Gaotu Techedu Inc. (GOTU) Q2 2019 Earnings Report, Transcript and Summary

Ladies and gentlemen, thank you for standing by, and welcome to the GSX Techedu Second Quarter 2019 Earnings Conference Call. [Operator Instructions] Please note this event is being recorded on Thursday, the 22nd of August of 2019. I would now like to hand the conference over to your first speaker today, Mr. Tip Fleming from Christensen. Thank you. Please go ahead.

Thank you, operator. Hello, everyone, and thank you for joining us today. GSX' earnings release was distributed earlier today and is available on the company's IR website. On the call with me today are Mr. Larry Chen, GSX' Founder, Chairman and Chief Executive Officer; and Ms. Shannon Shen, Chief Financial Officer. Larry will give a general overview and then Shannon will discuss the financials. Shannon will be available to answer your questions during the Q&A session that follows. Before we begin, I'd like to remind you that this conference call contains forward-looking statements as defined in the U.S. Private Securities Litigation Reform Act of 1995. These forward-looking statements are based upon management's current expectations and current market and operating conditions and related to events that involve known or unknown risks, uncertainties and other factors, all of which are difficult to predict, and many of which are beyond the company's control, which may cause the company's actual results, performance or achievements to differ materially. Further information regarding these and other risks, uncertainties or factors is included in the company's filings with the SEC. The company does not undertake any obligation to update any forward-looking statement except as required under applicable law. It is now my pleasure to introduce Larry. Larry, please go ahead.

Thank you all for joining us for our second quarter earnings call and our first call as a public company. We are pleased to report another profitable quarter of growth, with net revenues of RMB 353.7 million, a 413.4% increase on a year-over-year basis. The strong growth was propelled by robust performance across all of our business lines. Our K-12 courses, the cornerstone of our company, continued its rapid expansion, which speaks to the popularity of our services and the tremendous potential of this industry. According to Frost & Sullivan, the K-12 student population is expected to reach 241 million by 2023. 45% of these students are expected to attend online after-school tutoring courses, which is up from just 9.5% in 2018. As the industry leader, we are confident that we are uniquely positioned to reap the benefits of this fast-growing market. Amid a sea of industries in China currently with a strong macroeconomic headwind, online after-school tutoring is one of the few islands enjoying strongly favorable condition, including governmental policies. The recently announced opinions on regulated online after-school tutoring allowed for greater operational flexibility for online service providers compared with their off-line counterparts. This not only gave the parents and students more room to make personal choices, but we think it also will strengthen the competitive advantage of industry leaders such as ourselves. We relentlessly pursue the high end of quality with our course offerings. This remains the overarching focus of our growth strategy, which has been reinforced by our intense brand recognition as an NYSE-listed public company. Delivering results-oriented, high-quality education to our students will continue to be at the heart of everything we do and will serve as a key to deepening our economic moat for sustained, profitable growth. Now I will hand the call over to Shannon, our CFO, to walk you guys through the details of our financial and operating results.

