Back when I was a teacher, I used to tutor 3 year-olds on the Gifted and Talented (G&T) test to supplement my abysmal income. This was one of those jobs that makes you reassess your values because it undermines the spirit and principle of G&T programs.
Tutoring can be an equalizing force when used for students who need remediation. At the same time, it’s a multi-billion dollar industry that exacerbates inequality.
Let’s pick this thing apart to understand why there’s so much frenetic movement in the space, what’s going on with China (including why their tutoring companies are shedding value [above]), and why everyone is racing towards AI.
High Level Trends in the Space
Tutoring startups are raking in funding recently:
Ignite! Reading raised $10MM Series A (February 2023)
Scoolhouse.World (Khan Academy’s peer-tutoring company) received a $MM (undisclosed sum) donation from Citadel (Jan 2022)
GoStudent raised $95M (August 2023) on top of $340M Series D in 2022
Preply raised $70M (July 2023) bringing their Series C to $120M
Departments of Education are partnering with companies to offer tutoring services.
Rhode Island + Khan Academy
Ohio + Amplify.
Little Rock ISD + Ignite! Reading
Newark Public Schools + Khanmigo
Teach for America launched their Ignite Fellowship, an AmeriCorps program for college students to become tutors
In 2021, China banned all private academic tutoring (worth over $310Bn). We know how this type of thing goes (Prohibition, anyone?). Blackmarket tutoring is on the rise, with prices up 50% - 200%, depending on the subject (source: Bloomberg).
Lots of movement from Private Equity. It’s a particularly appealing space because there are tons and tons of opportunities for efficiency, it’s a cash-cow, and tutoring companies will pay out the nose later on for customer acquisition and consolidation.
Primavera Capital Group (a reputable Chinese PE firm) acquired The Princeton Review and Tutor.com (Note: Tutor.com has a Department of Defense contract and this remains a concern) (May 2023)
Alpine Investors acquired FEV Tutors (July 2022).
IK Partners just sold Studienkries, which they acquired in 2017, to GoStudent (Dec 2022)
Meanwhile, PE firms in China had opposite momentum—leaving the tutoring space in China. PE alone had about $120Bn invested in Chinese tutoring/education companies, and those funds are now in a twilight zone as their portfolio companies try to pivot.
What are they deploying capital on?
Companies are investing in AI-powered tutors:
Khan Academy is, IMO, the leader in this space with Khanmigo.
Preply is investing heavily in AI, leveraging GPT-4 to personalize learning. Full disclosure, I have a wonderful Spanish tutor from Venezuela and their platform is amazing (it even tracks my talk-time).
Duolingo, un producto fantástico (10 day streak, baby!), launched their Duolingo Max, which is powered by GPT-4 and which lets you roleplay conversations with the generative AI. Pretty neat use-case
Tutoring companies are also acquiring other companies—primarily for inorganic growth in what is a super competitive market. For the most part, tutoring has been relatively undifferentiated and commoditized. There wasn’t/isn’t much to distinguish Mathnasium from, say, Varsity Tutors.
GoStudent, based in Austria, acquired Studienkreis, a traditional tutoring company founded in 1974, to roll up their students and tutors. They just gobbled up 125,000 students and 1,000 physical locations.
Byju spent $1Bn ($600M in cash) for Aakash, a 33-year old chain of traditional tutoring centers (which Blackstone bought in 2019)
Why are tutors suddenly springing up?
Well, that question is actually a bit of a misdirection. Tutoring has been around forever —but it’s been a BUSINESS since the early 2000s. Mathnasium, Kaplan, Princeton Review. They have largely been a mechanism for (1) remediation, and (2) to game the system (like SAT tutors, or G&T tutors).
Khan Academy’s Schoolhouse, Amplify and Khanmigo are squarely in the space of remediation. COVID exacerbated demand. NAEP scores in Math and Reading had their steepest decline since 1971, erasing decades of progress.
Source: Nation’s Report Card
Historically, access to tutoring has been extremely expensive (and for most of human history, limited to royals and rulers). But soaring demand paired with technology has allowed tutoring to scale at a relatively cost-effective way, although this is still the biggest constraint for public education tutoring. For the DOE, public education tutoring still costs about $2,000 to $3,000 per student (and is still 100% worth it) because it’s effective (offering the most exposure to a student’s Zone of Proximal Development).
Paying Less for More
Remote-instruction technology (Zoom, ClassIn, Google Classroom) has started to disassociate costs from geography. You can now get a tutor from across the world! Companies like EDGE Tutors (spun out of Edukasyon) arbitrage tutoring services, so that you’re paying a Filipino hourly salary for high quality instruction. Through Preply, you can get high-quality direct language instruction at a fraction of the cost of your neighborhood language class.
Another trend that PeerTeach and Schoolhouse are pioneering is peer tutoring—where students teach each other. We’re big fans of this method. It’s difficult to implement, since you have to train the students to be effective tutors, but it pays dividends for all students (sources: link to efficacy study on students in Nigeria; GreatSchools article on implementation).
These companies are reducing the cost of tutoring.
AI Tutoring
And then there is the rise of generative AI, which was the talk of the town at ASU-GSV. There are a lot of benefits to AI tutoring, and it was a natural progression from chatbots. It can scale easily and cost effectively. The marginal cost of tutoring a student drops from $2K to $100 (the price of Khanmigo). The drawback here, at least initially, is that the student lacks the socio-emotional connection, and studies show that person-to-person interactions are important for learning (source: PNAS study). Of course, at the rapid pace of development of generative AI, we could get to a point where an AI tutor can effectively mimic a visual human, complete with facial expressions and all.
