百度 (BIDU) 2017 Q4 法說會逐字稿

Hello, and thank you for standing by for Baidu's Fourth Quarter and Full Year 2017 Earnings Conference Call.

(Operator Instructions) Today's conference is being recorded.

If you have any objections, you may disconnect at this time.

I would now like to turn the meeting over to your host for today's conference, Sharon Ng, Baidu's Director of Investor Relations.

Hello, everyone, and welcome to Baidu's Fourth Quarter and Full Year 2017 Earnings Conference Call.

Baidu's earnings release was distributed earlier today, and you can find a copy on our website as well as on newswire services.

Today, you will hear from Robin Li, Baidu's Chief Executive Officer; Qi Lu, Baidu's Chief Operating Officer; and Herman Yu, Baidu's Chief Financial Officer.

After their prepared remarks, Robin, Qi and Herman will answer your questions.

Before we continue, please note that the discussion today will contain forward-looking statements made under the safe harbor provisions of the U.S. Private Securities Litigation Reform Act of 1995.

Forward-looking statements are subject to risks and uncertainties that may cause actual results to differ materially from our current expectations.

Potential risks and uncertainties include, but are not limited to, those outlined in our public filings with the SEC, including our annual report on Form 20-F.

Baidu does not undertake any obligation to update any forward-looking statement except as required under applicable law.

Our earnings press release and this call include discussions of certain unaudited non-GAAP financial measures.

Our press release contains a reconciliation of the unaudited non-GAAP measures to the unaudited most directly comparable GAAP measures and is available on our IR website at ir.baidu.com.

As a reminder, this conference is being recorded.

In addition, a webcast of this conference call will also be available on Baidu's IR website.

I will now turn the call over to our CEO, Robin Li.

Hello, everybody, and thank you for joining today's call.

2017 has been a transformative year for Baidu.

We took steps to exit non-core businesses, added new management talents and sharpened our strategic focus to strengthen our mobile foundation and lead in AI.

Our shift to make Baidu an AI-first company is paying strong dividends.

AI continues to be a main driver behind the growth of our twin-engine search-and-feed business.

By coupling AI computing with user data across Baidu's 20+ apps, many with monthly active users in hundreds of millions, we generate amazing user insight to power Baidu's search and feed algorithms, which in turn drive our top line growth and generate cash flow to fund our new AI businesses, autonomous driving and conversational AI.

We had a solid quarter for fourth quarter.

Total revenues reached CNY 23.6 billion, growing 29% year-over-year, which is a meaningful acceleration from the first half of 2017.

Baidu's fourth quarter revenue was driven by the robust growth of our feed revenue as well as strong traffic demand, particularly coming from e-commerce companies.

In December, we signed a strategic partnership with Huawei to cooperate in areas of technology, product and content.

Baidu will support Huawei in the development of AI-powered smartphones and make available our mobile apps such as our flagship Baidu app and Baidu Maps, which are gradually upgraded with AI features such as Baidu voice assistant, intelligent search and augmented reality.

Our comprehensive partnership with Huawei is at a new level of strategic cooperation, and we hope to form comprehensive partnerships with other smartphone players in the future.

By working with handset makers at the outset of product design, we hope to enable a new generation of AI-powered phones, which will significantly improve user experience, especially when combined with Baidu AI-enabled mobile apps.

I would like to point out that Mobile Baidu has been rebranded as Baidu app.

The rebranding reflects Baidu app as the main channel to experience Baidu services.

When we launched Mobile Baidu several years back, the app was considered an extension of our PC services and the mobile channel to access Baidu search.

Today, the Baidu app is the main channel to access Baidu's twin engine of search and feed and the container that offers the best Baidu experience.

On our new AI businesses, we launched the DuerOS 2.0 in November at our annual conference, Baidu World, and last month announced Apollo 2.0 at the Consumer Electronics Show in Las Vegas.

Baidu's autonomous driving platform, Apollo, has received great industry recognition, and we have formed strategic partnerships with the industry heavy weights such as Intel, NVIDIA, NXP and Renesas.

For Baidu's conversational AI platform, we continue to focus on growing the adoption of DuerOS at a rapid pace.

Last November, we released our first, first-party showcase hardware, the Raven H smart speaker, which received industry accolades, including "The Best of CES 2018" by The Wall Street Journal and one of "The Best Gadgets of CES 2018" by WIRED.

We now have over 130 DuerOS partners, who, together, have released over 50 DuerOS-powered smart devices.

And we are excited to see the DuerOS install base and usage grow very rapidly.

