Why Most EdTech Tutoring Startups Fail (And What the Survivors Did Differently)

In January 2024, Zenius, one of Indonesia's oldest and most recognized edtech brands. shut down after 20 years of operation. The company had raised $40 million from top-tier investors like Alpha JWC and Sequoia. At its peak, it served 16 million students and employed over 1,000 people. None of that mattered. When the funding dried up, so did the company.

We've been watching the online tutoring space in Southeast Asia for years, and Zenius's collapse crystallized something we'd long suspected: online tutoring, as most companies practice it, isn't actually a business. It's a customer acquisition channel masquerading as one.

The numbers tell a brutal story. Global edtech funding crashed 87% — from $5.8 billion in 2021 to just $712 million by 2023. The easy pandemic growth evaporated. And when it did, the companies with real unit economics survived. The rest became cautionary tales.

Here's what we've learned about who wins, who loses, and what we look for when we evaluate tutoring businesses in the region.

The Landscape: A $10 Billion Market with a Graveyard Problem

Southeast Asia's edtech market reached $10.7 billion in 2024 and is growing at nearly 15% annually. K-12 tutoring represents almost half of that: a reflection of the region's deep-rooted educational anxiety and competitive exam culture. Indonesia, Vietnam, and Singapore anchor the market, each with distinct characteristics that reward different business models.

But behind these growth numbers lies a graveyard of failed startups. Zenius is the most prominent casualty, but it's not alone. Snapask; a Hong Kong-based tutoring app that raised $35 million to expand across SEA, went dark in 2022. Topica, a Vietnamese pioneer that once employed over 1,000 people, has shrunk to just 18 employees. These weren't small experiments. They were well-funded companies with millions of users.

The pandemic created a dangerous illusion. When schools closed, usage exploded. Companies grew 3-5x almost overnight. Investors saw hockey stick charts and poured in capital. But that growth was artificial. students had nowhere else to go. When they did, many didn't come back. And the companies that had optimized for user acquisition suddenly discovered they'd never figured out how to make money.

The Math That Kills Companies

We've looked at dozens of tutoring businesses across the region, and the unit economics tell the same story over and over. Customer acquisition costs (CAC) for B2C edtech in Asia typically run $150-300 per student. Monthly subscription prices? $10-15 in Indonesia, $50-150 in Singapore. Average retention? Six to twelve months, if you're lucky — annual churn often exceeds 50%.

Do the math: if you spend $200 to acquire a student who pays $12/month for 8 months, your lifetime value is $96. You're losing money on every single user you acquire.

Before COVID, marketing spend was 20-25% of revenue for most edtech companies. During the pandemic, it dropped because growth was organic: schools were closed, and parents were desperate. But after the pandemic, CAC spiked to 70-80% of revenue as companies competed for the shrinking pool of engaged learners. The companies that had built their financial models on pandemic-era assumptions discovered those assumptions were fantasies.

This is what we mean when we say online tutoring isn't a business; it's a customer acquisition channel. The product itself, at the prices most markets will bear, doesn't generate enough value to pay for its own growth. Something else has to.

What Separates the Survivors from the Casualties

The companies that survived the funding winter share a pattern we've come to recognize. They didn't treat online tutoring as the product. They treated it as a wedge, a way to acquire customers cheaply so they could sell them something more profitable.

Ruangguru is the clearest example. Indonesia's dominant tutoring platform raised $212 million and serves roughly half the country's online K-12 market. When the funding environment turned, they didn't try to make their B2C subscription model work. Instead, they pivoted aggressively into B2B corporate training through their Skill Academy product. Today, they serve over 500 corporate clients. The unit economics of selling to enterprises. longer contracts, lower churn, no CAC per user — are fundamentally different from selling to parents.

Geniebook in Singapore took a different path. They recognized that Singapore's high-ARPU market could support premium pricing, but only if they delivered results that justified it. Their answer was hybrid: online platform plus seven physical learning centers across the island. The offline presence does several things: it increases perceived value (parents see where their money goes), improves outcomes (struggling students get face time with tutors), and dramatically increases lifetime value. Geniebook has now achieved three consecutive quarters of profitability. They're one of the few pure-play tutoring companies in the region to do so.

