I was mildly surprised last week to hear that Tencent is pulling out of a funding deal with edtech darling Vipkid, known for one-on-one remote tutoring. Sources close to the deal told Reuters that Tencent had already given verbal agreement, but changed course after stricter regulations for online education came into effect. Reuters reported that Vipkid was seeking funding at a valuation of $4.5 billion. The company’s previous round valued them at $3 billion.
The new regulations, issued in July, mandate that not only must all teachers on a platform have teaching qualifications, the platforms must also publish them along with teachers’ work experience. On top of that, over the last year, key management personnel as well as a large number of foreign staff have left as the company struggles to turn a profit.
Companies like Vipkid, however, represent only a small fraction of the edtech market. Similar to verticals like “medtech” and “proptech” (property), edtech is less a known market than a lot of entrepreneurs trying to figure out how to apply existing technology to profit in education. Early learning, K12, STEAM, professional development training, corporate training, as well as schools and classrooms, all look like fertile grounds for disruption. But in many cases, it’s still “solutions searching for a problem.”
A few weeks ago, reporters Tony Xu and Wei Sheng gave a general overview of Bytedance’s edtech endeavors. Intrigued, I decided to take a closer look.
Bottom line: The edtech market is huge. In 2018, Deloitte estimated that the market was then worth RMB 2.9 trillion (about $410 billion) and would be worth RMB 3.4 trillion in 2020. And one of China’s most successful content companies is poised to dominate. With their expertise in UX design and powerful AI recommendation engines, Bytedance is making a lot of different bets in different education areas—their most promising is the application of the recommendation algorithms it uses to keep viewers watching videos to online lessons, and using NLP to teach languages. While all of them are online, Bytedance’s education plays cover an interesting mix of pedagogical methods and design preferences. Of course, just because its Bytedance doesn’t mean they’re automatically successful; it just means that there’s a lot of money being thrown around. However, CEO and founder Zhang Yiming is personally passionate about education and, with his track record, is probably playing the long game.
A brief timeline of edtech at Bytedance
(A version of this section originally appeared in In Focus: Bytedance #12)
- December 2017: Bytedance holds education industry conference to talk about the integration of the sector with technology. The event is the first hint of pedagogical ambition dropped by the Beijing-based unicorn.
- March 2018: Bytedance acquires Openlanguage , an online English course provider.
- May 2018: Launches Gogokid, a one-to-one tutoring platform for Chinese children to learn English online with foreign teachers.
- July 2018: Launches Haohao Xuexi , a knowledge-sharing app that features content covering career advice, parenting, culture, and wealth management.
- August 2018: Bytedance leads $49.5 million Series C funding round for San Francisco-based education technology company Minerva Project.
- December 2018: Launches Aikid, a foreign teacher live-streaming platform.
- January 2019: Bytedance licenses patents from now-defunct smartphone maker Smartisan. The company indicates that they are meant to expand and develop online education business. By this time, Bytedance had reportedly spent over RMB 400 million on Gogokid.
- April 2019: Bytedance lays off half of Gogokid’s staff and reduces sales team to 200 employees. China media also report that Aikid had suspended operations four months earlier.
- May 2019: Launches K-12 online education platform Dali Ketang, which offers courses from primary school to high school. Chinese tech news outlet 36Kr reports that Bytedance acquired another online teaching platform named Qingbei Wangxiao to help with the development of Dali Ketang.
- July 2019: Tech Planet reports that Bytedance is testing short-video-based English-learning app named “Tangyuan English.” The app officially launches in August.
The state of education in China: Education has a high place in East Asian cultures. Japan, Korea, and China all put enormous amounts of pressure on their children to do well in school and on exams. However, for most, graduation does not end the educational journey:
- In a 2019 report, L.E.K, a management consulting firm headquartered in London and Boston, estimated that the children’s education market in 2018 was worth RMB 500 billion, with an average growth rate of 15% from 2013.
- In that same report, the firm estimated that the adult education market, including higher education and professional development training, in 2017 was worth RMB 633 billion with an average growth rate of 7.2% from 2013.
- They predict that between 2017 and 2020, the market for white collar professional training will grow 12.6% from RMB 281 billion to RMB 509 billion.
- Deloitte estimated that online education was 9.32% of the education market in 2018. In 2020, they predict it will be 10.41%.
- By 2020, the consultancy also predicts that K12 and STEAM education will make up 44.7% of the online education market while corporate training and professional development training will account for 10% and 18%, respectively
The tech in edtech: Let’s face it, most of us roll our eyes when we hear about the latest “AI+[insert sector]” or “blockchain+[insert sector]” company. Most of these companies do absolutely nothing with their buzzword tech. At this point, more often than not companies peddling AI aren’t talking about :the textbook definition referring to machine learning-driven algorithmic decision making systems. Read marketing-speak AI as a colloquial definition meaning anything that has a complex logic path, i.e. the same as “smart.”
Unfortunately for us, there are no word police in the tech industry (Well, almost none—it’s a lonely fight.). Marketing and sales people can walk around linking the latest buzzword with a product that has little to do with it. While there certainly is a lot of room for technology in classrooms, both real and virtual, it is hard to find much of it in publicly described use cases.
The Bytedance play: Bytedance, as a company specializing in developing AI systems, does actually have in-house tech that it can immediately apply to its edtech projects: behavior-driven predictive recommendation algorithms. At least two of their apps, Dubai Bei Danci and Tangyuan English, include content feeds similar to those found in Jinri Toutiao and Douyin/TikTok. Almost all of the giant’s education apps that don’t feature real people are also built with one of the most mature implementations of AI: natural language processing.
Tangyuan English, for example, has you learning spoken English via short and sometimes funny skits. After your virtual instructor, a pre-recorded video, gives you a line, you have to repeat it back. Based on how well you matched their database’s baseline for good pronunciation, your virtual instructor will react negatively or positively. When I tried it, there were a few instances where I was not able to move on because my pronunciation was slightly too far off from a Chinese English speaker, the dataset most English language learning NLP systems in China are trained on.
There is, of course, a lot of room for more sophisticated implementations of artificial intelligence: truly personalized programs and courses that adapt to the needs and learning style of the student. But the technology just isn’t ready to scale and there’s little evidence that Bytedance is exploring this currently. However, they may be the only company doing edtech that can legitimately claim to be using AI in their products.
Holding a bucket in a storm: To paraphrase a piece of economic wisdom, “edtech is recession proof.” The fact that one-on-one tutoring has taken a hit recently (a la Vipkid and Bytedance’s Aikid/Gogokid) doesn’t mean we’re seeing a downturn in the edtech vertical. Rather, they were a combination of over-specialization, poor management, and tightened regulations. Edtech is a huge space and there’s a lot of opportunity inside and outside the app economy. Bytedance is the best placed 2C company to actually apply real technology to education.