2017 has proven to be a very fruitful year for Baidu.

We look forward to continuing our momentum in 2018.

I'll now turn the call over to Qi to go through our business progress.

Thank you, Robin.

On our core business front, we remain very focused on strengthening our foundation for the long term, particularly through leveraging AI technologies to improve the user experience for Search, Feed and our flagship Baidu app, formerly named Mobile Baidu.

For Search, we are making steady progress on improving Baidu's search experience and expanding the scope of search with richer content such as video.

Specifically, the coverage of our Top-1 results, whereby we satisfy our users' intent with the first search result, has reached 30% of queries in Q4.

We have also made strong progress in improving the coverage and quality of video content for search, with 15% of our search page views returning high-quality videos, up from 11% in the previous quarter.

Note that in order to fully leverage the strong growth of video content, we introduced a more stringent quality measurement standard for quality video coverage starting in Q4.

Even more importantly, last November, we successfully launched Bear Paw account, a new platform for mobile web content that is designed to link the content from other platforms and deliver a significantly upgraded mobile search experience by enabling social tools such as users clicking to follow, businesses broadcasting features, transaction tools such as payment and content management tools.

The initial adoption has been very strong, with over 25% of searches within Baidu app returning content from Bear Paw account publishers.

For Feed, we are making strides in improving the core foundation across the board.

In Q4, our product quality continued to improve.

As a result, the daily content distributed by feed grew over 20% in Q4 over the prior quarter.

We continued to deepen user engagement with total daily time spent on the Baidu app growing approximately 30% year-over-year.

Video content has been a key growth driver not only as part of feed but also as part of our overall mobile product matrix.

Specifically, the percentage of video content distributed in feed has grown to 40% on average in Q4, up from 14% a year ago.

We continue to improve our Baijiahao platform, which connects content from self-media publishers with our feed users.

We saw the number of content creators grow from over 200,000 at the beginning of last year to 1 million currently.

We also provided content creators with cutting-edge AI tools through the launch of a Content Creator AI Platform, which helps automatically review material for edits, intelligently splice video into images and more.

Our flagship Baidu app, which offers the best Baidu product experience, continues to grow strongly, driven by the twin engines of Search and Feed.

We released Baidu app version 10.0 at Baidu World in November, with a much improved user experience, featuring smoother navigation and content viewing, expansion of enriched content such as video, user-generated content, vertical channels and light social sharing, to name a few.

In addition, we're trailblazing to fully leverage AI capabilities to drive more and faster pace of product innovation such as the brand-new text-to-speech version of the Baidu app.

On the advertising front, we continue to make steady improvements to our advertising product offerings and ad platform to drive sustained growth in Search and Feed monetization.

For example, we have seen strong adoption of dynamic ads.

Dynamic ads can be used in search and feed and help our customers reach the right users on the platform by the showing users highly relevant, personally tailored ads, driving significantly higher conversion rates.

Dynamic ads have been widely adopted by our customers in transaction-driven verticals with structured product catalogues.

In Q4, we drove higher adoption in existing verticals such as e-commerce and travel, and we expanded dynamic ads to new verticals, including auto and real estate with strong results.

We're in the early innings of dynamic ads and see very promising opportunities ahead.

We know that the ultimate end goal for our customers is leads conversion at scale on our platform.

In the second half of last year, we introduced Optimized Cost Per Click, or oCPC, on the merged search and feed ad platform.

Customers can give us their budgets and cost per conversion, and our ad platform will optimize search and feed traffic to deliver the highest number of conversions that fall within their budgets.

For example, in the online gaming vertical, customers can set a budget and target price per download.

Our ad platform intelligently optimizes bid price and other criteria for both search and feed traffic to guarantee as many app downloads as the budget allows.

The technology is a culmination of efforts dating back to 2013 and entails the use of big data and proprietary machine learning modeling on conversion tracking and traffic analysis.

Overall, we're working hard to strengthen our value proposition to advertisers, offering the ability to obtain leads conversion at scale through the breadth of our advertising platform, which includes search, feed and various forms of display ads, through innovative ROI-enhancing ad formats, which include dynamic ads, action ads, augmented reality ads and through constant upgrades we make to our technology and data-driven ad platform, of which oCPC is a prime example.

Lianjia and New Oriental, are 2 such advertisers who used a range of ad products to reach the users they seek.

Lianjia, one of the largest real estate broker chains in China, leverages Baidu's search ads, feed ads, Brand Zone and dynamic ads, mobile canvas ads, and image and video ads to promote their brand and services.

