Zuoyebang is committed to providing learning tutoring services in all subjects for primary and secondary school students across the country, and has independently developed a number of learning tools, including photo search, Zuoyebang live classes, composition search, etc.; Xuebajun focuses on online 1-on-1 tutoring, artificial intelligence, and photo search for primary and secondary school students. Tailor-made personalized learning plans to help primary and secondary school students find good teachers that suit them. The author of this article uses Xueba Jun as a comparison to analyze the product testing strategy of Zuoyebang.

1. Positioning of Zuoyebang

Zuoyebang is committed to providing learning tutoring services in all subjects to primary and secondary school students across the country and is a leading brand of online education for primary and secondary schools.

2. Started from the community, with a natural advantage in activity

The Zuoyebang project was incubated by Baidu and officially started independent operation in September 2015.

In the early stage of the project (before V4.0), it only supports students to post questions online and wait for netizens to answer their questions; or ask teachers one-on-one questions, and teachers provide professional video answers; at the same time, students can also answer other people's questions online.

There are question lists such as [High Reward] and [Difficult Question List] next to it, where you can earn virtual currency faster by answering questions. Virtual currency can be used to exchange for point cards, phone bills, etc. in the APP's built-in mall (the 10 in the upper left corner represents a 10 virtual currency reward).

When posting a question about homework, first select the age and subject category for which you are submitting the question. You can enter the question manually or submit a photo of the original question directly on the book.

Summary: In the early days of Zuoyebang, its main purpose was to solve the problem that students did not know how to write their homework and had difficulty getting timely answers. There were few online forum apps dedicated to students, which led to students' daily social activities being mainly offline, while it was difficult for them to find peers with similar interests online to discuss topics. It can be seen as a hybrid of the student homework version of "Baidu Knows" and the high student density of "Baidu Tieba".

At this time, due to the nature of the question-answering forum where strangers help each other, Zuoyebang already had the prototype of Baidu Tieba. Students can browse some "circles" of their interest, i.e. sub-forums, through the classmates circle interface, and post and reply in the sub-forums.

But there are also many problems:

• For example, due to the problem of mobile phone pixel or photography technology and environment, the picture may not be clear enough, which may affect the answer;
• There is no guarantee of the speed, accuracy, and completeness of responses to posts.
• There may be multiple questions being asked and answered repeatedly, resulting in a waste of resources;
• The wealth required for some gifts that can be exchanged for wealth is too high and somewhat useless. The gifts cannot provide sufficient incentives and are not enough to stimulate the response speed.
• The emphasis on community communication and the inconvenient homework questions also made this "learning" app somewhat off track.

Although the project team also introduced teacher video explanations, it is still a drop in the bucket compared to the number of students posting questions every day and the growing user base. At the same time, video explanations are not a good monetization model, which will be discussed later.

Image source: Baidu Index

As an early version of Zuoyebang, before the photo search function was launched, its overall search index showed a relatively stable and slow growth. It is worth noting that at this time, Xuebajun had already been online and had been polishing its photo-taking and question-searching function for a long time, while Zuoyebang was still at the primitive stage of posting questions.

Despite this, compared with Xueba Jun, which did not build a community during the same period, Zuoyebang still leads Xueba Jun in search index volume by nearly 3 to 4 times by relying on its community advantages.

The advantages that Zuoyebang, which started out as a question-and-answer community, has at this time also fully demonstrate the significant role that the core function of the community plays in promoting its specific activity indicators such as daily and monthly active users.

3. Seize the special period, accurately understand user needs and respond quickly

After adding the photo-taking and question-searching function on January 21, 2015, Zuoyebang experienced a brief explosive growth.

Comparing the time before and after the launch of the photo search function (i.e. before the release of version v4.00 on January 21, 2015), its Baidu Index quickly exceeded 10,000 from thousands and stabilized at around 2,500.

This is mainly due to several major updates after version 4.00 and before the release of version 4.3.2 on March 13:

1. Improve search accuracy - optimize important functions, improve user experience, and deepen the moat;
2. With the coming of winter vacation, more than 8,000 winter vacation homework answers have been added. Users can get the entire answer book by scanning the barcode on the back cover - accurately understanding the specific needs of users in a specific period of time and making full use of Baidu's existing data resources;
3. A new question reading function was added, which allows users to get accurate answers by just reading a few words. This function is designed to address the pain points that searching questions by taking photos is fast but subject to many constraints, the search results are unstable, and typing input is accurate but time-consuming and labor-intensive.
4. A new composition search function has been added, and the composition library has reached 8 million. Winter vacation composition is one of the required homework for primary, secondary and high school students. At the same time, its time-consuming and labor-intensive, high difficulty and lack of standard answers also mean that it cannot be easily searched out. The launch of the model essay search function by Zuoyebang at this time is an important update that deeply follows up on students' learning needs and quickly responds to the shortcomings of existing products.

Image source: Baidu Index

It is not difficult to see from the Baidu search index of Zuoyebang from January to May 2015 that after a series of actions taken by Zuoyebang, its active users not only experienced explosive growth in early March (near the end of the winter vacation, students began to do a lot of cram school homework and urgently needed standard answers and model essays for reference).

Although the index slipped slightly in the later period (the demand for homework answers and model essays after school starts is not as strong as during the winter vacation), it generally stabilized at 25,000 at the end of May, an average increase of about 14,000 compared with the average index in mid-February and a drop of about 15% compared with the average peak index of 29,328 at the end of February.

It is not difficult to see that Zuoyebang has converted most of the new users during the winter vacation into loyal users, and the user retention rate is extremely high.

