4 analysis models for high-conversion products!

Today, I would like to talk to you about the issue of "conversion".

Through communicating with my friends, I found that many people may have misunderstood the issue of conversion. Therefore, it is necessary to share with you the specific analysis methods and ways of thinking.

For example, we planned an event. Although we mobilized many channels and the user coverage was theoretically wide enough, the effect was not good. Very few users entered the product or made purchases.

How should we solve this kind of problem?

Many people will review the activity, see what needs to be improved, and then correct it. This is a common practice among many people. Is there anything wrong with doing this?

The problem is that it makes modifications based solely on one's own subjective opinions, changing whatever comes to mind without any rules to follow.

In fact, the scary thing when you encounter a problem is not that you don’t know what to do, but that you just do whatever 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.

So what is a reasonable 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 above activity case, the user coverage is wide enough, but very few effective users are generated. We can basically determine that this is a problem of low conversion.

So how to solve the problem of low conversion? Well, most activities have low conversions because of ABCD... There are several ways to solve low conversions: ABCD...

Defining the problem means finding a way to analyze the problem and a solution, rather than personal subjective assumptions.

So back to the beginning, why talk about conversion?

In the eyes of many people, the word "conversion" is just a data. Have you ever thought about where and in what process it should be used? We just looked at the data and found that the conversion rate was very low, so we thought we needed to adjust the product, and then that was it. We had to change it anyway.

No matter what work we do, we should follow rules and regulations instead of making blind guesses.

Next, we will list several models to try to help you analyze the conversion problem in detail.

1. Funnel Model

Many people know this model. The funnel model is a linear process. From the beginning to the end, users will lose data at every stage, just like a funnel.

The above picture shows the process of unregistered/non-logged in users searching for products and forming shopping conversions. As the process progresses, users will churn at every stage.

What we need to do is to reduce the user churn rate at each step. Let’s take a look at the steps of using the funnel model:

Step 1: Describe the process

Just like the picture above, draw the core process of your product conversion.

Step 2: List the influencing factors

Let’s take the “search keyword” node as an example. Its factors can be:

• ●What else can be optimized in the search box?
• ●Is the operating logic smooth?
• ●What is the degree of fuzzy search?
• ●Are the prompts in place?
• ●Are the search page recommendations reasonable?
• ●Can the search history be further optimized?
• ●Can the search results be further optimized?
• ●……

Only take the "search keywords" link as a reference. You can continue to give examples and dig deeper into other links. What we need is to list the influencing factors of each link. The more detailed the better.

Step 3: Optimize each factor one by one

After finding these influencing factors, we need to optimize them one by one to ensure a smooth process.

The funnel model is applicable to all conversion processes. When you plan an activity and attract a lot of new users or increase the number of users visiting the product, but fail to achieve the expected conversion effect, this model can come in handy.

When using this model, if we can think about these 3 questions, it will help you achieve better results:

(1) Can the funnel process be shortened?

The fewer process nodes there are, the less likely it is that there will be loss.

(2) Can the order of model process nodes be adjusted?

Adjusting the order can sometimes also help achieve product conversion. Let me take the "Operation Xiaokaxiu" website as an example. If I want to improve the user login data, I will move the "login" node forward and prompt the user to log in when browsing.

(3) Is there a gap in the funnel model?

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

Many products use email registration, but email verification requires users to jump to their personal email and log in to the email before verification. Because email login and verification is a relatively cumbersome process, it is easy to cause users to jump out.

If we change email verification to SMS verification, so that there is no need to jump to the page, will the registration success rate be improved?

Therefore, if the user is shopping without logging in and needs to log in to the account for checkout, do not jump to the page at this time, but 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 completing the login.

2. Focus Path Model

Unlike the funnel model, which is based on process analysis, the focus path model is based on a single page for analysis.

The focus path model mainly records the visual path of the user's attention points, pleasure points, and itch points. The color depth of the visual path varies according to the degree of attention.

As shown below, this is a common heat map:

Heatmaps are similar to interest path images.

The heat map uses the number of clicks as a statistical display, and the colors from cold to warm represent the frequency of users' clicks.

The color of the focus path image, from cold to warm, represents the user's attention level. These focus points can be combined into a visual path, as shown below:

It can be seen that the focus path is not drawn through data intelligence, but needs to be manually drawn in combination with the user's focus and visual trend. The red area represents the pain points that the user is concerned about, and the numbers and lines represent the transfer path of the user's focus.

So, how to use this attention path model?

① Filter out the areas on the page that users will pay attention to.

② Draw the visual path. Like the picture above. Of course, the visual paths of different users may be different, so we only focus on the main visual path, which is a visual path under comprehensive consideration.

③ Mark the degree of user concern (itch points/satisfaction points/pain points) and determine the level of optimization target area. For example, high-level pain points should meet higher-level requirements, while low-level pain points can be used and have little impact on user decisions.

④ Adjust the visual path. If you find that higher-level focus points are placed lower or to the side, you need to adjust them to more important positions on the page to reduce user search costs, decision-making time and loss.

3. Channel conversion model

For the product, when we do new user acquisition work, in addition to the requirement of the number of new users, we also have to be "precise", because high precision means high conversion.

If your product is aimed at high-end users, a watch costs several thousand yuan, and the users you attract are all students, you are actually wasting money on new customer acquisition activities.

The channel conversion model can help you verify the effectiveness of the channel very clearly.

As shown in the following figure:

We assume that the four channels can attract 50,000 users with the same investment of funds and manpower. From the perspective of conversion rate, the conversion rate of channel one is 10%, and the conversion rate of channel four is 50%. By comparison, we should increase the investment in channel four and reduce the investment in channel one.

Let's look at another situation:

Still using four channels, with the same investment in funds and manpower, Channel 1 can attract 800,000 users and Channel 4 can attract 50,000 users. Although the conversion rate of Channel 1 is very low, only 10%, which is not as good as that of Channel 4, the number of users attracted by Channel 1 is 3.2 times that of Channel 4. At this time, investment in Channel 1 should be increased.

The purpose of the channel conversion model is to compare the input-output efficiency of various channels and help you make the best channel selection. The prerequisite is that there needs to be data support.

4. Growth Conversion Model

The growth transformation model is similar to the points system growth model. The difference is that it can also be the user's target model within a certain time period.

We need to define the goals and requirements for each level, and after setting sufficient rewards, users will gradually transform from ordinary users to senior users.

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 interact more and grow rapidly on the platform, promote retention, and convert them 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.

Author: Operation Xiaokaxiu

Source: Operation Xiaokaxiu