Hujiang case study

SkyGlue helps optimize their signup form conversion rate.

Company Bio was born in 2001. It is the pioneering digital platform in online education in China. Hujiang specializes in online education and is dedicated to providing the most effective interactive education platform and the most comprehensive resource to self-learners. Hujiang has over 2 billion users and millions of members. Hujiang is the largest foreign language education platform in China.


No one questions the importance of signup form in the online conversion process, especially when you are spending money on PPC advertising to acquire new customers. Increasing the conversion rate directly translates to low cost and saving marketing dollars.

Let’s use as an example the registration process of a new user who is brought to the website through external advertisement. There are three possible outcomes:

• He left without doing anything.
• He tried to signup, but dropped out in the middle and did not complete the registration.
• He successfully registered an account.

The first scenario is called a bounce. If this happens a lot, it means either the message on your advertising does not match the content on your signup form or your ad is targeting the wrong audience. Even the load speed on your website could be a factor as well. Because visitors left the site right away, they did not leave any trace for you. To reduce bounce, we need to adjust the ad content and ad target.

For the second scenario, where a user tries to register, but for some reason abandons the form before completing the registration, we need to analyze the click path of these people and find out why they left. Since SkyGlue allows us to analyze individual user behavior and logs everything a visitor does, we used SkyGlue to perform this analysis.

Let’s take a look at the registration form first. Here is the url:


Users need to enter: username, password, email address etc. to complete the registration process.

We need to analyze the following:

How many users attempted to register, but aborted the process early?

Which input field is the most confusing and is causing registrants to give up?

First, we need to create a segment and find the registrants who aborted the registration process.


“360bulo-p” identifies the traffic from one of our advertising sources, after users have successfully registered, they will be automatically redirected to a page called “reg_active”. By excluding traffic re-directed to reg_active page, we can find the failed registrants. Since SkyGlue uses “Event Category” to store user actions, the third condition includes the visitors who performed actions on the form.

After creating the segment, we went ahead generated the “User Drill-down by Time” custom report SkyGlue pre-configured .


This custom report added two filters and filtered out actions users performed outside of the registration form. From this report, we found there are 349 visitors who attempted to register, but eventually failed.


Event Label is the visitor identifier. Total Events are the number of clicks. Each click action on any field on the form is counted as one event.

We exported the data into Excel. Here is the statistics based on the number of actions/clicks


We can see from the stats that more than half of the visitors only tried on the 1st and 2nd form field, then left. Only 1/5 of the visitors (68) worked hard to complete the form but still gave up in the end.

We examined closely a user who had 6 click actions on the form but still failed:


He came to the form, selected a language, then typed in username, password. He tried to input his username multiple times, but left in the end. The reason is because if someone has typed in a username that’s already existed in our system. We do not allow this username to be registered again. This obviously caused confusion for this visitor.

Next, we wanted to find out how many visitors who had good intention to register but got stuck on selecting proper usernames.

We went back to the “Users drill down by Time” report. We selected sub-category: Event Category. Then defined the actions as: “Click::Form user_name::” with clicks greater than 2. We consider a visitor being confused by the signup form if their actions are greater than 2 on form field “user_name”.


Below is the finding:


There are a total of 20 people who were confused by the signup form. When we work in this number with the previous statistics, the ratio is 29{40d197c9430f88f907d4c0a4eefc4c9a7dcebee593d7a5e3156d69ac601435ba} (20/68). So, about 1/3 of our clients’ encountered confusions when trying to complete the registration form. The rate is quite high.

(Below are the dropout rates each of the form field)

Input Field Element ID Visitors Ratio
User Name Click::Form::user_name:: 20 29.41{40d197c9430f88f907d4c0a4eefc4c9a7dcebee593d7a5e3156d69ac601435ba}
Password Click::Form::password:: 10 14.71{40d197c9430f88f907d4c0a4eefc4c9a7dcebee593d7a5e3156d69ac601435ba}
Confirm Password Click::Form::INPUT::re_password:: 3 4.41{40d197c9430f88f907d4c0a4eefc4c9a7dcebee593d7a5e3156d69ac601435ba}
Email Click::Form::INPUT::mail:: 8 11.76{40d197c9430f88f907d4c0a4eefc4c9a7dcebee593d7a5e3156d69ac601435ba}
Verification Code Click::Form::INPUT::vali_code:: 5 7.35{40d197c9430f88f907d4c0a4eefc4c9a7dcebee593d7a5e3156d69ac601435ba}

The form clearly identifies that username, password, email, verification code are the fields that created the most confusion. We gave the stats and feedback to our product development team and they improved the signup process.


1 Immediately after a user moves mouse away from the field, we check the username eligibility and provide hints to recreate a new username if needed

2. While a user is entering the password, we include strength hints for the password

3. Email field automatically populates the email services. This is an added convenience and avoids typos.

4. For the verification code, we add hints telling people upper case and lower case does not matter.

We deployed the changes on 8/31/2012. Compared with the statistics on the 30th, we found success registration rate went up from 5.94{40d197c9430f88f907d4c0a4eefc4c9a7dcebee593d7a5e3156d69ac601435ba} to 7.88{40d197c9430f88f907d4c0a4eefc4c9a7dcebee593d7a5e3156d69ac601435ba}, a 24{40d197c9430f88f907d4c0a4eefc4c9a7dcebee593d7a5e3156d69ac601435ba} boost. With the advertisement investment staying the same, the conversation rates increased close to one-fourth. This is a very significant achievement. We will continue to use SkyGlue to optimize our signup process and improve conversation rate further.



SkyGlue allowed us to analyze individual user behavior and helped us to optimize the process of our signup form by finding potential user experience problems/issues and boosted our conversion rate. The changes brought significant benefit to our business. We plan to continue to use SkyGlue in other product analysis and improve our site visitor experiences as a whole.