Thanks, Larry, and thank you, everyone, for joining the call. I will now walk you through our operating and financial results. Please note that all financial data I talk about will be presented in RMB terms. We entered 2019 with robust growth in the first quarter, and I am pleased to report that with our continued momentum across all our key operating metrics, which drove the strong financial performance during the second quarter. We are still laser focused on online, live, large class, which we believe is the most effective and scalable model to disseminate the scarce, high-quality teaching resources in China to aspiring students all over the country. Net revenues in the second quarter jumped 413% from the same period of 2018. Our gross billings, which is the metric we pay close attention to, increased by 462% year-over-year to CNY 599.4 million, up from CNY 106.6 million in the same period of 2018. Total enrollment, which refers to the enrollment are paying more -- paying more than RMB 9.9, reached a historical high, hitting the number of 592,000, 3.5x that of the same period of 2018. Period enrollments, which refers to enrollments for regular classes, increased to 363,000 from 94,000 in the second quarter. The across-the-board growth was primarily driven by the continued expansion of our K-12 after-school tutoring business, and to a lesser extent, our foreign language, professional and interest courses. Now let's break down our revenue stream. Net revenue from our K-12 courses increased by 463.7% to CNY 270.3 million and accounted for 76.4% of net revenues, increasing from 69.6% in the same period of 2018. This revenue growth was primarily driven by increase in paid course enrollments and K-12 students tuition fees. Our standard tuition fees for our K-12 courses increased an average of over 20% year-over-year. Another reason for the increase in ASP is that we introduced the new courses that have higher standard tuition fees. For example, we launched a critical thinking course in late 2018 that has been well received by parents so far. Gross billings contributed by K-12 after-school tutoring business were CNY 504 million compared with CNY 79 million in the second quarter in 2018. Paid course enrollments increased by 372% to 321,000, up from approximately 68,000 for the same period of 2018, demonstrating our superior education quality recognized by parents and students. The ramp-up in paid course enrollments is also a result of our high retention rate, evidenced by the reregister of 3-semester students in sequential semesters. Average gross billings for paid course enrollment increased from 1,100 in the second quarter of 2018 to 1,500 in the second quarter of 2019. Net revenue from our foreign language professional and interest courses accounted for 21.8% of net revenue, increasing from 17.4% in the same period of 2018. Net revenue increased by 542.7% to CNY 77 million, up from CNY 12 million for the same period of 2018. The significant year-over-year increase was primarily because we are able to raise tuition fees by optimizing our course catalog and promoting highly qualified teachers, all of which helped to increase paid course enrollments. For example, we were able to increase our standard tuition fee for foreign language courses by an average of over 40% and for special interest courses by over 30%. CNY 85.3 million in gross billings were generated by our foreign language professional and interest courses compared to CNY 20.7 million in the second quarter in 2018. Paid course enrollments were 42,000 in Q2 '19 compared to 26,000 in Q2 '18. Leveraging our know-how with online, live, large-class education, we will further expand into this large industry segment. Our core business has grown regularly as we continually improve retention and take advantage of word-of-mouth referrals, which also benefit from the strong and rising demand for online education. At the same time, we've managed to achieve and return profitable growth by improving our ability to control costs and operating expenses. Our cost of revenues increased by 281% to CNY 101.2 million, up from CNY 26.6 million year-over-year. This increase was primarily due to our increased recruitment of teaching staff, including instructors and tutors. As we expand our business operations, we expect our cost of revenue to increase in absolute amount in the foreseeable future as we fill more students and offer more courses. Non-GAAP gross profit margin, which exclude share-based compensation, increased to 72.6%, up from 61.5% in the same period of 2018. We operate a highly scalable business model. Teaching staff compensation, one of the largest cost add-ons, is declining as a percentage of net revenue primarily due to economics of scale. Average enrollments per class further increased from 980 in the first quarter in 2019 to around 1,200 in the second quarter. We are able to pay our teachers incrementally more while still enjoy greater operating leverage. The competitive compensation we provide to teaching staff, a by-product of our scalable business model, contributes to high teaching staff retention rate, which will benefit our students in the end. Selling expenses increased to CNY 169 million, up from 18% -- CNY 18.4 million in the second quarter of 2018. The increase was primarily a result of more marketing expenses spent to extend market share to attract new students, especially for summer campaign and for brand enhancement. Research and development expenses increased by 164.2% to CNY 41.1 million. We constantly work on ways to apply the latest technology to improve learning experience. This includes the efficiency and effectiveness of our teaching for data delivery as well as operational efficiency. We will continue to enhance our proprietary technology infrastructure, upgrade our system, optimize our IT tools and software, recruit talented personnel, and introduce and apply practical artificial intelligence technology. We also devoted more resources and plan to increase spending as we continue to develop our education content. Given that our revenue growth rate outpaces that of revenues net of R&D expenses, we believe we can still expand operating leverage despite the incremental investment in research and development. G&A expenses increased by 194.3% to CNY 26.1 million, mainly due to an increase in G&A head count and an increase in related compensation. Income from operations increased to CNY 16.2 million from loss from operations of CNY 0.5 million in the same period of 2018. Non-GAAP income from operation, which excludes share-based compensation, increased to CNY 31.1 million from CNY 0.2 million in the same period of 2018. Operating margin, defined as income from operations as a percentage of net revenues, rose 4.6%. Non-GAAP operating margin increased to 8.8% from 0.2% in the same period of 2018. We would like to highlight that operating margin shows distinct seasonality for K-12 after-school tutoring business, especially for online, live, large-class business model. We offer courses in 4 semesters, namely the 2 school semesters in spring and fall and 2 holiday semesters in summer and winter. The semesters do not perfectly match quarters on calendar. The revenue generated by winter and spring semester would swing substantially between first quarter and second quarter due to the floating schedule of spring festival. For instance, certain amount of net revenues moved from the second to the first quarter due to the early schedule of spring festival in 2019 compared to 2018, which results in a faster revenue growth rate in the first quarter of 2019. In the second quarter, spring semester usually ends in the first week of June, which means that nearly 1/3 of the second quarter barely contributes to net revenues. At the same time, summer marketing campaign acquires higher investment in sales and marketing expenses. Those may lead to a lower level of operating margin in the second quarter compared with the full year margin. The same trend also applies to operating margin in the third quarter. In the summer semester, we provide very low promotional pricing to students in certain grade, especially at the start of a new level of school, for example, first grade, the seventh grade and the 10th grade. Revenue contributions from these promotion classes in summer semester is negligible, while summer marketing campaign continually lead to higher sales and marketing expenses. The fourth quarter will contribute the largest portion of gross billings, revenue and operating income during the year as fall semester starts with regular priced classes and the marketing activities become less intense. Thanks to our strong organizational capability and operational efficiency, we have been profitable for 5 consecutive quarters from the non-GAAP perspective since the second quarter of 2018. We are one of the few, if not the only, leading players in the market that have achieved sustained profitability. In the future, we will continue to execute our pricing strategy, well-proven market strategy and provide students with the best-in-class learning experience. Our basic and diluted net income per ADS were both CNY 0.04 in the second quarter. Non-GAAP basic and diluted net income per ADS, which excluded share-based compensation, were CNY 0.11 and CNY 0.10, respectively. Net operating cash flow for second quarter of 2019 was approximately CNY 193.8 million. Now let's take a quick look at the key financials of balance sheet. As of June 30, 2019, we had CNY 60.1 million of cash and cash equivalents and CNY 1.7 billion of short-term investments. As of June 30, 2019, our deferred revenue balance was CNY 503.7 million. Deferred revenue primarily consists of the tuition collected in advance. Lastly, I also like to take the moment to introduce progresses we have made to comply with online after-school tutoring regulation. On July 15, 2019, the China Ministry of Education, along with 5 other ministries and administration, published opinions in regulating online after-school tutoring. The detailed regulations are broadly in line with our expectations. As Larry just mentioned, we view the regulation as beneficial to the online education industry, particularly leading players including ourselves. In terms of licensing, we are compliant with the ICP and telecommunications business operating license. In terms of teachers' qualifications, nearly 80% of our teachers are already qualified or have passed both written and the oral tests and are waiting for the registration. We have a robust pipeline of qualified teachers. In case certain teachers do not get their certificates before the grace period is over, we will be able to seamlessly substitute such teachers with minimal level of disruption to our students. In terms of tutoring time [indiscernible], we are already compliant as all of our K-9 classes' end by 9 p.m. and, hence, we expect minimal changes to our K-9 operations in this regard. In addition, it is worth noting that the new regulation limits off-line access to tutoring to 8:30 p.m., whereas online classes can end at 9:00 p.m., further underlying the benefits of our model compared to our transitional off-line peers. Also in terms of homework, off-line tutors are not allowed to assign any homework at all, whereas in online channel, there are only restrictions for the first and second grade. We believe our current curriculum and teaching plans do not require any significant changes. In terms of the 3 months or 6-day classes limit for upfront payment, we believe this regulatory stance provides greater clarity and consumer protections for both parents and students. It will motivate us to devise service and products of higher quality. We have promised to provide full refunds to parents and students at any time before the live class is over since we start this business. Considering our low refund rate and high retention rates, the reduction in payment period will have limited impact on us. Particularly, we believe we are more capable of guaranteeing the safety of prepaid tuition fees compared with a number of our peers, thanks to our profitable operations and the increasing brand recognition. As one of the leaders in the online education space, we believe we are among the least impact by the largest -- by the latest regulation as we have been conducting our business in line with these policy aims, and believe that these policies will, in fact, further enhance our leadership position. We will continue to monitor the development of online education in China to continue to provide our shareholders with responsible guidance and effective execution and to continue to deliver strong financial and operational results over the long term. With that, I now provide our business outlook. Based on our current estimate, net revenue for the third quarter of 2019 are expected to be between RMB 486 million and RMB 506 million, representing a projected increase of 390.9% to 411.1% on a year-over-year basis. These estimates reflects the company's current expectation, which are subject to change. That concludes my prepared remarks. Operator, we are now ready to take questions. Thanks.