Baidu's AI technology powers user targeting, knowledge graph, lookalikes and more.

It is fundamental to delivering dynamic, customized search and feed ads that significantly improve Lianjia's lead conversion.

Baidu delivers, on average, over 5 million high-quality traffic daily to Lianjia, a scale which is multiple times that of any other advertising channel.

New Oriental, one of the largest private educational services providers in China, partners with Baidu to bring leads to their platform.

New Oriental has been a longstanding customer of our search ads and has been an early adopter of feed ads.

Over the past year, New Oriental increased their feed marketing budget six fold.

Now turning to iQIYI.

iQIYI has continued to lead the online video industry with its unique blend of entertainment and technology DNA.

In December, according to iResearch, iQIYI's PC and mobile app daily average active users reached 76 million and 158 million, respectively.

Monthly time spent on iQIYI's mobile app grew to 408 billion minutes, 22% higher than December 2016.

Next, I'll talk about our progress on our AI-enabled new businesses.

For Apollo and our autonomous vehicle business, we've made big strides in Q4, especially with the launch of Apollo 2.0 at CES, which provides an unprecedented set of capabilities for simple urban road autonomous driving to our developers and our partners.

This elevates the Apollo ecosystem into a new level where we began an accelerated pace, which we would call China Speed, of productization and commercialization of autonomous vehicles.

Over the past quarter, we announced commercial volume production of L4 minibuses in 2018 and several commercial volume production of passenger vehicles in 2019 and 2020.

The Apollo ecosystem is growing rapidly, now with over 90 partners.

Furthermore, Apollo has been designated by China's central government as the national autonomous driving platform.

For DuerOS, our conversational AI platform, we launched DuerOS 2.0 at the Baidu World in November with much elevated platform capabilities and a series of new innovative DuerOS-powered products at the CES.

And more importantly, the DuerOS ecosystem is growing rapidly.

We have teamed up with over 130 partners and have released over 50 DuerOS-powered and branded hardware products on the market, including televisions, speakers, smartphones, story-telling toys, white goods appliances and car dashboards.

And as a result, the overall DuerOS install base of devices and aggregate users and usages are growing at a very rapid pace.

On our other AI-enabled business, the ABC intelligent cloud business continues its rapid growth in Q4, particularly in key verticals such as media, financial services and mobile phone OEMs.

Our Financial Services Business also grew healthily and substantially improved its core capabilities in both consumer credit loan as well as asset management.

We have also made good progress in establishing a new operating structure that will enable the FSG business to further innovate and pursue bigger opportunities.

Overall, we are pleased with our progress, and we look forward to continuing our work in 2018.

With that, I will turn the call over to Herman to go through the financials.

Thanks, Qi.

Hello, everyone.

Welcome to Baidu's Fourth Quarter 2017 Call.

Let me go through the financial highlights.

All monetary amounts are in renminbi, unless stated otherwise.

Q4 was a great quarter for Baidu.

Our total revenues grew 29% year-over-year to CNY 23.6 billion, which exceeded the high end of our guidance.

Non-GAAP operating income reached CNY 5.8 billion, up 104% year-over-year.

Adjusted EBITDA reached CNY 6.9 billion or 29% of total revenues and was up 78% year-over-year, illustrating the operating leverage potential of our financial model as we simplify and align Baidu's businesses around AI.

Let me give you more color on our financials.

During the quarter, we saw strong revenue growth coming from feed, and as a gateway to online traffic, Baidu benefited from the high season e-commerce activities.

Our mobile revenue reached 76% of total revenues and compared to 65% for the same period last year.

Baidu's online marketing services revenues in the fourth quarter grew 26% year-over-year to CNY 20.4 billion.

And we had approximately 460,000 online marketing customers in Q4, up 2% year-over-year.

Revenue per online marketing customer was CNY 44,300, showing a healthy increase of 25% year-over-year.

Other services revenues in the fourth quarter were up 53% year-over-year to CNY 3.1 billion, mainly as a result of robust growth in iQIYI membership and fees for our Financial Services business.

Total revenues in 2017 reached CNY 84.8 billion, up 20% year-over-year.

Online marketing revenues were CNY 73.1 billion, up 13% from 2016.

We had about 775,000 online marketing customers in 2017, down 21% from 2016 as we rebuild our search business.

Revenue per online marketing customer was approximately CNY 93,500, up 43% from 2016.

Other services revenues in 2017 were up 94% year-over-year to CNY 11.7 billion, mainly as a result of robust growth in iQIYI membership and fees for our Financial Services Business.