Vuihoc in Vietnam is still early-stage but following a similar playbook. After raising $6 million, they didn't try to out-spend competitors on user acquisition. Instead, they acquired The IELTS Workshop, a specialized test-prep provider. Test prep has fundamentally different economics than general tutoring; parents have a specific, time-bounded goal (pass the exam), willingness to pay is much higher, and outcomes are measurable. It's a natural upsell for students who came in for general supplemental learning.

Now contrast these with the casualties.

Zenius had 16 million users but couldn't monetize them. They tried acquiring Primagama, an offline tutoring chain, but the integration failed. They never built B2B revenue. They never figured out how to increase LTV beyond the basic subscription. When the music stopped, they had users but no business.

Snapask charged per question, students would photograph homework problems and get answers from tutors. The model had viral potential but zero stickiness. Once students got their answers, there was no reason to stay. CAC was high, retention was abysmal, and there was no upsell path. The company raised over $35 million and effectively disappeared.

The pattern is clear. Pure B2C online tutoring. where the subscription is the product and growth comes from marketing spend — doesn't work in Southeast Asia. The CAC is too high, the LTV is too low, and there's no moat. Survivors either found a way to monetize differently (B2B), increase LTV dramatically (hybrid/offline), or escape to a higher-margin segment (test prep).

What We Look For in Tutoring Businesses

When we evaluate tutoring companies now, we start with a simple question: where does the money actually come from?

We've learned to be skeptical of impressive user numbers. Zenius had 16 million users and still died. What matters is paying, retaining customers: and whether the cost to acquire them makes any sense. We want to see LTV:CAC ratios above 3:1. In pure B2C edtech, we rarely see that. The companies that achieve it have either cracked B2B distribution or built something that dramatically increases what customers are willing to pay.

We look for B2B revenue or a clear path to it. Schools, governments, and corporations don't need to be marketed to the same way parents do. A single school partnership can deliver hundreds of students at near-zero CAC. Enterprise training contracts have 2-3 year terms, not 8-month average retention. When Ruangguru pivoted to Skill Academy, they didn't just add a revenue line; they fundamentally changed their unit economics.

We pay attention to hybrid models, especially in markets that can support premium pricing. Online-only is efficient but commoditized. When everyone offers the same video lessons and practice problems, price becomes the only differentiator. Physical presence, whether tutoring centers, school partnerships, or blended programs. creates switching costs and justifies higher prices. Geniebook's seven campuses aren't just delivery channels; they're moats.

We're cautious about companies that rely on continued fundraising to fund growth. The 2021 environment isn't coming back. Companies need to show a path to profitability at current or near-current scale, not a hockey stick that requires three more funding rounds to achieve. The survivors of the funding winter were the companies that could tighten their belts and still function. The casualties were the ones whose models only worked with infinite capital.

And we watch for regulatory risk. China's tutoring crackdown in 2021 destroyed a $100 billion market overnight. TAL and New Oriental lost 80-90% of their value. 170,000 people lost their jobs. Southeast Asian governments are supportive of edtech today, but the anxiety that drives tutoring demand is the same anxiety that makes it politically sensitive. Companies with diversified revenue — especially B2B and corporate training: have a natural hedge. Pure K-12 B2C players are exposed.

Looking Forward

The online tutoring market in Southeast Asia is entering what we'd call a consolidation phase. The weak players have been flushed out. The survivors are profitable or close to it. Funding is returning cautiously to companies with proven unit economics.

We expect the winners to start acquiring. Ruangguru has the scale and the war chest. Geniebook is already making moves; they recently acquired AfterSkool and Best Physics Tuition in Singapore. The fragmented landscape of small tutoring providers is ripe for roll-up by platforms with distribution advantages.

The wildcard is AI. We're already seeing AI tutors cut tutor preparation time by 50% or more. If AI can deliver comparable learning outcomes to human tutors, the entire cost structure changes. Tutor labor is 60-80% of the cost of live tutoring sessions. Eliminate that, and the unit economics that killed so many companies suddenly start to work. The question is whether incumbents will adapt or whether AI-native entrants will leapfrog them.

For founders building in this space, the lesson is clear: don't build a tutoring company. Build a company that uses tutoring to acquire customers for something more valuable. For investors, the lesson is equally clear: user growth isn't a business model. Unit economics are.

We've watched 87% of edtech funding evaporate. The companies that survived weren't the ones with the most users or the biggest raises. They were the ones who understood that online tutoring is a channel, not a destination, and built accordingly.