, our Q1 2015 revenue guidance is $6.5 million lower, just due to the currency change from Q4 2014 to January 31, 2015.

Second, as you think about modeling quarterly [assets] in 2015, please consider the following factors. First, revenue in Q1 2014 benefited from the Olympics, which is obviously not going to occur again in 2015. Without the Olympics, our sequential revenue growth in Q1 2014 would have been minus-1%. Additionally, in the second quarter of 2014, we benefited from the World Cup. Without the revenue from the World Cup, our second-quarter 2014 sequential revenue growth would have been 15%.

We are sharing these additional considerations with you to provide the proper context as you think about each quarter of 2015. However, even with these considerations, our guidance still calls for continued robust growth. And we look forward to executing with excellence, in order to have another year of significant out-performance in 2015.

With that, we're ready to take your questions. Operator, would you please announce the first question?

(Operator Instructions)

Mark Mahaney, RBC Capital Markets.

Thanks. Two questions, please. On the slew of third-party deals -- Yahoo! Japan and Flipboard, Google -- congrats on getting those. How should we look for the impact of those in the P&L? Any broad details you can provide would be great.

And then on the engagement metrics going forward, Anthony, what should we be able to look at in order to track what's happening to engagement -- whether it's improving or declining -- if you're no longer disclosing timeline views? Thank you.

Thanks, Mark, this is Dick. Let me jump in there with a brief overview of the way we're thinking about these third-party relationships, and then I'll let Anthony chime in on the details and address the engagement question as well. First of all, on the third-party relationships, really think about those in two ways. There are deals where we're distributing our content to third parties that we're now starting to monetize, and that's new and exciting for us. That's the Yahoo! relationship and the Flipboard relationship. And then, the second kind of deal that we're looking at now and talking about are the Google-type deal.

So I do want to confirm that we have a relationship that we've agreed to with Google. I don't have any more details to share about it at this time. What's new about those deals is that they will drive traffic and distribute traffic to our logged-out experience. And that's the way we're thinking about those relationships now as differently than we were thinking about them previously. I'll let -- Anthony, why don't you jump in and take the rest?

Sure, Mark. In terms of the impact of these third-party deals, they were all considered in the guidance that we provided you today, so they're in that consideration. In terms of engagement metrics, as I mentioned, we're no longer going to provide the metric of timeline view. And the reason for that is it's really a measurement that doesn't reflect the initiatives that we're doing. In fact, if anything, we're taking specific initiatives and product changes that will hurt timeline view. As an example, the recently launched product, While You Were Away, will cause you not have to go through many timeline views to find something that was really important to you eight hours ago.

And so that's why we decided to eliminate the timeline view metric, given that we have specific product changes that will hurt that metric. More broadly, as we think about engagement, there are a number of different ways that we measure engagement -- there's no one perfect way. When it comes to advertising, it's going to be click-through rate. And it's actually different by each format. A mobile app download click-through rate is very different than a regular Promoted Tweet that could be either re-tweeted or favorited as a measurement of payment.

Additionally, on the consumer side, many companies use DAU to MAU. And while that is a long-term goal of ours, to become a daily product, today we have great variance in DAU to MAU across geographies. In our more mature markets, we have very high DAU to MAU, 50% plus. In the emerging markets, we have very low DAU to MAU, at 20% range. They all migrate up to a higher rate over time.

And so as we get to a point where we have a metric that's going to really reflect what we're trying to do, we'll share that with you. But, at this point, there's a number of them that we look at it, and no one metric to share.

Thanks, Anthony. Thanks, Dick.

Great. Thank you. Next question, please, operator.

Paul Vogel, Barclays.

Just two questions. One, first on the MAU number. As think about the first quarter, if you add back the $4 million you lost, the run rate getting back to where you were in the first three quarters is an acceleration. Is that just seasonality? You mentioned that the new products really haven't clicked in yet, so I'm curious as to the acceleration there, if that's seasonality or something else?

And then, on the advertising side, if you could just talk about the pitch you're making to advertisers on terms of monetizing the non-logged-in users, how that's being received, and how you're selling that advertising relative to what you sell on a core logged-in user? Thanks.