Moving on to cost of sales.

Cost of sales in the fourth quarter was CNY 11.4 billion, up 18% year-over-year.

Cost of sales in 2017 was CNY 43.1 billion, up 22% from 2016.

Content cost in 2017 was up 70% year-over-year to CNY 13.4 billion, mainly due to increased content purchasing by iQIYI.

In 2018, we expect content costs to step up at a similar pace to 2017.

We plan to continue to invest in self-produced content and high-quality licensed entertainment content to support iQIYI.

On the feed side, we plan to ramp up our content offering to support the growth of Baijiahao, or BJH, account.

Moving on to operating expenses.

SG&A expenses in the fourth quarter were CNY 3.6 billion, up 9% year-over-year.

SG&A expenses in 2017 was CNY 13.1 billion, down 13% from 2016, mainly due to the cut back in promotions from Baidu Deliveries and O2O initiatives.

With promotions from Baidu Deliveries and O2O initiatives bottoming in 2017, we expect SG&A in 2018 to rise on a year-over-year basis as we ramp up the promotion of our flagship Baidu app and other planned app releases.

The magnitude of our channel and other promotional spending increase will depend on our ability to find marketing resources that can generate positive ROI over the lifetime of new app users.

Since marketing revenue derived from a new user is spread over several quarters while related channel and promotional costs are expensed in the quarter that they occur, successful promotion of Baidu apps may temporarily cause our margins to decline in the quarter of heavy marketing spending.

R&D expenses in the fourth quarter were CNY 3.7 billion, up 25% year-over-year.

R&D expenses in 2017 was CNY 12.9 billion, up 27% year-over-year, mainly due to an increase in personnel-related costs.

In 2018, we expect R&D expenses to continue to rise, perhaps at a higher pace as we plan to step up our investment in AI initiatives.

Share-based compensation, which is allocated to related cost and expense line items, in the fourth quarter was CNY 977 million compared to CNY 632 million in the corresponding period in 2016.

Share-based compensation in 2017 was CNY 3.2 billion compared to CNY 1.8 billion in 2016.

Operating profit in the fourth quarter was CNY 4.8 billion, up 118% year-over-year.

Non-GAAP operating profit was CNY 5.8 billion, up 104% year-over-year.

Operating profit in 2017 was CNY 15.7 billion, up 56% from the corresponding period in 2016.

Non-GAAP operating profit in 2017 was CNY 18.9 billion, up 60% from 2016.

Other loss, net in Q4 was CNY 294 million compared to other income, net, of CNY 1.8 billion in the corresponding period of 2016.

Other income, net, in 2017 includes investment gain from the disposal of Baidu Deliveries, while other income, net, in 2016 included investment gains resulting from exchanging Uber China shares for Didi's.

Income tax expense in the fourth quarter was CNY 929 million compared to the CNY 401 million in the corresponding period of 2016.

Fourth quarter 2017 effective tax rate was 18% compared to 9% last year, the increase of which was mainly due to tax refund received as a result of newly granted preferential tax licenses for certain PRC subsidiaries in Q4 2016.

Income tax expense in 2017 was CNY 3 billion compared to CNY 2.9 billion in 2016.

Effective tax rate in 2017 was 14% compared to 20% in 2016.

The decrease in effective tax rate in 2017 was primarily due to a nontaxable investment gain.

For 2018, we expect our effective tax rate to be in the high teens.

Net income attributable to Baidu for its fourth quarter was CNY 4.2 billion.

Diluted EPS was $0.12 -- CNY 12.

Non-GAAP net income attributable to Baidu was CNY 5.2 billion and non-GAAP diluted EPS was CNY 15.

Net income attributable to Baidu in 2017 was CNY 18.3 billion, up 57% from 2016.

Diluted EPS in 2017 was CNY 52.

Non-GAAP net income attributable to Baidu was CNY 22.3 billion, up 68% from 2016.

Non-GAAP diluted EPS was CNY 64.

Adjusted EBITDA in the fourth quarter reached CNY 6.9 billion or 29% of total revenues and was up 78% year-over-year.

Adjusted EBITDA in 2017 was CNY 23.3 billion or 28% of total revenues and was up 48% year-over-year.

As of December 31, 2017, Baidu had cash, cash equivalents and short-term investments of CNY 100 billion -- CNY 100.5 billion.

Our net operating cash flow and capital expenditures were CNY 10.0 billion and CNY 1.2 billion, respectively.

Total headcount as of December 31, 2017, was about 39,300, down 14% year-over-year.