[Operator Instructions] Our first question today comes from Alex Yu with Crédit Suisse. Alexander Yu;Credit Suisse;Analyst: So I have 3 questions to ask. First one, I'd like to ask about our customer acquisition channels. How much of our new student enrollments were from external channels? And how much from our organic channels? And for external channels, what are major channels that we used and our expectations for the customer acquisition costs in the second half? On the second aspect, I would like to ask about our revenue and also enrollment breakdown between the primary, middle school and high school departments. And what strategies do we have to do better in the primary school segment? And the third aspect I would like to ask about is the contribution from low-tier cities versus Tier 1 and 2 cities. What are the major provinces or cities that contributes revenue of us?

Thanks, Alex. Your first question is about the customer acquisition. So actually, most of the new student enrollments we have acquired are through traffic acquisition. And the main channels we are using is basically the same as other online education companies, such as WeChat, like WeChat Official Accounts, WeChat Moments, news feed ad, and also with TikTok, Toutiao and Baidu. And also we are exploring new channels as well. And so talking about like the second half year's sales and also the marketing expenses budget. So it's -- this is worth noting that, also there is some seasonality increase for our business. Like in summer campaign, summer campaign usually starts in May, and it lasts through May until the end of August. So when we talk about the sales and marketing strategies in the second half, so we are talking about like the activities we are taking in the summer. And like I just mentioned in my prepared remarks, like the sales and marketing expenses we are... [Technical Difficulty]

Okay. You may continue.