Compared with Zuoyebang's quick response to the special question bank for students' winter vacation homework, Xuebajun, as an APP that targeted the photo-taking and question-searching function from the beginning of its launch, has already gained an advantage in the question bank after long-term data accumulation. From January to March, the major updates of Xuebajun are as follows:

1. New writing assistant function: you can freely combine a large number of sample essays to quickly solve your writing problems. At the same time, English composition materials are launched, which can solve the problem of Chinese and English composition with one click;
2. New practice function: Selected questions allow students to practice classic questions efficiently and improve their scores quickly;
3. Added text search function: to solve the problem of difficulty in searching when taking blurry photos;
4. Optimize search accuracy and application interface.

Image source: Baidu Index

It is not difficult to see that Xuebajun and Zuoyebang each have their own advantages in winter vacation updates and are equally matched.

However, by observing the data of Baidu Index from January to May 2015, it can be seen that Xueba Jun also reached its peak in early March, and then the average value stabilized at around 15,000 at the end of May, which was a decline of about 35% from its peak of 23,133 at the end of February.

Compared with Zuoyebang, Xuebajun has suffered a severe decline, with a large number of new users leaving and a high churn rate of new users - what is the reason?

In addition to major optimizations and bug fixes in April, Zuoyebang has also set its sights on another special period - the sprint to the college entrance examinations. Through the functions of "Secret Papers of Senior High School Entrance Examination", "Universal Materials for College Entrance Examination Essays", "Special Topic Updates for Senior High School Entrance Examination and College Entrance Examination" and "Comments from Famous Teachers", a large number of high-frequency updates are carried out specifically for candidates who are about to graduate.

Xuebajun, however, did not launch any special features or sections for that period.

It is worth noting that from the beginning of May to the beginning of July, the two indexes showed a trend of increasing and decreasing. This subtle change also proves from the side that:

1. Students have low loyalty to learning functional apps and may have two apps on their phones at the same time. Once students find other software that is more valuable during their studies, they may significantly reduce the use of another software or even uninstall it;
2. During special learning periods, students' demand for student aid apps is also showing an explosive growth trend. Learning to seize favorable opportunities is of great help in promoting apps and increasing user activity.

4. Relationship between core functions and additional functions

Another point worth noting is that before Zuoyebang released the "take a photo to search questions" function, Xuebajun's Baidu Index once surpassed Zuoyebang.

Looking back at the updates of Xueba Jun from the end of 2014 to the beginning of March 2015, in addition to the necessary updates for the winter vacation, Xueba Jun focused on the "optimization of photo search" function; while Zuoyebang tended to "supplement social functions", such as adding "people nearby" (making it easier to find and add friends), adding voice function to private messages, and adding task boxes to bind users to promote activation.

Although Zuoyebang is aware of the various drawbacks of posting questions to search for questions and has launched the "Ask the Teacher" function to remedy the situation, it is still difficult to conceal its sluggish growth trend compared to Xueba Jun.

This also proves once again that learning function APP must take "optimizing core functions" as the basic foothold for product development. Once the core functions are inferior to others, no matter how good the additional functions are, they will be of no avail.

But considering the whole picture, are additional features just icing on the cake?

——Not necessarily.

Looking at the starting points of Xueba Jun and Zuoyebang, Xueba Jun mainly focused its updates on "solving questions that users don't know how to answer", and based on this, it launched a series of functional explorations, such as "take a photo to search for questions", "online Q&A" and "built-in question bank" (there is no need to search for similar questions, it essentially solves a series of such problems), and continuously optimized related functions.

Over time, although product functions have been effectively improved, users' stickiness to the platform is not high, and there are not enough strong connections (such as friends on the platform, daily tasks, and timely access to fresh and useful learning resources). Once alternatives with similar or even better core functions emerge, the reasons that support users to continue to stay on the platform are challenged, and user loss is inevitable.

In order to cope with this challenge, products that use core functions as their main selling point need to continuously deepen their moats, maintain and update core functions, and put a lot of resources into building core functions, forming a vicious cycle:

Core functions are no longer top-notch - A lot of resources are used to maintain core functions - Additional functions are available with limited resources and difficult to develop quickly - Core functions are optimized - A large number of competitors update and optimize core functions - Core functions are no longer top-notch...

At the same time, as a new search function in 2015, taking photos to search for questions had certain technical barriers at the time.

However, with the advancement of science and technology and the accumulation of previous experience, the technical barriers of "taking photos to search questions" are becoming thinner and thinner, and it is gradually transforming from a "technological product" to a "database product". Companies no longer compete on whether they can take photos to search questions, but turn to competing on search speed, question bank capacity, etc.

As a result, the question banks of various softwares that use photo search have been continuously expanding and can already meet the vast majority of needs of the vast majority of students. Photo search products are once again becoming homogenized.

On the one hand, it is difficult for students to judge which APP has the most powerful question-searching function based on the photo-searching application interface alone, and students need to conduct in-depth testing for a long time; on the other hand, the performance of most current photo-searching APPs is already sufficient to meet daily needs, and the relative advantages of each in photo-searching are not obvious, so APPs with photo-searching as their main selling point will inevitably be lost in the crowd.

To sum up in one sentence: Xueba Jun’s success is due to taking photos to search for questions, and his failure is also due to taking photos to search for questions.

They mistakenly set functions with low future technological barriers as the first goal of product replacement, thus falling into a vicious cycle; they also failed to keep up with user needs in special periods, and were complacent, always focusing their product updates on "solving problems that users don't know how to solve."

While Zuoyebang, Xiaoyuan Search Questions and others launched traffic pool monetization models such as live courses and achieved good development, Xueba Jun still stuck to its positioning as a "question search software" and the eliminated "1v1 video online question and answer" monetization model, and eventually fell behind the industry leader and was forced to adopt a contraction strategy.

Author: Don't worry, you are right

Source: Don't worry about Hong Kong