中國教育業(yè)報告：線上教育熱潮

1、3Asia Pacific Equity Research11 March 2020DS Kim(852) 2800-8597ds.kimTable of ContentsOther key points (contd)2Our two-cents on online education fever 4Whats happened? 4Whats warranted, and whats not? 5A TAM conversation 9The numbers 12Whats in the price? (Well, a lot) 15What does Price-to-Sales tel

2、l us? 16How much growth is the market looking for?17Our stock calls20Appendix 2: Recapping the products 24Offline: Size matters24Online: Different mode of delivery 26GSX Techedu 29Beautiful story, strong numbers, not such great risk/reward29 Our forecasts and key assumptions 31Valuation and Price Ta

3、rget 33Koolearn36Great for parents, not so great for investors36Our forecasts and key assumptions 38Valuation and Price Target 42New Oriental Education 45Model Update 45 Our forecasts and key assumptions 47Valuation and price target48TAL Education 52Model Update 52Forecasts and key assumptions 54Val

4、uation and Price Target 574Asia Pacific Equity Research11 March 2020DS Kim(852) 2800-8597ds.kimOur two-cents on online education feverExtremely disruptive events such as COVID-19 and school closures lead to a lot of behavioral changes, some short-term, some structural. We recap whats happened so far

5、, whats likely to stick, and whats still debatable as we try to assess potential opportunities and challenges ahead.Whats happened?Following the school closures since Chinese New Year, major online after-school tutoring (AST) companies have launched nationwide free courses for K-12 students during s

6、chool hours (Table 2), which are delivered in a massive-scale live-steaming with (very) limited tutor support. Most of these courses are also broadcast on popular online media/video platforms for free (in addition to proprietary apps and websites), such as CCTV video (央視頻), Xuexi Qiangguo(學習強國), Tik

7、Tok, Blibli, Tencent video etc., which has helped to extend their reach significantly.Some went even further. Koolearn offered its entire K-12 large classes for the spring semester for free, including normal-priced courses with full tutoring support.Obviously this free for all promotion was only pos

8、sible because Koolearns opportunity cost is only a tiny fraction of its bigger peers (i.e., Koolearns large- class revenue was only RMB0.1B last year, vs Xs RMB4.0B or GSXs RMB1.7B). But, in any event, Koolearn was able to garner 1m+ enrollments with this promotion alone, which would not have been p

9、ossible if not for this opportunistic (and clever) promotion.Table 2: Summary of national free live broadcast courses offering and datesCompanyAnnoucement dateOffering datesTAL (X)26-Jan1 Feb until schools resumeKoolearn27-Jan7 Feb 21 FebZuoyebang27-Jan3 Feb 6 MarchYuanfudao29-Jan3 Feb until schools

10、 resumeYoudao29-Jan1 Feb late FebGSX (Gaotu)4-Feb6 Feb 14 FebNew Oriental6-Feb10 Feb until schools resumeSource: Company data, J.P. Morgan5Asia Pacific Equity Research11 March 2020DS Kim(852) 2800-8597ds.kimFigure 1: X free live broadcast courseSource: J.P. MorganFigure 2: Yuanfudao free live broadc

11、ast courseSource: J.P. MorganFigure 3: Free live broadcast courses on Tencent VideoSource: J.P. MorganFigure 4: Free live broadcast courses on Xuexi Qiangguo appSource: J.P. MorganWhats warranted, and whats not?Positive #1. It is very clear that these platforms (and online courses in general for tha

12、t matter) gained significant awareness across China given the following:Most major operators say they were able to gather tens of millions of enrollments from free courses since the COVID-19 outbreak (Table 3; though definition of enrollment can be different across companies). The sheer scale of use

13、r data should help to save significant amounts of traffic acquisition costs.Our checks on search queries on both web (mobile + PC) and WeChat do indicate a surge in interest for most of major companies. Its also interesting to see some players have garnered much more attention than some others. The

14、pictures say a thousand words see below.6Asia Pacific Equity Research11 March 2020DS Kim(852) 2800-8597ds.kimTable 3: Summary of managements commentary from calls/mediasSource: Company data, China TimesFigure 5: Wechat Index: comparison between online AST playersSource: Wechat Index.Note: WeChat ind

15、ex reflects the popularity of different key words on WeChat. It integrates both search and browsing behavior data on WeChat.CompanyCommentsDateSourceGS XRight after the Coronavirus broke out, we instantly took actions and thanks to our strong organizational capabilities, we have launched free live c