Also let me recap with our marketing strategy. So let me talk about marketing strategy. Within our organization, there is a phrase we use frequently. We call it effective growth. So we don't usually see marketing as a expense. We actually see it as an investment. So one of our major advantage is our high efficiency at our customer acquisition. The student -- the customer acquisition efforts while -- like the value chain is very long. So it's less from the traffic acquisition at the beginning and then it's the free promotion class we provide. And also follow that is the sales products that we have a really strong sales force that can help convert the parents from a promotion class to a regular class. So since the value is really chain, that company needs to be equally capable on each of the links that to make the whole sales chain more efficient. And so we have been trained our staff in the past 2 to 3 years that they -- we train them to use money wisely because we haven't raised any money since like in 2 -- since 4 years ago. Then -- so then for example, if we can be 5% better growth in the space than others, then maybe we can deliver an acquisition efficiency twice or even several times as high as our peers. So we actually -- and we believe we developed a really efficient technology to closely monitor our customer acquisition process, so they calculated ROI closely that the system we developed is very efficient. By doing the peak season of our traffic distribution, they can monitor the traffic on hourly basis. So we monitor our links and the barcode we send out through all our sales channels. And we can obviously mine those channels if we do see like they are being slow or there's a latency when the customer once registered the classes. And also when we talk about the seasonality of our sales and our marketing expenses, in the -- from July to August, we are still invest in the market to acquire new students. And usually, in the fourth quarter, and especially when the fall semester starts, there will be a live marketing activities there. But since -- you may know that this year, the competition in the summer campaign kind of like heated up a little bit, so we need to react quickly to what's happening on the market. And that's what we have been training for the past 2 years. And talking about the budget for the second half of the sales and marketing budget, we will adopt what's going on at the market and quickly act to what's going on in the market. So I think that is your question. Then your second question is about the enrollment split from different segments, and how to do better in the elementary school. So in this quarter, around 25% of our paid course enrollments are from elementary school, 30% are from middle school and about 45% are from -- were from high school. So in terms of student number, elementary and middle school are catching up quickly. It is -- sometimes, it's because in high school, students tend to buy multiple subjects that require the cross-selling, that they are contributing more enrollments. And elementary and middle school increases faster than average, especially for elementary. So we are having a better performance in this market sector, and we are expecting to see this positive trend to continue in the following time. We believe that elementary school students are vital to our business for they are the most fundamental target customers in K-12 market. Also they are the biggest market because there are 6 grades in elementary. The total -- it's like the total of middle school and high school, and they are likely to contribute much more to our business as they have a potential larger lifetime value, and we can really build a brand image in them. So -- and also we put also elementary courses development as a strategically high priority. The head of our elementary school department has intensive experience in education and used to manage over 500 employees. We also put a lot of energy on recruiting capable teachers. The instructors are either from the top 2 schools, like Tsinghua University and the Beijing -- Beida, Tsinghua and Beida or other top 10 universities. That or -- we also recruit teachers with intensive teaching experience, like around 2 to 3 years of teaching experience. Further, that we are standardizing the curriculum, centralize the content development and building like revisions, reviews to make the curriculum more fun of the class. And our teaching philosophy is to ignite the interest. We also -- I know like in the industry, a lot of the companies are using -- like they are recruiting more fresh college students or like from all those prestigious universities and provide training courses. After that, these teachers can deliver a really good class. But from our perspective, even for children at a younger age, they still need to follow those teachers that have strong confidence, that can ignite their interest in the courses and also build confidence in the subject. So we do have the confidence that we can build -- like our elementary school and middle school can contribute more to revenue and to our business. And your third question is about the cities and the provinces we are covering. So actually, the enrollments are quite evenly distributed across tiers of cities. For the first tier city and the new first-tier city and second-tier cities, it's basically contributed about 50% of the enrollment. And for the lower cities, they are contributed another half. And actually, we did a really good job on penetrating lower-tier cities. So if you do a comparison between 2018 and the first quarter in 2019 and the second quarter, so the first quarter in 2019, actually, the penetration rate -- the enrollments contributed from the third tier cities or the fourth tier cities and the fifth tier cities increased about 5% in total. And in this quarter, we do see a 3% quarter-over-quarter increase in the third tier city and the fourth-tier cities. So which means we basically we did a really good job on penetrating lower-tier cities. Also like I want to -- I also want to remind you that based on the enrollment number we have right now compared to the 200 million students out there, it's still a really small number. And it may not have -- to take away meaning for that how we are doing on low-tier cities. But we do see different demand from the first-tier cities and the lower-tier cities. For instance, for the first-tier cities and the new first-tier cities, parents are more sophisticated ways in those classes, so which means our teachers, including our field staff, needs to be very familiar with the product they are promoting. And for lower-tier cities, a lot of parents may have never got a chance to see so many good teachers. So they are very thrilled that they have this chance or this exposure to the good teachers. So these are the different demands from first tier -- from the higher-tier cities and the lower-tier cities. And talking about the distribution on provinces, it basically, it consists with the population distribution. Like there are top provinces we are carrying, like Guangzhou and Shandong, Zhejiang, Jiangsu and also like Hunan, Hubei. These are all the provinces that have the largest student base. I think that covers your questions. Thanks, Alex.