Sure. Thanks, Paul, this is Dick. In Q1, I would say it's a combination of seasonality, a return to organic growth, and the set of product initiatives we've created to drive growth. Again, at a high level, I'd like to say that I'm thinking about growth and our product as, these changes we're making now as helping us grow across logged-in, logged-out, and our syndicated audience across the web and third-party mobile apps.

The user numbers we saw on January, again, indicate that our MAU trend has already turned around, and that Q1 trend is likely to be back in the range of absolute net ads that we saw during the first three quarters of 2014. So we're in a great place there. And, again, I would stress that it's seasonality, a return to organic growth, and product initiatives, all taken together.

Paul, one thing I want to clarify is, you mentioned the word acceleration. And so, for everyone's benefit, we added 14 million monthly active users on a net basis sequentially in Q1 of 2014. In Q2 of 2014, we added 16 million net monthly active users. And then, in the third quarter, we added 13 million. When we say that Q1's trends are likely to indicate that we back towards the trend of absolute net ads in Q1 through Q3 of 2014, we're referring to those specific numbers -- 14 million, 16 million and 13 million, not anything else from a percentage basis.

Paul, Dick again. To address your second question specifically, it's going to be the very same ad unit that we brought to our Twitter-owned and -operated properties, and now are syndicating into these third-party experiences that we will be bringing to the logged-out experience on Twitter. It's important for me to highlight that we are first going to be very focused on delivering a delightful user experience.

I talked in my prepared remarks a little bit about the fact that it will be a combination of algorithm-generated timelines and curation. We want to really nail that down. But once that's nailed down, we're already set to go on delivering that promoted Tweet ad unit into the logged-out experience.

Great. Next question, please, operator.

Ross Sandler, Deutsche Bank.

Just two quick questions. Dick, we know it's early, but, given the recent product releases, can you just talk about which are showing the most traction, and how they are impacting engagement?

And then, Anthony, on the full-year guidance -- so, about 70% at the high end, including FX and the lack of Olympics and World Cup, that suggests not a lot of deceleration. Can you just talk about where you expect ad load to be in 2015? And what that growth looks like from a same-advertiser basis versus new advertisers? Thanks.

Ross, it's Dick. In answer to your first question, it's simply too early to have any data to share as to the impact of those features we released in January. I would say that, when I think about them in the context of how they're growth drivers, the features we released in January -- like native video and group VMs -- make Twitter a richer and more enjoyable experience, both for new users and as an increasingly daily use case for existing users.

I'm also, finally, particularly excited about Instant Timeline. When we talked about that at Analyst Day, one of the reasons all of us here are so enthusiastic about it is because we know that the follow action is a high-friction action for new users. And removing that action completely and dropping users into a really high-quality timeline right away, we're hopeful we'll drive immediate engagement and create long-term users out of that large group of people we have coming into the top of the funnel.

And then, Ross, as relates to your advertising question, at the Analyst Day, I talked about an opportunity over the long term to get to 5% ad load. And we didn't say that was the limit -- that was an opportunity over time in how we would get there. We're so far from that 5% today; that's not how we actually model the business. The way we model the business is really on a demand basis, which you alluded to, which is the amount of spending of existing advertisers plus advertising spending from new advertisers.

And I can tell you the bottoms-up forecast and top-down forecast that we've put together for 2015 and that we shared with you on revenue, we still won't be near that 5% to get there. Obviously, the amount of inventory in that 5% will grow with our user base. But we're far from getting to that 5% opportunity long term. And our forecast reflects the demand side of the equation relative to the supply, given we have such access supply compared to demand.

Great. Thanks, guys.

Thank you. And the next question, we're going to take from Twitter. It comes from the Twitter account of Dan Ernst at Hudson Square. And he asks: Why wouldn't a logged-off visitor to Twitter be greeted by a curated, best-of-breed and set of trending tweets, and also a brand ad?

Hi Daniel, it's Dick. That's exactly correct. That's exactly what we will be doing. I would say once again that it's important to me that we really deliver a delightful user experience, first and foremost. One of the reasons we launched that logged-out experience on web first is so that we can iterate quickly as we learn how users are engaging with it. And once we're ready, we will be delivering Promoted Tweets from all sorts of our advertisers into that experience.

Great, thank you. Operator, next question, please.

Doug Anmuth, JPMorgan.

Two things I wanted to ask. First, can you use the algorithms from Instant Timeline? And then also, While You Were Away, to let people develop more customized timelines within their own experience and take that beyond just the logged-out or just the initial experience through Instant Timelines?