Turning to first quarter 2018 guidance.

Let me first talk about ASC 606, which is a new revenue accounting standard taking effect on January 1 this year.

ASC 606 requires Baidu to report revenues on a net basis, excluding value-added taxes.

For fiscal 2017, our revenues would have been approximately 5.6% lower on a net basis.

For the first quarter 2018, we expect revenues, net of VAT, to be between CNY 19.86 billion and CNY 20.97 billion, representing a 25% to 32% increase year-over-year or a 6% to 11% decrease on a sequential basis.

This forecast is our current and preliminary view, subject to change.

I would like to open the call.

Operator?

(Operator Instructions) Our first question comes from the line of Wendy Huang from Macquarie.

I just wonder if you can give some color on your news feed revenue target for 2018.

And also, can you give us some update on your volume production target for the minibus collaboration with King Long?

Wendy, let me answer the first question and then Qi will answer the second question.

With regards to our feed, as I mentioned in our prepared remarks, feed was one of the main drivers of our revenue growth.

And we're seeing sequential growth of over 20%.

And in 2018, our plan is baked in a sequential growth at that rate.

You mean the sequential growth will be like around 20% for each quarter in 2018?

Yes, in the fourth quarter, we saw that sequential growth.

And currently for 2018, we plan to grow at that kind of rate.

Okay.

Let me answer the second part of your question, which is the planned L4 minibus volume production with our partner King Long Bus.

First of all, the volume production schedule for the product to be on the market is second half of this calendar year.

That's number one.

Number two, we're already seeing very, very healthy inbound interest of early bookings for that product.

That's number two.

Number three, what I want to emphasize here is the accelerated pace of Apollo business opportunities, especially in this particular category, L4 vehicles.

The key is to think about L4 with lower speed and in a more constrained environment.

Once you have that perspective, all of a sudden, you will have much, much more commercialization and productization opportunities.

L4 minibus commercial production for this year is a strong first step, but the inbound interest from our partners will have similar or different approach.

Think about logistics.

For example, a logistic vehicle doesn't have to drive at higher speed.

It can drive at a certain limited hours, let's say, from midnight to 4 a.m.

in constrained paths and planned set of routes.

So the overall is we're super, super happy and excited about the set pace of commercialization and productization of our autonomous driving platforms.

You can think about King Long minibus in 2018 is the first step of many such and more expanded L4 opportunities for our platform.

Our next question comes from the line of Alicia Yap from Citigroup.

My question's also actually related to the autonomous partnership and also the delivery of the vehicles.

Specifically for these L4 buses with King Long, should we start to assume some service fee that Baidu able to generate starting second half this year?

And in terms of other OEM partner, what would be the monetization potential once the delivery happens later in the '19 and the 2020 level?

And then separately on the housekeeping questions, on the sequential decline in the online marketing customer, what was the reason for that?

And how what will the trend be in 2018?

Okay.

Let me take the first part of the question.

With regard to the auto platform, essentially, our business model, we've been super clear from the get-go what's our business model, which is our business model is we'll be selling services, a suite of expanding set of services that's essential to power our autonomous driving vehicles of different types.

For example, HD maps is the first service that we're commercializing, and we are building a sales pipeline with inbound set of commercial opportunities.

On that parallel, you can think about Apollo and Android is very similar from business model perspective.

Android OEM, let's say Samsung, can take Android free to build the phones, but you have to use Google Search services, which is Google's primary vehicles for monetization.

And here for Apollo -- at this stage of Apollo, it's very similar.

HD maps is our first services, but we'll have more and more services that's being planned and in development.

So over time, we see just the services alone over the coming years will be growing to a very material and significant revenue source for the company.

Having said that, the key caveat is that this market will take time to build.

We are focusing on building the health of the ecosystem, elevating the technological capabilities, ensuring more and more partners embrace our platforms and over time, sustainably, healthily build up that business.

And the last thing I will say is the Apollo platform, the Apollo ecosystem preserves many different options of future business model expansions.

So overall, we are super bullish about the long-term economic earning capabilities.

I will emphasize long term, emphasize our focus on building a healthy ecosystem at this stage.

Let me answer your second question with regards to sequential drop at our customers.

So during the quarter, we started to wind down the Local Express business, which is you can search, for example, for local restaurants and so forth.

That business is probably 1% of our revenue, but it has customers in the tens of thousands.

So I don't think that is meaningful from a total customer perspective.

I think it's more meaningful as a pattern that we have throughout all of 2017.

Just look at where our core businesses are, and those that are not as relevant to our total core business, we started to tune back their budget so that we can focus on businesses that are focused in AI.