The next question today comes from Gregory Zhao with Barclays.

A very strong quarter, and congratulation on the IPO. So this is your first earning call after the IPO. So I want to ask 3 questions, if I may. The first one is about the big picture of the industry. So we know the market has very huge opportunities considering the potential of the user base and addressable market. So is there any way you can help us to quantify where the market opportunity, such as the potential to penetration rate and ARPU upside? So this is my first question. The second one is I want to understand the competitive landscape and your user acquisition strategy. So we heard your competitor also very aggressively are expanding into the online space even they have some traditional some lapsed issue in the off-line. And we know they invest a lot in online acquisition. So just want to know your thoughts on that front, how to compete with them. And also how do you differentiate your course content from your peers? The last one is a quick follow-up on the higher-tier city and lower-tier city questions. So you just reported very impressive penetration rate in the lower-tier city. As your courses are given online, so I assume that the prices across all the different tier cities should be the same. I just want to confirm that.

Thanks, Greg. And yes. Your first question is about the quantitative market opportunity. So K-12 is actually a tremendous market. Like it's children-level market has a huge opportunity. Like for instance, in 2018, the online market penetration rate was only 9.5%. So in 5 years, it will reach to 45.5% based on Frost & Sullivan's report. So -- but we are in the industry. Actually, we do see the penetration rate is much more higher than what then, like the outsider has observed. So let's just do a quick math. For instance, it's like a trillion-level market, then in 5 years, 60% is online, then within the online business, probably 60% is for large-class business. Then that already gave you like a 300 million market. It's really huge. Then let me talk about online live large class. It has a higher entry barrier than other like players because like all the technology infrastructures, you need to have like the live broadcasting technology, the big data and the AI technology. And also you need to have a large group of people to serve such a large group of students. So that entry barrier that kind of like decides, like the top players will have a higher market kind of concentration rate. So let's say, if the top 3 players hold that 50% of the market share, then that already gives you like an annual revenue of -- net revenue of about CNY 20 billion. So that's a really huge market opportunity for us and for like other players right now. And the deep thinking behind why we are doing the online tutoring in large-class format is that think about like what's the real demand from the students and the parents? They are always searching for the best teachers. And if they look at a teacher's supply chain, the online teachers, they can now be the best. Like college students can do like one-on-one tutoring. And if we can find some like graduates has -- have good potentials like we train them up, they can be a very good model class teachers. But think about like only the top-notch teachers can deliver the teaching content in a large-class format, which means the large-class teachers are always the best teachers. If 1 day, all the students and the parents know that the best teachers are all in online and they are all in large-class format, they will follow the teacher. That's why we are quite confident that large -- online-large-live class has a huge and tremendous opportunity there. So then your second question is about the competitive landscape. So the competition was kind of like heated up in the summer, especially for this year. We all heard the news that top players in the industry, they kind of like invested over 1 -- over RMB 10 million per day since June. And we did see like different dynamics this year. First, like market penetration rate is much higher and it grows even faster than they expected. If we put together the enrollments across all of the online platforms, the number added up to a big value real quick. At early stage, it's a really good thing to have multiple players to educate the market together. We see promotions from both online and off-line like posters and videos from subways, bus station and television. And we do see more and more parents are getting to know online education, especially for lower-tier cities. And the second scenario is -- and second is -- and then also a very interesting scenario is that parents were signing off on multiple platforms. In a larger extent, this is a -- this is kind of like a new and it's different from last year. We did this way with our parents. A lot of them are paying for promotion classes on 2 to 3 platforms or even parents, they are signing for 4 companies. One reason is that the promotion classes are not that costly. And the other reason maybe, perhaps, offering on social platforms. Besides that, education companies have similar exposure to the same group of parents. And this way, the parents cannot -- can be more selective than ever before. That's what's happening in the summer. So then -- which means like talking about the competitive advantage that a company like the teaching product we are providing needs to be excellent, needs to be -- excel. It cannot be just average or slightly above average. When the parents have become more picky, you need to -- we need to grab the -- their attention at the first place and let them know that we are kind of like providing the best teaching products in the market. So that's why we spend a lot of time on training our teachers, help them transitioning -- help them transition their teaching model from off-line to online or from their old model to the new model. So we also have our teachers to adopt a lot of new like teaching skills. Like they need to adopt those online live broadcasting skills to attract the students in a higher level. That's the like competitive landscape right now. And there is a phrase that we usually says, like the start of our strategy should not be our competitor, it should always be our customer. And that's why we put customer achievement as the first sentence in our corporate value. We believe if we can continually provide our customer with the best teaching quality, that our customer can follow us and they will keep -- retain with our platform. I think your third question is about course content development. So the course content development for large-class teaching format is kind of like different from those centralized small-class content development process. The course content development should be from upside down that because we may hire the teachers with intensive experience in the education industry, we are leveraging other skills, other experience they have that has been proven in their past experience. So the way we are doing our content development right now is like all of our instructors are part of the content development process. So they can -- they do group discussion and they routinely organize those meetings to discuss -- to come up with the most efficient teaching method. So we can centralize -- we can standardize them. And we also have a support team to support those instructors, like they are doing the PowerPoint for the teachers. They are doing [ variations, radios ] to help teachers to deliver the licensing in more efficient way. And also talking about like the teacher training, we are building up a strong like instructor pipeline like the -- they kind of like -- we are building the ladders like we have those -- top teachers has 10 -- over 15 years of teaching experience. Then also we have younger teachers that they can constantly learn from their role models. And your next question is from -- like it's much like the pricing strategy, I think, like the price we charge with our customers from different tiers of cities. So off-line, advertising strategy for off-line is kind of like different from online. Because when we talk about off-line business, they have different operational costs, first-tier cities and lower-tier cities. Like in first-tier cities, like the rental is higher and the teaching compensation is higher, but that costs are lower in lower-tier cities. So first, for us, since we are doing business online, the costs are basically the same for all of our customers. But we are developing a variety of products that can adapt to lower-tier cities maybe with a lower price that can be more affordable to those parents. I think that concludes your question.