And then, secondly, as you enhance the experience for logged-out users, particularly with the launch of Homepages, is there a concern that you could go too far in delivering a good user experience for logged-out that would not let people -- or drive the incentive for them to log in as much? Thanks.

Hi, Doug, it's Dick Costolo. I think that, that algorithm that really powers both that recommendation engine for While you were away and Instant Timeline and others, is an algorithm that we've been working on for some time. It's gotten solid enough that we really like the instant timelines it delivers.

But I think it will be a while, as we continue to learn and rollout Instant Timeline, before we use it in some new and additional way that's beyond what we're using it for now. So, I do think it will be a period of learning around Instant Timeline and iterating on that and improving it, because we have such a great opportunity there with new users that come to the platform.

As it regards -- is the logged-out experience going to cannibalize the logged-in experience in any way? -- actually view it flipped around. It's the case today that we've got this really incredible opportunity around the people that hear about Twitter or are aware of Twitter and are driven to the platform because they see it on TV or they hear about it in the press somewhere. And we throw up this wall in front of them today and force them to go through all of these hoops in order to get started.

So, I'm actually of the mind that, that logged-out experience will be a way to show them what Twitter is and how valuable it is, and all the amazing content that's on the platform. And then, as they want to engage with it, they can become logged-in users.

Okay. Thanks, Dick.

Next question, please.

Heath Terry, Goldman Sachs.

Dick, I was wondering if you could give us a sense of the engineering priorities for this year. How important will product introductions like While You Were Away be versus improving the algorithms around personalization for existing products like the Discover timeline?

Sure. Thanks, Heath. That's -- I can be very straightforward answer about that. The algorithms and the back end, if you will, around the capabilities that are driving these algorithmic While You Were Away, Instant Timeline, et cetera, is a number one priority. It drives so much of the high-quality engagement on the platform that, that's probably first and foremost where we'll devote a lot of resources. Having said that, lots of kinds of product experiences will emerge from that work. The Magic Recs product experience emerged from that work.

As I mentioned, once we get through Instant Timeline, there are lots of other things we want to do with that capability. As another example, we've got an experience out to users who come back to the platform who registered a while ago, haven't been with us, and come back. They previously would come back into whatever -- however many or few accounts they'd followed before. Now we can drop them into a resurrection Instant Timeline, if you will. And we'll focus on that more as we like the results of Instant Timeline. So, that back end is going to power a lot of experiences, and we're spending a lot of time and energy there.

Great. Thanks, Dick.

Great. We'll take our next question from Twitter from the account of Matt Zolan. And he asks: Do you see impact to revenue in 2015 as a result of strength in the US dollar versus other currencies?

Thank you, Matt, it's Anthony. Our 2015 revenue guidance of $2.3 billion to $2.35 billion is, in fact, impacted by currency. On a constant-currency basis to January 31, 2015, it would have been $55 million to $56 million higher than what we actually gave as guidance.

Thank you. Next question, please, operator.

Eric Sheridan, UBS.

Anthony, two for you on the guidance going forward. The last couple of quarters have really demonstrated a tremendous amount of leverage in the business, in terms of the margins and the beats on the EBITDA line. Any update you can give us on what, other than just incremental revenue, is driving that leverage? And whether it's informed any new way you're thinking about the long-term profitability of the Company?

Second, on CapEx, that was a bigger increase than we thought, year on year, and the way you're forecasting 2015. Any way you can carve out the priorities for those dollars in the capital budget in 2015? Thanks so much.

In terms of leverage, at Analyst Day, we took the time to give a new perspective on our long-term margins. During the IPO, the Company had articulated a long-term margin profile of 35% to 40%. And, at Analyst Day, given the leveraged outperformance that we achieved as a public Company during the first year, and having more information about the business more broadly, we had raised our long-term adjusted EBITDA margin to 40% to 45%.

The leverage that we're getting today is primarily due to the outperformance in revenue and not being able to re-invest it fast enough. And while we have contingencies in the quarter, on the positive side, the investment in certain areas, if things do outperform -- it's just a really hard thing to do in a real-time basis.

There are productivity initiatives, as you'd imagine, in addition to just operating leverage. But that contributes a very small component to the overall operating leverage that you're seeing today, at this point. But it is a long-term opportunity for us, and that's why we gave the long-term margin of 40% to 45%.