16、ourses on multiple platforms, and we have attracted nearly 15 million in enrollments18-Feb-20 4Q19 earnings callYoudaoThere have been over 10m free course enrollments, since we launched the campaign on Jan 2427-Feb-20 4Q19 earnings callEDU (Koolearn)We have acquired more than 10m enrolment via Koole

17、arn free online courses offering so far02-Mar-20 Business update callTALWe saw strong free enrollment traffic and incremental online revenue of a few hundreds million RMB for the winter semester since the COVID-19 outbreak02-Mar-20 Business update callYuanfudaoOn February 3, the first day of our fre

18、e live broadcast courses, we saw in total 5m ofprimary and middle school students attended the course on the single day. As of now, 06-Mar-20 ChinaTimesthere have been more than 20m enrollments for our free coursesZuoyebangAt present, we have served more than 30 million free course enrollments, the

19、highest in the industry; meanwhile, we have served more than 1.1 million regular-price winter students.Most of the free class students are new users.06-Mar-20 ChinaTimes7Asia Pacific Equity Research11 March 2020DS Kim(852) 2800-8597ds.kimFigure 6: Baidu Index: comparison between online AST playersSo

20、urce: Baidu IndexNote: Baidu index reflects the popularity of the key words via Baidu search. The Y-axis shows the daily search volumes of certain key word via Baidu on PC and mobile.Figure 7: Sougou Index: comparison between online AST playersSource: Sougou Index.Note: Sougou index reflects the pop

21、ularity of the key words via Sougou. The Y-axis shows the daily weighted average search volumes of certain keywords via Sougou.Positive #2. Another unquantifiable benefit was the behavioral changes among students and parents. Our view is that the COVID-19 outbreak provided online K- 12 operators an

22、opportunity to showcase how their courses “could actually work” and is a potential driver for accelerating penetration. More on TAM later in the note.Positive #3. Perhaps most important, in our view, could be that a big chunk of new users came from low-tier cities. Recall, low-tier cities have been

23、the toughest area to crack for online K-12 operators, similar to most other internet verticals, due to relatively low online/mobile adoption rates, higher traffic acquisition cost, and competition from local and low-cost players.Put differently, the COVID-19 disruption seems to have expanded the ser

24、viceable addressable markets (at least temporarily), which wouldve otherwise taken much longer or cost heavy sums of marketing dollars in our view.8Asia Pacific Equity Research11 March 2020DS Kim(852) 2800-8597ds.kim but there exist a great deal of uncertainties, too. All that said, this does notnec

25、essarily mean online K-12 companies could enjoy a big jump in revenues/profits as cautioned by operators themselves. Its too early to assess the quality of new enrollments, i.e., we do not yet know how difficult or costly it will be to convert them into paid users, which requires a few challenging s

26、teps. First, operators will first need to keep/engage students in free broadcasting classes until course completion. Second, theyll then have an opportunity to convert them into short- term/promotional course users. Lastly, operators will need to entice them to enroll and pay for full-priced normal

27、courses.Its difficult for us to estimate accurately such quality (thus the scale of benefit) at this point, without access to actual trends (such as attendance rate, drop-out rate, early conversion rate, etc.). But we do think that the conversion process could prove relatively more challenging than

28、that seen in the past given two potential issues.Quality of traffic. Many (if not most) of users are likely double counted across different platforms, considering how easy and accessible it was to enroll for such courses. Not only that they were free and readily available across different media/chan

29、nels, all major K-12 players offered essentially the same type of products of similar quality, at the same time, for the same purpose. Many parents (your author included) have used the opportunity to try out different platforms & products, and we suspect the attendance and completion rates of these

30、courses to be much lower than typical promotional courses.Quality of experience. Based on our observations, free courses offered since the outbreak are not as interactive or engaging as normal courses, with regard to both in-class experience (lack of two-way interactions) and post-class experience (

31、lack of tutor supports, which is crucial for the class review, QnA, and parents consultation) the quality of those courses seems to us more akin to old- fashioned pre-recorded courses than popular dual-teacher live-streaming mode.This was totally understandable (well, for us) given a bottleneck in t

32、utor capacity (despite the significant hiring spree in recent quarters) and bandwidth, versus the sheer scale of enrollments during the period. But students and parents especially first-timers can be less forgiving even for the free courses, and it could actually prove more costly to revive/reactiva