The next question today comes from Jeffrey Chan with CLSA.

I have three questions. The first one, it's just a recap. Can management help me to recap a bit on the gross billings per paid course enrollment by category, which means by K-12 and by the interest and foreign language course? And my second question is, what is the subject exposure now? Is it math still accounting for the highest proportion? And how about English and Chinese? And my last question is, is there any specific strategy for management in recruiting new instructors given the higher competition landscape?

Sorry, Jeffrey. Can you repeat your last question?

My last question was, is there any like specific strategy in recruiting new instructors given the higher competition landscape?

Okay. So your first question is about the average spending for enrollments. So I also want to remind you that, that also has a seasonality. For instance, yes, in the second quarter, like a lot of students that are registering for like summer class and for part of the fall class, then some of the spend, it leads down to what we called entrance-grade level, like the first grade in elementary or the first grade in middle school, the price of the class is lower. It's basically only 90 -- CNY 49,000 per enrollment. So that way, we kind of like track down the average spending per enrollment. But if we talk about -- if we move the scenario to the first quarter, then in the first quarter, students are enrolling for a spring semester. Usually, the average class price, the ASP is higher. So that gives you a higher average spending per enrollment. So that's -- if we compare quarter-over-quarter, the average spending per enrollment doesn't change a lot. This kind of like write-offs the impact on the ASP increase. It's about CNY 1,500. And your second question is about the subject exposure. Yes. So in this quarter, math, K-12, English and Chinese combined contributed around like 80% of our paid course enrollment of our K-12 courses. Math do contributes the highest proportion. It's around -- it's about 30%. But we did really good job on English and Chinese as well. These 2 subjects contributed about 40% as well. So that's another advantage we have that it has -- we have. Our revenue distribution on these 3 major subjects are kind of like very, very new ones. So that -- which means we have great, good teachers that are evenly distributed on those subjects, which allows the students to have more alternatives. And they can do a better job with the cross-selling. And that's another reason why we can achieve our profit volume in the long run. And also like, we can see a decreasing trend of math proportion and an increasing trend of English enrollment. So in the future, the paid course enrollment I expect to spread more evenly over these 3 subjects. And also I want to let you know that we also did a really good job on those we call the smaller subjects like physics and chemistry. Because sometimes, even for those off-line institutions and for other online institutions, it's really hard to find a good teacher to cover these subjects. So a lot of students when they cannot find a good teacher in physics and in chemistry, they are -- they tend to search online for those teachers. So that's why we have exposure to them. And once they register with us, we also recommend like other like math and chemistry to them. So that's why we have a really cross-selling rates in our high school sector. So I think that's your second question. And also we talked about the strategy of recruiting our instructors. Yes. So we do see like as a education company, it is always the ability to recruit, to train and to motivate the teacher. That is the core ability. And for us, we have a really high bar to recruit instructors. In 2018, the interview acceptance rate was lower than 2, which means, within 100 interviewees, we will only accept less than 2. Only less than 2 can pass our interviews. And so actually, we -- really, we don't see the competitive landscape has like too much impact on our teaching recruitment process, thanks to our public initial offering in the United -- in the NYSE. So that, we did a really good job on promoting our brand awareness, which makes it much more easier for us to recruit new teachers right now. And in the past, we kind of like we need to persuade the teachers that we are a really good company, that they have really high operating efficiency. But right now, with all the transparency, with all the like public data, we can provide to the publicity that allows teachers -- a lot of teachers, they come to us and -- because they see like how efficient we are running our business. And also I wanted to address that our teachers has a really -- our instructors has a really low -- has a really high retention rate and a really low turnover rate. In 2018, only 9 of our instructors actually left our company. Two of them were public school teachers because they are not allowed to teach online anymore. And the rest, the 7, it's mainly because like when we monitor their operating data like the sales conversion rate and the retention rate, their data is not that satisfying. So we convert them into our content development team. So we did a really good job on training our teachers as well. We had a teacher -- a teaching quality supervision team that are delegated to do that job. And we are probably the only company in this [ down ] world that we do over 100% supervision of all our online live large classes. And our teaching quality supervision team where we generated a supervision report in 24 hours. And we request all those teachers to respond to those issues in 24 hours. And sometimes, the report is also copied to our CEO, just like -- just want you to know like how much attention we pay to those teachers and we -- and all the efforts we put in keeping improving our teaching quality. I think that answers your questions.