As it relates to your CapEx question, there's two big investments in CapEx in 2015. One is a third data center. We have two data centers now. And the third data center will, importantly, allow us to have greater redundancy. Today, we have to completely replicate what's in one data center into the other data center. Having a third data center will allow us to get some productivity leverage as well.

The second thing that's impacting CapEx in 2015 is real estate capital expenditures. We're building out additional space in San Francisco and in New York, and those are two big expenditures as relates to building out CapEx. But we're also in the process of continuing to upgrade our locations from a capacity standpoint internationally. And so we're seeing CapEx go up quite meaningfully, over 100%, from real estate.

Great. Thanks for the color, Anthony.

Thank you. Next question, please, operator.

Justin Post, Bank of America Merrill Lynch.

Just looking -- and I know you don't disclose the churn. Can you just give us your high-level thoughts on user churn? Maybe some kind of numbers around what percent of people have churned off of Twitter or any thoughts you can provide there?

And then, what initiatives you are doing to get those people back that might be working so far. And, are you optimistic you can bring a lot of these people back this year? Thank you.

As a relates to churn broadly, we don't release the specific number. But I would tell you churn has improved in November and December compared to the prior months, and we're encouraged by that. Our growth team has a number of initiatives that are specifically addressed at converting unhealthy users to healthy users very early in their lifetime on Twitter. And, moreover, Dick talked to an Instant Timeline. The whole goal of Instant Timeline is to make the new user a healthy user, day one. Second one, for that matter.

And so, there's a lot of initiatives against the whole concept of having healthy users from day one. And if they're not, addressing that very quickly after they've come onto the platform, before we miss the opportunity to improve that. So we're encouraged by the trend in churn; there's a lot of initiatives there.

Thank you.

Great. And we'll take our next question from Twitter. It comes from the account of Kevin McGoldrick. He asks: How do you see Vine's relationship with the new native video option? And how to you see that interaction playing out?

Thanks, Kevin, this is Dick. Let me pan back and talk about video across the Twitter ecosystem between Vine, consumer video on Twitter, professional video on Twitter, et cetera. Our mission is to give people the power to capture and share ideas and information instantly, without barriers. And that word, instantly -- capturing life or a moment or an event as it happens, and communicating it as it happens -- is how we think about Twitter video being vital to consumers.

It's one of the reasons we created that simple capture-and-share capability in the Tweet Composer. You can think about all those benchmark moments on Twitter -- New Year's Eve, Ferguson, et cetera. Imagine being able to capture those on video and sharing them instantly with the world. That's where we think we're really going to stand out.

I think of Vine as really a creative medium in its own right, an artistic medium in its own right. We see the emergence of all these Vine stars and what they've been able to do on the platform as distinct from but complementary to native mobile video across the Twitter ecosystem.

Thank you. Operator, next question, please.

Anthony DiClemente, Nomura.

Just one for Dick, and one for Anthony. Dick, just a follow-up on your commentary just now about video. Can you just elaborate on where professional video fits into that? Maybe give us an update on Amplify and your professional video efforts?

And then, Anthony, just on monetization, you mentioned that CPE grew 10%. That's a strong number. You did mention that part of that was attributed to mix shift, and part of it was growth in same-format CPE. I'm just trying to get a little bit more -- dig into those two, in terms of what was a bigger driver.

And then, moving forward, what are the higher-priced ad products? Is there a mix shift to a higher-priced ad product, be it video or another, as we head into future quarters that will potentially continue to drive CPE growth, as driven by mix shift? Thanks.

Sure. This is Dick. To your first question, regarding professional video, I think of there as being a continuum from -- at one end of the -- are content partners like the NFL and the NBA and the broadcast networks. Maybe at one end of the spectrum, to you and me and our friends at the other end of the spectrum. And then all these folks like the Vine stars, if you will, in the middle there, who are increasingly becoming professional and migrating from consumer users of video to professional. We want to provide tools across that spectrum.

So what you're seeing from us now is, one, broadening the rollout of our Amplify program globally to more than 130 partners now. And layering onto that, capabilities like SnappyTV and video. Twitter.com for other kinds of content creators, who may not be a broadcast network or the NFL or the NBA but seek to monetize their content and distribute longer than 30-second video on our platform.

