How much do you know about product conversion: 4 analysis models for high conversion

What is conversion

Conversion includes two things: conversion rate and conversion number.

1. Conversion rate

Let’s talk about conversion rate first. The official definition of conversion rate is the ratio of the number of expected behaviors to the total number of visits. Simply put, it is the following formula.

The formula is: Conversion rate = number of expected behaviors/total visits.

In other words, conversion rate is a ratio indicator of all desired actions.

2. Conversions

The conversion number is easy to understand, it is the expected number of behaviors.

Conversion can be a single link, such as login, click, and payment; it can also be a process consisting of many links, such as the e-commerce purchase process and the registration and card binding process for financial products.

Why do we say conversion instead of conversion rate? Leave it in suspense and you will see it later.

Why talk about conversion?

To give a simple example, we organized an activity, and after we started promoting it, we found that the effect was not very good. Although we used many channels and theoretically the user coverage was wide enough, the number of users who were actually attracted to the product and generated purchasing behavior was very small. So, at this time, you need to modify and improve the activity, correct the bad aspects of the activity, and make the activity have explosive power.

What is the general practice? Look at the activity again carefully, see what needs to be improved, and then change it. This is common practice, is there anything wrong with doing it this way? Yes, his mistake was that he made changes based on his subjective will, changing whatever he thought of without any method.

So what is the correct approach? First define the problem, apply common analysis methods for such problems, then break the problem down and optimize each subdivision one by one. In the case of the activity above, when the user coverage is wide enough, very few effective users are generated. It can be basically determined that this is a problem of low conversion. So how to solve the problem of low conversion? Oh, most activities have low conversion rates because of 1…2…3…. There are several ways to solve low conversion rates: 1…2….3….

What is the difference between these two solutions? It lies in the composition, completeness and details. The most terrifying thing when encountering a problem is not that you don’t know what to do, but that you just do what you want to do, only focusing on the points you are concerned about, forgetting the methods, and forgetting to understand things that you have never thought about but should know.

Back to the subtitle , why do we talk about conversion? The word conversion is used a lot, but have you ever thought about where to use it? Do you just look at the data, look at Baidu statistics, the penetration rate is very bad, and then you think you need to adjust the product, and then it’s gone, no matter how you change it. In the eyes of many people, conversion is just a piece of data, a piece of data that they just look at.

Why do we need to talk about conversion? What is the significance of talking about this? Define the problem.

When the user input of our product is large enough but the value of users is not high, we determine that this is a conversion problem, which means defining the problem, finding a way to analyze the problem, and finding a solution instead of letting it run wild.

What is the analysis and solution of conversion?

1. Funnel Model

Look at the picture below. If you already know this model, you can jump directly to the second model, or continue looking at it. There are three additional details.

Taking the search and shopping conversion of unregistered or non-logged in users as an example, the funnel model is the core process steps that users must go through in the shopping process. The number of users gradually decreases during the process. What we need to do is to reduce the user churn rate in each step.

The specific steps of using the funnel model are as follows:

1. Draw the process

You need to draw the core process of your product, as shown above.

2. List the impact factors.

Find out the detailed factors that affect each process, and search for keyword nodes as an example:

The impact factor of a search keyword can be:

• What else can be optimized in the search box itself?
• Is the operation smooth?
• How fuzzy is the search?
• Are the prompts in place?
• Are the search page recommendations reasonable?
• Can search history be further optimized?
• Can the search results be further optimized?

…

I have only listed the impact factors of the search keyword link. You can list more for each other link. What we need is to list the impact factors of each link.

3. Optimize factors one by one.

Find the factor and optimize it.

The funnel model is suitable for all conversion processes. When you receive a task, such as an activity that attracts a lot of new users or the product itself has enough visiting users, but the expected behavior conversion is not high, then this model can come in handy.

In addition, when using this model for optimization, pay attention to three issues in the product process:

First: Can the length of the funnel be shortened ? The fewer process nodes there are, the less likely there is to be loss.

Second: Can the order of model process nodes be adjusted ? Helps with product conversion or has other goals. Take the "Want to Go" website that I visited as an example. They want to improve user login data, so they move the login node forward and prompt users to log in when browsing.

Third: Pay attention to the disconnection of the funnel process . Be careful not to interrupt the entire funnel model. If users jump to other pages during the process to perform other operations, they are likely to be lost.

Mingdao collaboration software used email registration in the early days, and the registration success rate was only 40% (roughly this data, I forgot the exact number). After analysis, we found the problem: each email verification requires the user to jump to their personal email, and the email can only be verified after logging in. Since email login and verification is a relatively cumbersome process, it is also possible that the user forgets to register because of checking other emails. Later, email verification was changed to SMS verification. When operating in the current interface, the registration success rate increased to 80%.

For example, if a user is not logged in when shopping, but needs to log in to his account when checking out, it is best not to jump to the page at this time, and directly pop up the login box. Even if the page needs to jump, the user should jump directly to the checkout process to continue the checkout after logging in. This is very common.

2. Pain Point Path Model

Unlike the funnel model, which is based on process analysis, the pain point path is based on a single page analysis. It mainly records the visual path of users in the process of paying attention to pain points (I will call pain points and itch points collectively as pain points). The color depth in the visual path varies according to the degree of the pain points.