Yes. This is very helpful and congratulations on the outstanding results.

Thanks.

The next question today comes from Sheng Zhong with Morgan Stanley.

I just have one question about the teachers. So until now in the K-12 segment, how many instructors and tutors do you have? And for your top teachers, what's the current student-teacher ratio? And because it's -- continues to increase. So when you monitor this along with the ratio increase, do you see any change in the key metrics that you see, sales conversion rate, retention rate? And also how much potential room do you see can improve the student-teacher ratio?

Thanks, Zhong Sheng. So your first question is about the number of our tutors and the instructors. So by the end of the first quarter, we have 169 instructors. And by the end of the second quarter, that number adds up to less than 200. And by the end of the first quarter, we have 552 tutors. And by the end of the second quarter, the number adds up to around -- over -- slightly over -- [ less than 1,000 ]. So our organization actually grows really quick. So that's the privilege we can enjoy because we have a really strong management team that Larry used to manage over 30,000 employees. So it's kind of like we are going into his comfort zone that when our organization became a really large one, that he knows exactly how to train up these tutors that they can provide a better service to our parents and to our students. So your second question is about the teachers and student ratio. So we have 2 ratios. One ratio is for instructors and student ratio and the other ratio is the tutor and student ratio. So for the instructor and student ratio, I've just mentioned that the average enrollment per class during my prepared remarks. And we did pay close attention to the class scale. We monitor class scale at 600, 800, 1,000, 1,200 just to keep a close attention to -- when the class scale becomes larger, wider, the -- like the ongoing KPIs will become worse. So we closely monitor a lot of ongoing KPIs, such as class attendance, class completion rate, homework submission rate, homework correction rate. And from our observation, the ongoing KPIs is largely depend on the teachers' teaching quality. Sometimes, the teacher is quite capable, even though his class has over like 2,000 students in. We can still see these numbers are getting better and better. But for some younger teachers or teacher in early stage or they did not take enough training, their numbers may become a little bit worse. I won't be using the word worse. We just don't see the number going up probably with the class scale of 600 or 800. So for those kind of teachers, we will providing additional training to them and closely monitor how these ongoing KPIs goes. And talking about the tutor and student ratios, like currently, in the second quarter, the average students a tutor is taking in the same time period is less than 300, it's about 270. So from our view, it's still very important to keep the intimacy between students and the tutors. The reason why we are doing it in the tutor-teacher class -- classes, it's like from one side, we've solved the demand that we are providing the best instructors, we are providing the best teaching quality. And from the other side, we need also to provide the best service and provide the emotional support as well as the intimacy that off-line institutions usually can do better. So that's why we pay close attention to recruit our tutors as well. We do see them as a valuable asset to our company. So we wanted to keep the ratio at certain extent. And also a lot of the investors may ask like, is 300 a little bit too much? But leveraging all the technology we have right now, like the apps, the homework grading system, like leveraging all the graphic recognition, voice recognition, now, right now, our tutor can fix a homework in 50 seconds. That saves them a lot of time to do all those human contacts, all those emotional connects with our parents. And your last question will be the retention rate. So -- yes. So the retention rate we observed from the second quarter is really impressive from our perspective. So for the past 5 quarters, we have been able to improve our retention rate on quarterly basis. And we are quite confident that we did a really good job on the high school sector and for middle school and primary. As mentioned, with the sector, we are improving the numbers really quickly. And we didn't disclose the retention rate, it's because like we want the other KPIs to be the really simple ones like -- hey, retention rate is we call the tune-out KPI in our company. We use the retention rate to evaluate instructors' performance, to evaluate tutors' performance. And we have a really conservative calculation of the retention rate. The retention rate in our company is only calculated by 1 course, the same instructor and the same tutors by the students in a class and they're registering for the subsequential semester. So which means if a student, they switched class from 1 instructor to another or switch from 1 tutor or another, that will now be calculated into the retention rate. And we -- I think like someday when the industry come up with a -- like a -- more acceptable ways to calculate the retention rate, we will be much more happy to disclose that number. And also I want you to pay attention to like another factors might impact the retention rate. First is the price of the class. So when they do reregistering, we didn't do discount. And sometimes, we even slightly increase the ASP. And the second is like the time window we provide for the retention period, we still use the traditional way. There are 4 days for the retention period. It's usually from Thursday to Sunday. And we do see a lot of the companies in the industry, they kind of like each time that their retention time -- window to 10 days or even to 15 days, which makes the KPI less meaningful to compare. So I think that addressed your questions.