When I mentioned briefly earlier in my prepared remarks that we will also be providing more and more capabilities there, that's what I was referring to. That middle tier of content creators who are looking to do more, will see more and more capabilities and tools from us on a platform.

And then, Anthony, as it relates to your question on CPE versus engagements, engagements was the larger driver of our overall growth, up 70% year over year. From a CPE perspective, it's somewhat of a complicated answer. From a pure mix perspective, our faster-growing newer units are higher-priced units, but lower click-through rate unit.

As an example, a mobile app download or video, they have different mechanisms that are the pay-for-performance mechanism that drives the click-through rate. And we're charging differently for them, based on a click-through rate that's lower in the overall funnel from a marketing perspective. And so CPE should improve as that mix shift continues in that direction. And click-through rate could, in fact, decrease as we do that. Which is what we saw this quarter.

But there's one underlying economic trend that also matters, and that's the balance between supply and demand, which is driven by our load factor. And so there could be circumstances where our supply is infinitely greater than our demand, and you wouldn't see that increase in CPE, because of the excess supply relative to demand. In this quarter, that was not the case. In this quarter, we did increase load factor. But demand was still greater than supply, and that's why we saw a CPE increase, both due to mix and same-format CPE.

Got it. Really helpful, thank you.

Thank you. Operator, next question, please.

Ben Swinburne, Morgan Stanley.

Anthony, in the guidance, do you assume much contribution from the new syndicated advertising partnerships you have and will add? And, if you do, could you just help us think about gross versus net, if there's any kind of rev share associated with that?

And then, Dick, can you just give us some color on why Twitter ended its Google integration years ago? Why it made sense then and doesn't make sense now, and now it makes sense to get together? Some color there would be great.

Sure. On the first part of your question, Ben, what I'd say is, these deals were considered in our guidance. I didn't provide a perspective on whether there was revenue or how much, I said they were considered. In terms of how revenue will be recognized as it relates to these syndicated monetization deals, it really comes down to what channel it goes through. We have two channels. One is an exchange, and that revenue is recognized on a net basis. If it goes through an ad network, that revenue is recognized on a gross basis. And it depends on where the demand's coming from and who the winner of a specific ad is.

Hi, Ben, it's Dick. Let me try to answer the Google question this way, if I may. We've obviously had a relationship with Google over the course of the years. We all -- a bunch of the executives here and a bunch of the executives there obviously know each other quite well.

I would say that, the way we think about the Google deal now -- again, without going into any of the details -- distinct from the kind of relationship we had in the past, is that we've got the opportunity now to drive a lot of attention to and aggregate eyeballs, if you will, to these logged-out experiences, topics, and events that we plan on delivering on the front page of Twitter. And that's one of the reasons this makes a lot more sense for us now.

Thank you.

Thank you. Operator, next question.

Ron Josey, JMP Securities.

I wanted to come back a little bit on product improvements. I know we talked a lot about it. But specifically on the on-boarding process, I'm wondering if the revised process has helped to really convert the logged-out users to logged-in users. And, if that is the case, if that's giving you all confidence in seeing 1Q on the [U trends] return to where they were a year ago? Thank you.

I would say that none of the work we've done to-date around logged-out users or experiences for users who aren't logged into Twitter has had any impact at all yet on the kind of outlook we provided to you for Q1. And then, again, over time, we think about our ability to get people excited about Twitter content and experience Twitter content by coming to the front page of the site, or other ways they might get to a logged-out Twitter experience as an on-ramp or an on-boarding vehicle into eventually becoming a more engaged logged-in user.

Great, thank you.

Thank you. We'll take our next question from Twitter from the account of Wayne [Hua]. The question is: Do you have forecasts as to how beneficial the new Google deal will be with regard to user growth?

We're real excited about the new Google deal. Everyone's asked questions about it, so thank you. Dick mentioned very clearly that we're doing this because we have a great opportunity to leverage the tremendous audience they have and their interest in relevant, real-time information, and bring that to our logged-out home page.

A logged-out home page doesn't necessarily those users will translate into monthly active users. If we do our job and give them the right type of content, and cause them to be engaged, and give them the right experiences over time, that will happen. But, today, that's not what we're focused on, nor what we expect. Over the course of time, that is an opportunity though.

