Interface News Reporter | Chen Zhenfang
Zhou Yuan, founder and CEO of Zhihu (NYSE: ZH), once had three wishes: traveling around the world, filming animal documentaries, and building a school. A senior told him, "Everyone's problems can be attributed to education." This planted a seed in his heart.
"Community and vocational education are two forms of Zhihu. Vocational education complements the other half of Zhihu's entrepreneurial vision and forms the 'complete body' of Zhihu. On January 11, Zhou Yuan announced at the 2024 Zhihu Education Conference that the vocational education brand 'Zhihu Zhixue' will officially operate independently.".
The vocational education business has become the second largest growth curve for Zhihu's commercial monetization. According to Zhihu's 2023 third quarter report, the revenue of vocational education business reached 145 million yuan, a year-on-year increase of 85.6%, accounting for 14% of the total revenue.
Zhihu's vocational education originated from community Q&A. Zhang Rongle, Senior Vice President of Zhihu and CEO of Zhihu Zhixue, revealed that as of the end of 2023, the number of views of vocational education related content on Zhihu has exceeded 10 billion.
Since 2019, Zhihu has been involved in vocational education for four years, and has established a vocational education business architecture system through various means such as joint operation, self operation, and external investment. By the end of 2022, the Zhihu Zhixuetang brand and app will be launched. On January 11, 2024, Zhihu Zhixue officially became independent in operation.
As of now, Zhixuetang has gathered six sub brands, including Quality and Vocational Education, Puppet Education, MBA Master, Tongyi Public Examination, Tongyi Teacher Examination, and AGI Classroom, covering more than 30 categories of courses, such as postgraduate entrance examination, public examination, teaching resources, CPA, CFA, MBA, and workplace skills. It has provided vocational training to more than 20 million students in total.
Zhang Rongle introduced that among the more than 30 sub category courses, self operated business accounts for one-third, mainly concentrated in the skills and interests section, and the course system is relatively similar to the teaching system of Zhihu Live. And certification courses are developed through investment, mergers and acquisitions, and cooperation.
On January 11th, Zhou Yuan stated at the meeting, "The entire (vocational education) industry is still in the development stage before the real housing supply of Lianjia. There are famous teachers in advertisements, but it was discovered during class that there were none at all. After paying for classes, the institution ran away in just two days."
During a media group interview, Zhang Rongle also stated, "The vocational education industry is in a mixed stage. Due to low efficiency, low threshold, and low level of digitalization, there have been no large-scale comprehensive education platforms in the industry."
"The vocational education industry is in a critical period of transformation, with digitalization and platformization becoming two major trends. The AI big model will reshape vocational education, and comprehensive education platforms will usher in a period of high growth." Zhang Rongle pointed out.
Last April, Zhihu and Facewall Intelligence jointly developed the Chinese large model "Zhihai Map AI". Both sides are promoting the application of cutting-edge technologies such as big models and AI agents in vocational education scenarios.
Firstly, through digital empowerment, improve the vocational education service system and delivery standards. Secondly, through open collaboration and integration of industry and education, we aim to promote high-quality employment in vocational education.
Zhang Rongle revealed that Zhihu Zhixue is exploring the digitization and AI ization of the entire process of job selection, question answering, and homework grading. In the next one or two years, Zhixue may achieve the full coverage of AI models in vocational education.
The implementation of the universal large model still faces difficulties.
"For the use of the business side, there is a significant gap (gap) in the universal big model." Zhang Rongle said, "The application cost of the universal big model is very high, and the cost of a single retrieval cannot cover the actual business benefits, and even exceeds the cost of labor. This is also the reason for the low utilization rate of the universal big model."
At the practical level, a lot of small models need to be developed to reduce the overall cost and achieve an effect of 80-90%, rather than directly empowering the business with universal models.
In terms of data, the bidirectional interaction and training between users and models in big model data are still insufficient, and more data is needed for training.
"We can see that the universal big model cannot solve the problems of multiple enterprises. It must be an industry level, even a Pass level solution, and even an exclusive model to better solve business problems," said Zhang Rongle.
The reason why the digital transformation of vocational education is slow is because the chain from service to customer acquisition and sales is long and complex. After the emergence of large models, it helps to achieve personalized changes in education, and the industry is in the early stages of technological change.
In the past few years, young people have faced difficulties in finding employment, tightened spending, and many vocational education institutions are facing closure.
Zhang Rongle told Interface Education, "In the past period of time, users' sensitivity to prices has been changing. The course is clearly moving in two directions, with lower and higher unit prices performing better, and the target audience for middle unit prices is not particularly clear."
He also mentioned that the labor cost of the vocational education service system accounts for a high proportion, which is also the reason for the expensive course prices. The best way to reduce labor costs is through AI. In the future, low-priced courses will be more user-friendly, while high priced courses will pay more attention to personalized needs and service requirements.
At present, Zhihu Vocational Education has not entered the offline stage on a large scale and is still in the exploration stage, without clearly stating specific offline business goals.
"Focusing on the future, vocational education must be closely integrated with offline education, and there are also many possibilities at the technical level, such as the iteration of AR technology. We pay more attention to the use of technology in the entire scene," said Zhang Rongle.