33、te disappointed users than to newly acquire users from scratch, in our view.We are not detracting from significant potential benefits that COVID-19 could bring to the online education industry (quite the contrary); but we would simply caution against extrapolating too much from current levels of tra

34、ffic and being overly bullish.9Asia Pacific Equity Research11 March 2020DS Kim(852) 2800-8597ds.kimSource: Frost & Sullivan, J.P. Morgan estimates. Note: a deceleration in industry growth in 2018 was due to a significant yet temporary slow-down in expansion amid stricter AST regulations.Then how big

35、 can the online K-12 AST market become in 3-5 years?Heres some food for thought:A baseline. The trend growth of K-12 AST industry (offline + online) has been extremely steady at 12-13% in the past decade or so, which in our view is unlikely to accelerate or decelerate significantly.Offline vs online

36、. Offline AST has grown at a 10% CAGR in the past three or five years, while online AST has exhibited around 40% growth p.a. As such, revenue penetration of online AST has more than doubled to 8%+ in 2019, versus only about 3% five years ago.Penetration. This implies much higher online penetration i

37、n terms of enrollment or student, given the price disparity between the two, i.e., ASP of online large-class live courses (by far the most mainstream product in recent years) is around RMB30-50/hour, which is about half of regular offline classes (RMB60-100/hour, typically depending on the city-tier

38、s). Put differently, online penetration based on student numbers is probably at over 15% for Chinas AST industry last year (Figure 11).Its impossible to say with any conviction how big online K-12s total addressable market (TAM) is, not least because China is already one of the, if not the, most adv

39、anced and penetrated country for online after-school tutoring.2.7% 3.3% 4.2% 5.3% 6.5% 8.4% 11.5% 14.8% 18.5% 22.2% 25.8%Online penetration2853183543934264664935285575876143091127168214293328370415455508425576461968375582814%12%10%8%6%4%2%0%900700500300100-100A TAM conversationIt is not easy to asse

40、ss the size of the AST industry, let alone online AST, given its fragmented nature and varying statistics from different sources and institutions. That said, based on figures from consulting firms (e.g., Frost & Sullivan, iResearch), we estimate online K-12 AST revenue to be RMB42.5B in 2019, more t

41、han quadrupled in five years since 2014 at a +40% CAGR. More importantly, this implies online has penetrated about 8.4% of the Chinas K-12 AST market (based on revenues), a significant jump from 2.7% five years ago.Figure 8: China K-12 AST: Market size and online penetration (revenue-based)(Rmb B)20

42、142015201620172018 2019E 2020E 2021E 2022E 2023E 2024EOfflineOnlineY/Y growth (RHS)10Asia Pacific Equity Research11 March 2020DS Kim(852) 2800-8597ds.kimFor instance, even in your authors home country, S. Korea, which has one of the highest AST penetration AND internet penetration rates, online cour

43、ses accounted for only 3.1% of K-12 AST revenues in 2019. Also, only 7.5% of K-12 students in Korea have participated in online AST during 2019, versus a remarkable 75% of AST participate rate for total (Figure 10). Granted, this may not be a fair comparison given the relatively inferior product qua

44、lity in Korea, in that most of online courses are offered in a pre-recorded format with limited tutor support, in turn possibly deterring online adoption in Korea.Figure 9: Korea AST: Market size and online penetration (based on revenue)Source: KOSIS; Note*: AST excluding extracurricular activitiesF

45、igure 10: Korea AST: Participation rates based on studentsSource: KOSISLast but not least, notwithstanding clear advantages of online (much better accessibility and affordability), we are of the view that offline-based products still offer meaningfully superior quality than online, when it comes to

46、efficacy of delivery (i.e., the level of attention and engagement) and teaching content (more localized curriculum for offline, vs nationwide/standardized online courses). We strongly believe online-offline is not mutually exclusive in education; we expect online and offline tutoring to co-exist in

47、any foreseeable future, complementing each other to tap into the huge pool of Chinese students.Please refer to our in-depth report, “China Education: Whats Better Than Learning?” (pages 14-31) for our detailed analysis and discussions on online vs offline market and products.1.6% 1.5% 1.3%1.1%1.1%2.