Can I ask what is the Local Express?

Is it related to Nuomi?

Yes, it was part of our Nuomi business unit.

Our next question comes from the line of Gregory Zhao from Barclays.

First, I have a quick follow-up question to Wendy's question.

So for the 20% revenue growth in the news feed, do you mean in Q1 '18 and in Q2 '18 you both expect the 20% sequential news feed advertising revenue growth?

That's my first question.

The second question is online marketing customers.

So your online marketing customers showing a turnaround trend with like 2% year-over-year growth compared to 1 year ago.

So just want to understand, which business category or segment are you seeing the strongest marketing customer rebound?

On feed revenue growth, I think you brought up a good point.

As I mentioned earlier, we saw a pretty good feed revenue growth on a sequential basis over the last 2 quarters.

In Q4, it grew over 20 some percent.

I would expect that, going into 2018, we will be able to see that in the near future.

With that said, as you know, historically, when you're looking at our advertising business from Q4 to Q1, you typically have a seasonal change.

From our guidance, you can see that we're actually guiding down on a sequential basis.

So you would not expect feed growth to be going from Q4 to Q1, but after that, at least in our plan, we expect to be growing at that kind of rate once we pass this seasonal sequential drop.

On your question of marketing customers, the growth that we're seeing was seen from different areas, we're seeing from retail, we're seeing from e-commerce, we're seeing from games and so forth.

Let me add to just what Herman said.

The growth on the customer category, the key from our perspective is to drive the capability of our ad platforms.

If you look at the dynamic ads, look at the different types of programs such as oCPC, we see substantial increase of budgets increase our key customers in e-commerce, in autos, in travel, in real estate categories because of their ROI increase.

Their conversions increase and their budgets expand.

So the key driver will continue to sustain because these are driven by more data, better platform capabilities and a richer ad format, more in line with ad formats.

Ultimately, it's a win for advertisers, and something that's a win for our ad platform.

Our next question comes from the line of Juan Lin from 86Research.

My question is on the partnership with mobile phone manufacturers, could you elaborate on your partnership with Huawei, I mean, which aspect of AI will be working with Huawei?

And how does your search partnership benefit you in terms of favorable apps pre installation costs.

The second question is on DuerOS.

So DuerOS is the most advanced and now key platform in China, and you have already made great progress in partnering with hardware manufacturers.

Going forward, are you planning to an increased investment in content behind the DuerOS as some of the older content such as music has already been dominated by other players.

I'm curious what is your differentiated strategy on this front?

Okay.

So let me take the second part first.

With regards to DuerOS.

It is absolutely the case, DuerOS is the leading conversational AI platform in China.

Our platform scale, our partnership size and our product innovation momentum speaks for itself.

The question about content is indeed a good one.

Music content at this stage for speaker products is a key part of the product experience.

The key thing I want to point out is the content licensing rights up to this point from content labels, rights owners, they are more distributed through the delivery vehicles, mobile apps and websites.

The content distribution rights for smart devices is a new era, and in many ways, it's up for grabs.

Think of that way.

So we are actively engaging with labels, content producers because we have leading platform.

We have more partners using our platform through devices.

And we will be able to build out over time a set of strong content rights, access to compelling content to music to offer users, to offer to our platform partners and with a good set of economics because this is a new era.

We are helping rights owners to monetize more through new type of endpoints.

So this is different from the traditional mobile app, so this is an important perspective to keep in mind when we think about content for conversational AI devices.

And then with regard to partnership with Huawei or other future OEMs, the key to have a content support is to follow.

AI is such a transformative technology wave that would elevate phone to what we call a new mobile era.

Each phone will have increased set of sensing capabilities, optical sensing, audio sensing, and also neural computing units.

Apple iPhone X and the Huawei Mate 10, all are lead indicators of the future direction of this type of new mobile devices.

Imagine the future phone will be able to see the environment, will be able to hear at the much, much grander sensitivities.

And that's a great opportunity for Baidu because Baidu's AI capability in speech recognition, in 3D visions are able to help our OEM customers to build better capabilities, more future interactions such as speech-based interactions or vision-based interactions such as AR.

And that can connect with core Baidu services such as search and feed.

So there will be opportunities to lower channel cost and also more importantly, bring our content services to a set of new interaction modalities, enrich the user engagement, and so truly that engagement will lead to future monetization opportunities.

This is the heavy caveat, this is also very early.

We're super happy to have a strong partnership with Huawei, but we fully intend Baidu will continue to build out those AI capabilities, coupled with our services, to bring to more of our users.

