The three levels of conversion analysis: do you understand both product and operation?

Conversion analysis is the most core and critical scenario in the analysis of Internet products and operations . Taking shopping on e-commerce websites as an example, a successful purchase behavior involves multiple links such as searching, browsing, adding to shopping carts, modifying orders, settlement, and payment. Problems in any of these links may lead to the failure of the user's final purchase behavior. In the context of refined operations, how to conduct conversion analysis has become a science.

What is a conversion? A conversion occurs when a user performs an action in the direction of your business value point. The business value points here include but are not limited to completing registration, completing user information, completing a purchase, etc. Every major conversion includes several small conversion links. We usually use a conversion funnel to show this process.

1. Every step of basic analysis transformation needs to be polished

The basic stage of conversion analysis mainly involves the analysis of conversion steps and the monitoring of conversion rate trends. The figure below is a conversion funnel that shows the overall conversion effect and the conversion rate of each step, as well as the trend of the conversion rate of each step over time.

Conversion Funnel

Taking the registration process as an example, the information “total conversion rate 9.54%” is not very helpful for us to optimize the registration flow and improve the conversion rate. However, with the help of the funnel chart above, it is not difficult to find that the conversion rate from the first step to the second step is only 18.5%, which is significantly lower than the previous and next two links. Once we discover the problem, we can optimize the registration flow specifically to increase the registration conversion rate as efficiently as possible.

At the same time, real-time monitoring of each conversion rate can help us detect sudden problems in the product in a timely manner.

Conversion rate trends

One day, the conversion rate of the second step of the registration process dropped significantly, and this step was exactly the step of filling in the mobile phone verification code. After checking, it was found that the agent of the SMS verification code automatically stopped the SMS verification service due to arrears, so the service was restored to normal after timely recharge.

2. Advanced analysis and comparison of conversions in different dimensions

User experience is affected by many factors and directly affects conversion rate. In order to better improve the conversion rate, it is necessary to consider factors of different dimensions, including but not limited to the user's operating system, browser type, access source, operating platform, access source, etc.

GrowingIO multi-dimensional conversion analysis

Taking the user's browser as an example, we compared the conversion rates of different browsers one by one and found that the conversion rate of Safari browser is lower than the overall. After research, engineers found that the reason was that the website used a new Java architecture that was not compatible with the Safari framework, resulting in a very poor user experience in this environment and a very low registration conversion rate.

Common factors such as the browser user's operating system, PC or mobile terminal, access source, etc. may affect the conversion rate. The more advanced the product or the operator, the more detailed the considerations should be and the more they should polish the product from the details to continuously improve the conversion rate.

3. High-level analysis and multi-dimensional cross-analysis support continuous product iteration

The process of discovering problems often needs to be split many times. At this time, you need a funnel that supports multi-dimensional cross-analysis.

When an e-commerce website used the GrowingIO funnel to measure transaction conversion, it was found that the number of users on the APP was higher than that on the website, but the conversion rate was lower than that on the web page.

Multi-dimensional cross-analysis

From the specific steps, we can see that the conversion rate from user submission to payment is significantly lower than that on the web page. It is worth noting that users who have submitted orders have a strong desire to buy and are a group of users with great potential to be recalled. But they chose to go back to the previous step instead of paying.

By comparing the information structure of the payment pages on the website and APP, it was found that the payment page on the APP lacked detailed descriptions of the ordered goods, recipient address, contact information, etc., which caused many users to return to the previous step for confirmation and made users hesitate, resulting in a decrease in conversion rate.

The product manager then referred to the website’s information structure to supplement detailed information and at the same time tried to recall lost users during the payment process.

After optimization, the conversion rate increased significantly

The effect of optimizing the payment link is monitored from the trend chart of the funnel. The conversion rate from order submission on the APP to the payment link has been significantly improved, even slightly higher than the website conversion rate, and the overall conversion rate has also been increased. At the same time, users who are selected for recall in the funnel are used as target users to observe the changes in conversion rate after recall in order to evaluate the effectiveness of this recall activity.

Such subtle conversion problems are difficult to discover based on intuition alone. It requires product or operation personnel to have a high degree of data sensitivity and sophisticated business skills. This is also the performance of the advanced stage of conversion analysis. After discovering the problem, product optimization is carried out and then returned to the funnel to monitor the optimization effect. The product grows steadily in continuous iterations.

Advanced thinking and tools for conversion analysis

Improving conversion rates requires both a data-driven mindset and proficiency in certain data analysis tools.

First of all, data analysis tools are essential for lean operations. Currently, popular funnel analysis tools on the market include Google Analytics, Mixpanel, GrowingIO, etc. “If you want to do your work well, you must first sharpen your tools” is exactly what this means.

Secondly, you need to have a strong data-driven awareness and be familiar with the business. Conversion rate is not only a data indicator, its essence is a true reflection of user experience. As we continue to expand the dimensions of user experience analysis and deepen our thinking on our products and user behaviors , we are also making continuous progress in conversion analysis.