48、7%3.1%Online penetration17.616.916.215.614.914.413.913.513.614.315.417.316.716.015.414.714.213.713.213.313.914.912%8%4%0%-4%-8%051015(Won T) 2020092010201120122013201420152016201720182019OfflineOnlineY/Y growth1.3%1.0%2.3% 2.3%71.7%69.4% 68.8% 68.6% 68.8% 67.8%71.2%72.8%74.8%3.7%3.7%3.0%2.8%2.5%2.4%

49、2.2%5.1%5.2%0%6.70%60%50%40%30%20%10%0%80% 75.0% 73.6%7.5%2019AST total2009201020112012201320142015201620172018Online courses11Asia Pacific Equity Research11 March 2020DS Kim(852) 2800-8597ds.kimSource: Company websites, J.P. Morgan1099468616360606046403837-12010080604020TALEDUTALEDUTAL(Peiyou Live)

50、EDU(Koolearn DFUB)GSX (Gaotu)YoudaoTALYuanfudaoEDU(Koolearn)ZuoyebangFigure 11: Course ASP comparison by product type (2020 Spring Semester Grade 7 Math)Rmb/hrOffline normal classOnline localised contentOnline live large class (dual-teacher)Offline dual teacher1096863469461603860604037-1201008060402

51、0OfflineDual-teacherPeiyou LiveXOfflineDual-teacherDFUBKGSX(Gaotu)YoudaoYuanfudaoZuoyebangSource: Company websites, J.P. MorganFigure 12: Course ASP comparison by company (2020 Spring Semester Grade 7 Math Subject)Rmb/hrTALNew OrientalOther online players12Asia Pacific Equity Research11 March 2020DS

52、 Kim(852) 2800-8597ds.kimThe numbersTable 4: China K-12 AST: Industry revenue and our forecasts(RMB bn)20142015201620172018 2019E2020E 2021E 2022E 2023E 2024EOffline AST284.5317.6354.1393.1425.0 465.4493.3527.8556.9587.5613.9Y/Y12%12%11%11%8%10%6%7%6%6%5%Online AST8.010.815.521.929.742.564.191.4126.

53、6167.7213.8Y/Y31%35%44%41%36%43%51%43%39%33%28%Online penetration2.7%3.3%4.2%5.3%6.5%8.4%11.5%14.8%18.5%22.2%25.8%K-12 AST revenue292.5328.4369.6415.0454.7 507.8557.4619.2683.4755.2827.7Y/Y12%12%13%12%10%12%10%11%10%11%10%2.7%3.3%5.3%8.4%14.8%18.5%22.2%25.8%5.3%6.4%8.0%4.2%10.0%6.5%12.3%15.4%11.5%20

54、.6%25.7%31.3%36.3%10%20%30%40%Source: Company data, Frost & Sullivan, J.P. Morgan estimatesReflecting a likely boost from COVID-19 disruptions, we now model what we believe to be a bull-case scenario, projecting the online K-12 AST market to triple in three years to RMB127B ($18B) and quintuple in f

55、ive years to RMB214B ($31B).This is predicated on the assumption that online ASTs revenue penetration can double to 18-19% by 2022E and triple to 26%+ by 2024E (versus 8%+ in 2019). This is an aggressive target, in our view, as it essentially implies nearly 45% of AST enrollments will be taking plac

56、e online (from 15%+ last year), considering the online-offline price gap is unlikely to narrow given fierce competition and heavy promotions for online; this seems possible yet quite aggressive to us.Figure 13: China K-12 AST: Online penetration rates based on revenue vs enrollment50%41.1%0%20142015

57、2016201720182019E 2020E 2021E 2022E 2023E 2024ERevenue basedEnrollment basedSource: Frost & Sullivan, J.P. Morgan estimatesPlease refer to company pages for detailed forecasts. But from a market-share perspective, we have modelled the following for our covered companies (K-12 only, excluding non-K12

58、 businesses).TAL: We assumed its X to retain its leadership during our forecast period, though the gap with No 2-3 players (namely Yuanfudao and GSX) could narrow. Were comfortable modeling TAL to outgrow the industry, given its unparalleled brand equity, high-quality curriculum, scale benefit, as w

59、ell as best- in-class management. We assume its revenue to grow at a 63% CAGR in the next three years, pushing up its market share to 14% by 2022E from 10% in 2019. We then assume the company to grow on-par with the industry at around 30%p.a. for the remaining forecast period.13Asia Pacific Equity R

60、esearch11 March 2020DS Kim(852) 2800-8597ds.kimGSX: We now give it the benefit of the doubt that its star-teacher strategy and aggressive viral/SNS marketing will stay as its sustainable advantages against peers, in turn driving its revenue to grow at a 95% CAGR over the next three years and more th