Education / KnowledgeChina
← Sticker Wall
JOURNEY · DETAIL
Offcn Education: From Li Yongxin's Dorm Room Startup to a 100-Billion-Yuan Collapse and Restructuring
Founded: Li Yongxin, Lu Zhongfang · Offcn Education Technology Co., Ltd.
JOURNEY
Key Fields
FIELD STAMPSIndustryEducation / Knowledge
RegionChina
ScaleGiant
ChannelOther
Origin
In 1999, Li Yongxin graduated from the School of Government at Peking University and chose to start a business rather than pursue a guaranteed postgraduate spot. At the time, universities across China were undergoing massive enrollment expansion, with graduates increasing from 847,600 in 1999 to 1.8775 million four years later. As job competition intensified, the civil service exam began to gain popularity. Drawing on his own job-hunting and exam experiences, Li identified training for government-sector employment as a blue-ocean market with rigid demand. His first two ventures (in general education and a top-scorer lecture tour) failed, leaving him so poor he couldn't afford a 1.6-yuan fried pancake. In 2001, following his mentor's advice, he focused on civil service exam training, and in 2003, with 30,000 yuan in startup capital from his mother, Lu Zhongfang, he officially founded Offcn Education.
Milestones
1999
Exploration Failure
After graduating from Peking University in 1999 and giving up his postgraduate spot, Li Yongxin started his business in a 6-square-meter room near the university with a monthly rent of 600 yuan. He attempted various projects, including general education, top-scorer lecture tours, and math tutoring, all of which failed. At his lowest point, he couldn't even afford a 1.6-yuan fried pancake, and the team relied on loans to survive. It wasn't until 2001, when he pivoted to civil service exam training on his mentor's recommendation, that he found his direction. This phase lasted from 1999 to 2001.
2001
Entering the Civil Service Exam Market Turning Point
In 2001, the team officially entered the civil service exam training market, pioneering the industry. In 2002, they launched Offcn.com, China's first portal for civil service exams, using online content to drive offline course enrollment. In 2003, his mother, Lu Zhongfang, provided 30,000 yuan in savings as startup capital, and Offcn Education was formally established. Early growth was driven by university lectures and specialized interview training. This phase lasted from 2001 to 2003.
2005
National Expansion Growth
Offcn Education expanded nationwide using a direct-operation branch model, distinguishing itself from competitors' franchise models to ensure teaching quality. By 2012, it covered 31 provinces and cities, with approximately 440 branches and 6,000 instructors. They also launched their first series of civil service exam practice books, creating a three-pronged driver of book publishing, in-person classes, and online schools, firmly establishing themselves as a leader in the industry. This phase lasted from 2005 to 2012.
2014
Diversification and IPO PMF
Offcn expanded its offerings from civil service exams to over 100 categories, including public institution recruitment and teacher certification. They pioneered the 'agreement class' (tuition refund if failed), where students paid tens of thousands of yuan in advance, with the prepaid funds forming a massive capital pool to fuel expansion. In 2018, they went public via a backdoor listing through Yaxia Automobile, completing the deal in 2019 to become the first A-share listed civil service exam training company. This phase lasted from 2014 to 2019.
2020
Peak Growth
In November 2020, Offcn Education reached a peak market capitalization of approximately 265.8 billion yuan. That year, revenue was about 6.2 billion yuan, with over one million students trained. Li Yongxin and his mother, Lu Zhongfang, topped the education industry's wealth rankings with a net worth exceeding 140 billion yuan. Li publicly pledged to donate 10 billion yuan to Peking University, marking the height of his success.
2021
Collapse Failure
Impacted by the spillover effects of the 'Double Reduction' policy, a wave of refund requests from agreement classes, and blind expansion (spending heavily on land and regional headquarters), the refund rate soared above 65% starting in 2021. Performance plummeted, leading to massive losses, widespread refund defaults, and student lawsuits. The market cap fell by over 90% from its peak. Li Yongxin's pledged shares were forcibly liquidated and judicially frozen. This phase lasted from 2021 to 2023.
2024
Restructuring for Survival Pivot
Starting in 2024, the company saw a slight turnaround. In 2025, revenue reached 2.237 billion yuan with a net profit of about 49 million yuan, but outstanding refund liabilities remained at approximately 508 million yuan by the end of 2025, with an asset-liability ratio of about 86%. By 2026, the stock price had fallen about 95% from its 43.58-yuan high, closing near 2.36 yuan, with a market cap of only about 14.5 billion yuan. The controlling shareholder's stake dropped to about 12%-13%, branches shrank from over 1,600 to about 800, and staff decreased from over 40,000 to about 7,300. The company is now betting on AI job training devices and vertical large models. This phase lasts from 2024 to 2026.