And this is net new opportunities because the mobile phone industry is driven by AI into what we call a new mobile era.

The phones will be able to see and hear and offer new modality.

One thing also to add, Juan, on your first question in terms of content, what we're seeing with DuerOS is not just talking to smart devices on music.

You're also seeing usage in other areas, content domains where Baidu has historical strategic advantage, things like Baidu Knows, things like Baidu Post Bar, Q&A stuff, things like video with iQIYI and so forth.

So when users are using, it's a whole wide variety of content.

A lot of that, Baidu has built through strong foundation in the past.

Our next question comes from the line of Karen Chan from Jefferies.

My question is regarding on how we should think about traffic acquisition cost trend going into 2018 as we step up our channel marketing apps with handset manufacturers?

And secondly, given the accounting treatment change, how should we think about that impacting the overall growth in operating margin?

Two questions.

First one is on TAC.

I think when you look at TAC bidding in the past, you're seeing that there's been a fierce competition.

And in the past, we focused on profit margins when we're bidding for TAC.

I think with our gross margin higher due to our scale advantage, we're considering, on an opportunistic basis, to put more emphasis on incremental profitability.

So on the one hand, we think there's an opportunity for TAC when we are focused more on profitability than on margin.

On the other hand, we believe, ultimately strategically, it's more important to grow the Baidu App, our flagship app.

So we're also focusing on growing our overall channel cost where we see a positive ROI.

With regards to ASC 606, revenue recognition or revenue accounting change, we're now going to be reporting 2018 net of value-added tax.

Obviously, when you have a lower base of revenue, when excluding VAT, our margins should be improving as a result of that revenue accounting standard change on an apples-to-apples basis.

But when we're looking at the overall margin of 2018 compared to 2017, as I said in the prepared remarks, there are 3 areas that could potentially cause our margin to be lower, and we'll have to look at that closer as we approach each quarter.

Number one is the content cost.

We think that, in addition to funding the great content, the iQIYI producers and also the ones that they license, Baijahao, BJH accounts, was starting to beef up that ecosystem.

Secondly, on the marketing cost and promotional cost, which we talked about and lastly, R&D, we're probably going to grow similar pace to 2017, potentially higher as we look for opportunities to increase our investment for AI.

Our next question comes from the line of Ming Xu from UBS.

I have 2 questions.

First is how should we think about the competition in feed business and the video business, particularly on the content side.

Secondly is how should we think about the margin of the feed business.

And how does it compare with the core search business?

This is Robin.

Let me address the question.

For the feed business, we consider it part of the total service package for the Baidu App, which means the search plus feed, a twin engine, or search for all kinds of things that interest will be reflected not only in the search results but also in the feed content.

So our feed is more personalized, more relevant to the users and have been growing very, very quickly, as you can tell from our financial results as well as our traffic results from third-party research firms.

And for video, iQIYI continues to lead in the market.

We are #1 in terms of DAU.

We are #1 in terms of average time spent.

We are #1 in terms of number of paying subscribers, and we are #1 in terms of profitability.

Although it's not profitable yet, but we lost a lot less than the competition.

So it continues to do very well.

And in terms of margin, at this point, the margin for feed is lower than search, but longer term, I think the margin for feed will be very close to search because it's a very, very fragmented on the supply side.

So unlike long-form video, we have lots of choices to choose from.

And the platform has a lot of power to direct traffic.

The key competency for the feed business is how can you efficiently distribute content instead of what kind of content you have.

And the short-form video or text or images, those kind of content, the price is not that high.

The key value add is how you find the right users for the right content.

Our next question comes from the line of Alex Yao from JPMorgan.

Appreciate your color on the feed ad growth outlook into 2018.

Can you give us at least some preliminary thought on how you think about the core search will be growing in 2018, especially in the backdrop of a slowing down industry traffic, particularly on the mobile side?

What's the strategy to drive core search revenue growth in 2018?

I think couple of things.

Qi mentioned earlier how we're growing our core search.

I think what you're seeing is a lot of the Baidu technology, such as oCPC, such as dynamic ads leveraging data.

I think that, on the one hand, you're seeing a lot of new feature sets that we're doing, that we're allowing the users to come on our search in addition to be able to do transactions and so forth.

So that's on the one hand.

On the other hand, what you're also seeing is us focusing on our flagship Baidu App so that as we're getting more users on the app themselves, they'll maybe come to look at our feeds, and sometimes, they'll maybe come to do searches.

So you're seeing that as a combine.