As shown in the figure below, many people may be familiar with it. This is a common heat map. Heat maps are often displayed based on the number of clicks.

The image presentation of the pain point path is slightly similar, but the difference is that the color from cold to warm represents the pain level of the pain point, and there is an additional visual path of attention, as shown below:

The pain point path is not a picture drawn intelligently by scale data, but needs to be manually drawn in combination with the user's pain points and visual trends. The red area represents the pain points that users are most concerned about, and the numbers and lines represent the visual path of the pain points that users are concerned about.

How to use the pain point path model?

1. Filter out the user pain points on the page and mark the pain levels of the pain points. Pain points can be divided into four levels according to their importance.
2. Draw the visual path that users take to focus on key pain points. As shown in the picture above, mark the visual paths of the pain points that users are accustomed to paying attention to and mark the order. Because the visual paths of different users may be different, the main visual path we focus on here is a visual path under comprehensive considerations.
3. Determine the optimization target based on the pain level of different pain points. Optimize pain points according to four levels: basic usability, satisfactory, very satisfactory, and exceeding expectations. If high-level pain points are met, higher-level requirements will be met. Low-level pain points only need to be usable, which has little impact on user decisions.
4. Optimize high-level pain points. Optimize pain points according to level requirements.
5. Adjust the visual path (optimize the location of pain points). If necessary, because the high-level pain points are placed lower or to the side, these pain points need to be mentioned in important positions on the page to reduce the inconvenience of user search, reduce decision-making time, and reduce loss.

The pain point path model and the funnel model, one focuses on the process and the other focuses on the page. Using the two in combination can achieve better optimization effects.

3. Channel conversion verification model

Anyone who has done new customer acquisition activities knows that for a product, in addition to quantity, there is another important requirement for attracting new customers, and that is precision. High precision means high conversion. Your product is for high-end users, and a suit costs several thousand yuan. If the people you attract are all students, you will only waste money on new customer acquisition activities.

What is the significance of the channel conversion verification model?

It is to compare the conversion efficiency of users attracted by each channel. The model is as follows:

In the above figure, assuming that the four channels have the same investment in funds and manpower, and can attract 100,000 users (for ease of understanding, simple integers are used), the conversion rate of channel 1 is 10%, and the conversion rate of channel 4 is 80%. Then, after comparing the four channels, what we should do is increase the investment in channel 4 and reduce the investment in channel 1.

Of course, this is considered from the perspective of conversion rate.

Another situation is as shown below (note the change in numbers):

Assuming that the four channels have the same investment in funds and manpower, Channel 1 can attract 1 million users and Channel 4 can attract 50,000 users. Although the conversion rate of Channel 1 is very low, only 10%, which is far lower than the 80% conversion rate of Channel 4, the number of users attracted is 2.5 times that of Channel 1. At this time, investment in Channel 1 should be increased.

The purpose of the channel verification conversion model is to compare the input-output efficiency of various channels, help you make the best channel selection data support, and make the input-output more efficient.

4. Growth/Goal Conversion Model

The growth/goal conversion model can be understood as a growth model similar to the points system . The difference is that in addition to being a user points growth system, this model can also be a target model for users within a certain time period.

Let’s talk about the growth conversion model first. As shown in the figure above, the goals and requirements for each level are defined. After setting up sufficient bait, users gradually convert from ordinary users to senior users. So in this transformation process, what we need to pay attention to is whether the rewards and requirements of each node are reasonable, and whether the bait or demand satisfaction between every two nodes is sufficient. Taking the common growth system of games as an example, in the initial stage of the game, various small tasks, various easily obtained red envelopes and various easy-to-achieve goals are often set to encourage new users to have more interactions and grow rapidly on the platform, thereby promoting retention and conversion into senior players. In the middle and late stages, the difficulty of user growth will be gradually increased to promote the payment conversion of game users.

The reason why the goal growth model, goal conversion model and growth conversion model are discussed together is that they have a common logic, which is to design multi-level goals, guide and induce users to achieve corresponding goals, and promote users to complete the established expectations of the product.

As shown in the figure below, this is the incentive policy implemented by Uber during the peak period of its early conquest of the domestic market.

Peak rewards. It can be seen from the reward policy that at the current stage of competing for private car drivers with several other competitors such as XX private car, the drivers' income for each order is at least twice the usual amount, and can reach 3 times during peak hours, which encourages drivers to compete for orders. With just a little effort during peak periods, you can double or even triple your income.

The benefit of rewards for completing orders is that it keeps drivers on the road forever. As long as they work a little harder, their income will go up to a higher level, so they will always have motivation. There are many more about Uber’s incentive policies, and I will just give a brief introduction here. Although Uber has now merged domestically, there is still a lot to learn from its former bravery, but that is beside the point.

Summarize

Therefore, when you encounter a situation where the user data is good but the value of the user is not high, or if you want users to use the product according to your product design, then first define this as a conversion problem, and then use the following four analysis models to enhance the value of your product.

• 1. Funnel Model
• 2. Pain Point Path Model
• 3. Channel conversion verification model
• 4. Growth/Goal Conversion Model

Hope it is useful to you!

Mobile application product promotion service: APP promotion service Qinggua Media advertising

The author of this article @水果 is compiled and published by (Qinggua Media). Please indicate the author information and source when reprinting! Site Map