Two Competing Narratives: China vs. India
This piece personally interests me (Chris), as an Asian, but also as a history nerd, and both China’s and India’s cram cultures are so ingrained in history.
China
In China, the civil service examination traces its roots back to the Sui Dynasty (518) and eventually became dominant during the Song Dynasty (960 - 1279). Passing this civil service exam guaranteed you a well paying job in government. During the Qing Dynasty, there was a 1% pass rate. The wealthy would pay for tutors, bribe examiners, or cheat. At its best, the civil service exam distributed talent, increased social mobility, and led to the rise of the technocrat. At its worst, it exacerbated inequality and incentivized test-taking over talent/skill (here’s a phenomenal article on the impact of these exams—and their subequent abolition, by Ying Bai and Ruixue Jia).
Sounds kind of familiar?
Image: A painting of the Imperial exam dating back to the Song Dynasty.
Today, wealthy families pay for tutoring, proxies, and everything in between for LSATs, GMATs, and the all-important Gaokao, the Chinese college entrance exam. Your score on the Gaokao effectively determines the quality of university you attend. In 2017, 9.4 million people took the test, and about 10% were accepted into first-class admission universities.
India
In India, the Civil Service exams today trace their roots back to the British Imperial Service Tests—specifically the Indian Civil Service for the subcontinent. In 1853, Parliament opened up the ICS to all British subjects. Today, the Civil Service exams take about 32 hours and nearly 1 million Indians take them each year. There is a 0.2% pass rate. National entrance exams, which dictate college acceptance, attract even more applicants. The medical and engineering exams were taken by 2.74 million Indians this past year for 64K spots: More than 2.6M people failed the exams outright.
Image: Orientation day at the Allen Career Institute, a “coaching” school in Kota that tutors 1.25M students per year with more than 2,000 instructors. For context, this is more students than are enrolled in the NYDOE.
Image Source: Zishaan A Latif for the NYT
China’s Response vs. India’s
In 2021, the CCP cracked down on private tutoring as part of their “double education” reforms—essentially, trying to stop unfair advantages and also limit the exhausting toll that cram-schools were having on their youth.
New Oriental, previously the largest tutoring company in the world, and which trades on the NYSE, reported an $876M loss after the crackdown.
Obviously, this hasn’t panned out. A black market tutoring industry has supplanted the formal one, driving up prices by about 200% (Source: Sixth Tone) and exacerbating affordability.
This has also pushed Chinese companies to seek new revenue streams. ClassIn, one of the largest remote education platforms (and used extensively by tutors) has been making an insane push into foreign markets. They had a huge booth at ISTE and were wandering the floor, taking notes.
Whereas Chinese companies are forced to look outwards, India is looking internally. South Asia is the largest growing private education market. Their companies are looking to solve both sides of the spectrum—remediation, and high-end exam tutoring. There are 444 million adolescents (age 18>) in India. Vedantu, a tutoring startup (which has raised money from Omidyar, GGV, Tiger Global) is now a unicorn. They have hundreds of thousands of subscribers and more than 25M free users who pay anywhere from $1.40 for a short session/short form content to $700 for a longer class. But students can also hire tutors across a range of subjects, primarily focused on Gr 2 - 6.
India’s tutoring companies are growing FAST, eating up demand inside the country. China’s edtech companies have shed value. Yuanfudao, another Chinese edtech/tutoring company, was valued at $15.5Bn right before China’s reforms. Zuoyebang was worth more than $10Bn. Now? Oof. Now Yuanfudao is trying to get into exercise, and investors are trying to liquidate their capital. Publicly traded New Oriental (earlier) has seen their stock value plummet from $195 at its peak in Feb 2021 to about $55 today.
What does this ALL MEAN!?
Two weeks ago we talked about how the higher education system is bifurcating and top-tier universities will only become wealthier and more exclusive. The tutoring industry will continue, as it always has, to cater to affluent families trying to get an edge. This is where most of the money has been historically, and so you’ll see continued consolidation here.
But declining NAEP scores, COVID learning-loss, and tech-enabled scale are causing a HUGE boom on the other side of the spectrum: Remediation tutoring. This is where most of the action will be, in my opinion, because it’s a) a huge customer base, b) historically been underserved (due to high OpEx), and c) is going to be on the receiving end of capital from governments and foundations. 62% of students in the US aren’t proficient in Math. That’s a lot of people to serve!
You’ll see generative AI tutors deployed here, too. Maybe at a slower rate with districts and states, simply due to some of the response-risk of generative AI (“My tutor said WHAT!?”). But a few schools will pilot (such as Newark and Khanmigo) because they’re either forward thinking or need a cost-effective solution and will cross that chasm.
Affordable options and subscriptions like Vedantu, low cost arbitrage providers like Edge Tutor, etc., have competitive advantages in this space.Peer tutoring and more scalable/remote tutoring is also really exciting. Like, REALLY exciting. Because we NEED tutoring! Our educational system is breaking at the seams. It should be all hands on deck—whether it’s a tutor from across the world, or an AI.
 | A guest post by
|
 | A guest post by
|
 | A guest post by
|