So I think, from an advertiser perspective, also from a user perspective, they benefit the 2 at the same time.

So you're probably going to see more and more of those 2. If one grows, it's probably going to pull the other because it's all wrapped up in one container.

So let me just add a little bit to what Herman said because search is a fundamental service for human needs, and there's longstanding growth perspective.

There's no question.

But when you think about -- for search product, it's all about driving long-term growth of users and usages and more content.

There's 2 fronts.

One is we're getting more and better and richer content into search.

We talk of videos and strong growth of high-quality videos.

And we are doing a lot of investment to steadily grow better quality, richer content as part of search.

The other is enable search to be done more with different type of modalities.

There's keywords-based modality search, but there's a lot of growth potentials for voice, for image-based search capabilities.

That's another front of search user and usage growth.

And there's monetization growth.

We talked about dynamic ads, talked about oCPCs.

For each of the key search verticals, we believe there's opportunities to go deeper to harness better yields.

And the last thing I want to emphasize, what Herman also said, we're going to strongly grow Baidu App because we think Baidu App will have better opportunities to deliver a highly richer, differentiated search experiences and for Baidu App, we have zero TAC.

So it's all our traffic, all our monetizations.

So there is these pivots that driving long-term growth, that emphasizes long-term user and usage growth and long-term profitability growth.

Our next question comes from the line of Thomas Chong from Crédit Suisse.

I have a question on Internet finance initiative.

Given the facts that we see this business line is growing at a very fast momentum, can management highlight about the regulatory headwinds and the time line for profitability?

You're probably referring to the regulatory headwinds that started in December, where the government are now banning products that are over 36% APR.

With regards to that specific policy, for Baidu, our financial services, we leverage Baidu's data.

So because we're leveraging a huge set of data, we're leveraging our AI computing power, we're able to actually be able to compete at a more efficient level, and we're able to offer products that are sub-36% APR.

With that said, obviously, with the overall regulatory environment, trying to get funding from banks, it has been harder since the announcement in December.

So in that sense, we are impacted from that.

With regards to profitability, we're still at a loss, but we're seeing the trend to be pretty positive.

If you look at it sequentially, for the last few quarters, our negative margin is getting less and less.

So we're pretty bullish on the future of our financial services.

Our next question comes from the line of Jin Yoon from Mizuho.

Couple of questions regarding content.

I think, Herman, on your prepared remarks, you said that content cost should increase by the same rate in '18 versus '17.

I apologize if I missed this.

But is that on a percentage basis or in absolute terms?

And in Q1, should we see the greatest -- or the most significant quarter-over-quarter jump because of the revenue push out in iQIYI and the sequential jumps for the rest of the year should be a little bit more stable?

Is that how we should look at it?

With regards to content, yes, for 2018, I would expect our content cost to grow at least at a rate that we saw from 2016 to 2017, number one.

And with regards to sequential content cost, yes, I would expect Q1 to be a little bit higher than Q4 because Q4, as you recall, iQIYI was impacted by certain events at the beginning of October.

As we're going into Q1, we no longer have that kind of content restriction environment.

So as a result, you'll be able to probably see a better advertising, and we will probably have to have similar content costs.

Herman, if I could just quickly follow up.

That jump that you're talking about, is that on a percentage basis or an absolute dollar jump on the '16 to '17 and '17 to '18?

Percentage basis.

Our next question comes from the line of Natalie Wu from CICC.

My question's regarding the search and feed ads.

What's the current standing from KA versus SME, I mean, the contribution wise?

And what's the trend going forward should we be looking at?

Historically, we have not disclosed the exact percentage contribution between KA and SMEs.

But a clear trend is that revenue contribution from KA or larger customers have been growing faster.

That's in part because of our rich data and coupled with our AI technology.

And also increasingly our customers, online customers are more willing to open up their data or exchange their data with our user data so that we can leverage the data on both sides to optimize conversion using our AI technology.

That resulted in faster growth of the KA revenue.

Robin, just wondering is there any difference that you have witnessed between the search and feed ad products regarding the KA and SME spending preference?

Yes.

There are overlaps between search and the feed customers, and there are also differences.

Some of the advertisers, they would prefer a more specific request of using search queries.

Others would probably target more certain specific demographic group, for example, aged between 20 to 25 or resides in third tier cities, things like that.

We use the feed to better target those kinds of advertisers.

Real estate, travel, these are all suitable for customers for feed.

We are now approaching the end of the conference call.

Thank you for your participation in today's conference.

You may now disconnect.

Good day.