I'll just chime in there, one more piece of information. This is not something that you're going to see launched or rolled out for several months. Just to make that very clear.

Great. Thank you. Operator, next question, please.

Mark May, Citi.

I had two. And, Anthony, I'm actually going to ask another one on Google. I believe that you're generating data licensing revenue from that deal. And, if so, given Google's scale, it would seem that, that could be a material incremental new source of revenue. Can you give us a sense of how that's impacting Q1 and for your revenue?

And then, if you could provide a little more color on what exactly happened with the iOS 8 update that you referenced? And is it your expectation that you'll be able to recover most, if not all, of those 4 million users that you talked about in the new year?

Sure. As it relates to Google and any revenue considerations, we're not providing any perspective on that at this time. The one perspective relative to your question I would provide just echoes Dick's points that this is a relationship that we'll roll out in several months from now. Which means not Q1, since we're already in February.

In terms of your second question, let me just give you some specific numbers as it relates to iOS 8. We said we lost 4 million monthly active users due to the iOS 8 integration. 1 million of those monthly active users were Twitter-owned and -operated monthly active users. And 3 million were on Safari -- what we call auto-pulling MAUs, and we lost those. We don't expect to get the 3 million auto-pulling MAUs in Safari back, and that's a non-Twitter-owned and -operated auto-pulling MAU. The 1 million -- that number was actually higher at a different point in the quarter, and we were able to bring it back down just to 1 million.

Let me add even a little bit more color there that might be helpful. We obviously have a great relationship with Apple. I've talked about that at length over the course of the last few years. On the second part of what Anthony talked about there, there was an unforeseen bug in the release of iOS 8 as it relates to the specific Twitter integration into iOS. That's why it was particular to us. Once we understood the issue, we moved as quickly as we could on multiple fronts to minimize its impact. But it wasn't a one-size-fits-all fix, which is why you've seen some of the complexity that we talked about here, in bringing those users back. The problem was complex and affected different users differently.

Thanks.

Thank you. Operator, next question, please.

Peter Stabler, Wells Fargo Securities.

This is Steve filling in for Peter. Some very nice developments on the content side here this quarter. Hoping you can provide an update on commerce efforts and retail initiatives in the fourth quarter? We know that brands love to tell their stories to their Twitter users. But we also think the ability to prove out ROI through e-com and direct transactions and increased efforts in the area of attribution can play a larger role. Would you comment on successes with direct response advertising, yield-driven communication, or click-to-buy from tweet?

Sure, Peter, this is Dick. I'll comment on commerce, generally. And then, Anthony, if you have anything you want to add, chime in. We continue to experiment with buy now and offers in commerce on Twitter. You may have seen some of the things we were running during the Super Bowl, with some of the people and events and groups participating in the Super Bowl. We don't have anything new to add on top of that. I don't have anything additional to announce right now, other than those continued experiments that we're running.

And as it relates to direct response, in my prepared remarks, I talked about the fact that the SMB channel saw a second consecutive quarter of faster year-over-year growth rate earn acceleration, so we're encouraged by that. But it's still very early days. We had shared with you the number of advertisers that we had in that channel, at Analyst Day. The number's growing very nicely. It's very encouraging, but it's still sub-100,000. And so there's still a big opportunity for us as it relates to DR.

Great. Thanks for the color.

Thank you. Operator, I think we have time for just one last question, please.

Youssef Squali, Cantor Fitzgerald.

Two quick questions. First, what was the impact of the self-service ad products overseas this quarter? You guys opened a whole bunch in the last quarter. Can you comment on the mix between the direct sales and self-service ads overseas?

And then, second, the impact of -- I know this is early -- but just the impacts of Promoted Video in the quarter, and pricing you're seeing there?

As it relates to self serve, that exists in both the MMS channel and the SMB channel. We haven't broken out those specific revenue segments, and we're not prepared to do now. But as I mentioned, they both performed very well and grew faster than our overall growth rates.

As it relates to self-serve international versus domestic, domestic self-serve is still bigger on an absolute dollar basis than international. But we haven't broken out those segments either.

Video?

Thank you all for joining us. We appreciate your time and look forward to speaking with you again next quarter.

Ladies and gentlemen, thank you for participating in today's conference. This does conclude the program. And you may all disconnect. Have a good day, everyone.