Turning Points
- In 2001, following a mentor's advice, Li abandoned general education to focus on civil service exam training, tapping into the rigid demand for government-sector jobs following university enrollment expansion.
- Pioneered the 'agreement class' with tuition refunds, using the prepaid capital pool to support rapid expansion, which both fueled the 260-billion-yuan market cap and planted the seeds for the collapse.
- In 2018, the backdoor listing via Yaxia Automobile turned the company from an offline training provider into a capital operation platform, where expansion logic became hostage to market cap expectations.
- The 2021 refund wave, compounded by 'Double Reduction' panic, caused the financialized model of agreement classes to backfire, plunging the company from a wealth-creation myth into a debt crisis.
- After 2024, the company cut agreement classes, returned to standard pricing, and bet on AI hardware, beginning a difficult recovery process while resolving debts.
Failures & Pitfalls
- Two consecutive startup failures between 1999 and 2001; neither the general education nor the top-scorer lecture projects were viable, leaving the founder unable to afford a 1.6-yuan meal.
- Blind, asset-heavy expansion during the peak period, spending heavily on land and regional headquarters, which tied up capital in non-teaching assets and exacerbated the cash flow crisis.
- The 'agreement class' model essentially used student tuition as an interest-free capital pool. When the refund rate exceeded 65% in 2021, the cash flow backfired instantly, triggering nationwide refund lawsuits.
- The controlling shareholder's high-ratio share pledging led to forced liquidations and judicial freezes after the stock price plummeted, reducing his stake to about 12%-13% and putting control at risk.
- As of the end of 2025, approximately 508 million yuan in refund liabilities remained unpaid, and revenue has shrunk by over 60% from its peak due to the collapse of brand trust.
关键成功要素
- Capitalized on the 1999 university enrollment expansion and the rising popularity of civil service exams, using a Peking University background and mentor resources to establish professional credibility.
- The direct-operation branch model ensured standardized teaching quality, covering 31 provinces and 440 branches by 2012, creating a channel barrier.
- Created a low-cost customer acquisition loop through book publishing, online traffic from Offcn.com, and offline in-person classes.
- Used the 'agreement class' financial design to lower the barrier for enrollment, rapidly scaling cash flow and size in the short term.
- Decisively cut agreement classes and compressed the branch network to about 800 locations after the crisis to restore operating cash flow.
Lessons
- Prepaid tuition is not profit; using student fees as a capital pool for expansion is essentially leveraging, and once refund rates spiral out of control, it triggers a run on the company.
- There is a severe mismatch between the cash flow characteristics of asset-heavy real estate investment and asset-light training businesses; assets that are beneficial in good times become burdens in bad times.
- Scale expansion in policy-sensitive industries must leave a safety margin for policy shifts; one should not mistake cyclical dividends for one's own capabilities.
- High-ratio share pledging by founders creates a 'death spiral' when stock prices fall, as margin calls threaten corporate control.
- Cutting financialized products and returning to standard pricing and real delivery is painful but is the only path to rebuilding trust.
Core Data
- 市值peak：2020 peak market cap of approximately 265.8 billion yuan (based on public data, independent verification not performed)
- 市值now：June 2026 market cap of approximately 14.5 billion yuan (based on public data, independent verification not performed)
- revenuepeak：2020 revenue of approximately 6.2 billion yuan (based on public data, independent verification not performed)
- 2025年营收：2025 revenue of 2.237 billion yuan, net profit of approximately 49 million yuan (based on public data, independent verification not performed)
- refundliability：Approximately 508 million yuan in outstanding refund liabilities at the end of 2025 (based on public data, independent verification not performed)
- debtratio：2026 asset-liability ratio of approximately 86% (based on public data, independent verification not performed)
- staff：Staff reduced from a peak of over 40,000 to approximately 7,300 (based on public data, independent verification not performed)
- branches：Direct-operation branches compressed from a peak of over 1,600 to approximately 800 (based on public data, independent verification not performed)
Competitors / Peers
Offcn Education's main competitors are Huatu Education and Fenbi. Huatu also started in the early 2000s and was known alongside Offcn as one of the two giants of the civil service exam market, but it was more restrained in promoting agreement classes and was less affected by the refund wave. Fenbi, which originated from the Yuanfudao system, entered the market with low-priced online question banks and systematic courses. It listed on the Hong Kong Stock Exchange in 2023 and used internet-based tactics to erode Offcn's offline market share. After Offcn's collapse, much of the vacated market share was captured by Fenbi's regional partner classes and smaller local institutions. By 2026, the focus of competition has shifted from branch density to AI-personalized learning and employment services; Offcn's launch of AI job training devices is an attempt to catch up with Fenbi's online efficiency model.