Yes. And I want to follow up, please, the teachers' student-teacher ratio. So the top teachers is very good resources. So do you think there are still how much room to increase the top teachers' student-teacher ratio?

So for the top teachers, there is a thing I also need to remind you. So I think that is another question that a lot of people they may pay attention to our sales concentration rate, that our top teachers may contributed over 40% of the revenue in 2018 as well as in the first quarter in 2019. So there's a reason for the concentration because like in 2018, even though by the end of the year, we have 163 instructors on hand, only half of them are in their full capacity because we provide longer than 6 months of training to those teachers. And there is a [ parameter ] for K-12 business is that when -- that needs time for instructors to grow the scale of his class -- his or her class. So because like there's -- for instance, if a teacher joins us in 2018 that, by the end of today, he already experienced the 4 or 5 times of retention period. And for a new teacher, he doesn't really have that time to grow his class scale. And for our top teachers, like I just said, we are closely monitoring their ongoing KPIs. And we also wanted to create a really healthy environment for those teachers that they are teaching the class scale that they are -- they can handle and they are feeling good about it. So right now, we do see our top teachers like they are teaching a class scale at around like low single digit of 1,000, like 2,000 or even like 1,200 students in the same class. And we will closely monitoring all the KPIs to see whether we can put their number up.

The next question today comes from Carson Lo with Nomura.

So congrats on the strong quarter. So I just have a few questions. First, to follow up about the teacher-student ratio. So based on the current KPI monitoring, so how -- yes, I just want to follow up on how -- what is the ceiling for the top teachers -- teacher-student ratio for the top teachers based on your monitoring? So to determine the room for the maximum students in the same class. And the second question is about the -- more about cross-selling. So I want to see for a cohort of users, then how many subjects do they normally apply in our platform?

Okay. So for the setting of the scale of the instructors, theoretically there's no limits for the class because our live broadcasting technology now supports 100,000 students in the same classroom without a significant latency. And in reality, we did host a class with a -- of students of over 26,000. And still, it's -- without any significant latency. But for the -- another reason or another parameter we are evaluating this optimized scale of the class is like the instructor and the tutor group as well as the content development team, also together with the technology team and the sales team is a group, which means like the instructors need to work closely with his group to make the class or to make the learning experience better. And when the group grow really quick or grow into a large group, the communication efficiency might be lower, which means sometimes like there's the optimal scale of these classes. And again, we are closely monitoring like all the ongoing KPIs. And right now, we did see for those top teachers, probably 1,000, 2,000, they are -- like they are in really good shape to handle that scale of class. And also for cross-selling, cross-selling did see different parameters across high school sector, elementary school sector and middle school sector. For instance, from each -- for elementary schools, they only learn 3 subjects at school. So -- and there -- sometimes, the parent won't be that brutal to them to have them like learn all the 3 subjects. They may just focus on the weakest one. So like for high school sector, they learn like 6 subjects at high school. And basically, they have no time to do leisures or holidays like all of their time are concentrated on all the tasks and study. So they tend to have a higher cross-selling rate. So that's kind of like a different across all the sectors. And that may subject to change whenever the enrollments' distributions across all of the school segments had to change. So usually, I mean right now, we did a really good job on cross-selling, like I just said, because we have good teachers evenly distributed on different subjects. So we will continue to encourage our students to register for different subjects with our platform.

So can I follow up quickly on the -- for example, on the high school students? So now, at this moment, on average, how many subjects were normalized with students users registered on our platform?

That also like show seasonalities like usually, that number tends to be high in the first quarter. Because if we take the scenario of the whole school year, like the school year usually starts in summer. And that when it comes to next year's first quarter, it comes to the end of the school year, so which means probably the students stays with the platform with longer time, then that's the time they're trying to do taking more courses because they have built a really trustworthy relationship with our platform. And the number in summer or in fall semester tends to be low because that's usually the time that students are newly registered and they are probably not familiar with the platform. So they tend to just try 1 subject, and they will be expanding in the subsequential semesters.

This concludes our question-and-answer session. I would like to turn the conference back over to Tip Fleming for any closing remarks.

Okay. Thank you, operator. Thank you, everyone, for joining the call today. If you have any further questions, please don't hesitate to contact us or the company directly. Thank you very much.

Thank you.

The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.