Data Operations: How to use data analysis to improve conversion rates?

As the Internet enters the second half of its development, the traffic dividend is gradually fading and the cost of acquiring traffic is getting higher and higher. Under the same conditions, it is urgent to improve the traffic conversion rate. Especially in certain fields, such as e-commerce and financial management, conversion rate is one of the important indicators to measure the quality of products.

How to improve the conversion rate of traffic is an old problem that has always plagued Internet companies. Therefore, it will be of great benefit to marketing, operations, product and data analysts to see through the secrets of conversion rate. Today we will talk about how to use data analysis to improve conversion rates.

Product conversion rate refers to the ratio of the number of users after using a single interactive behavior and overall function in the product to the number of users before using it. We usually call the former behavioral conversion rate and the latter overall functional conversion rate.

Behavior conversion rate = expected number of behaviors / total number of behaviors

Overall functional conversion rate = behavior conversion rate 1 * behavior conversion rate 2 behavior conversion rate 3 * … behavior conversion rate N

Expected behavior is the behavior that you want the target to do, such as reading an article, completing a registration, binding a bank card, successful payment, etc. In the click-through rate, "click" is the expected behavior, and in the forwarding rate, "forward" is the expected behavior. Similarly, there are download rate, activation rate, purchase rate, open rate, transaction rate, repurchase rate, etc.

Because the total number of people is fixed within a certain range (constrained by the size of the activity and the channel), we must try our best to increase the number of expected behaviors to increase the conversion rate (Conversion Rate Optimization).

The essence of conversion analysis is to promote the circulation of a company’s core business and enhance its profitability. for example:

Sales = Traffic conversion rate per customer

If you need to increase your company's sales, there are three ways: 1. Spend money on promotion to expand traffic; 2. Optimize the registration and purchase process to increase conversion rate; 3. Increase the average order value through methods such as full-discount and bundled products.

It is obvious that increasing the average order value is less feasible in the short term, but adjusting product pricing is a complicated task. A substantial increase in the average order value in the short term will easily lead to a decline in market competitiveness. At the same time, the brand and reputation will also be negatively affected, causing irreparable losses. As for expanding traffic, if the company has sufficient marketing budget, it is feasible to spend more money on promotion. However, for many companies, if they are not in the growth or mature stage of the product life cycle, the strategy of burning money is not sustainable. Therefore, the most cost-effective way for enterprises is to improve the conversion rate.

The benefits of improving conversion rate are: ① You can achieve better results with the same amount of money and increase ROI; ② Optimize the conversion process to improve user experience and brand satisfaction; ③ The optimized conversion process can play a long-term role and is sustainable.

Marketers use Data Analytics to create a registration conversion rate dashboard

Since conversion rate has so many benefits, how do we analyze it? Conversion rate is a scientific data that can be quantified and proven, and has the following functions: 1. Measure the strength of user demand; 2. Compare the benefits of two functions and optimize similar functions; 3. Verify channel quality and partition channel weight. So we can analyze the conversion rate in a data-based way.

For example, a social app has developed a new function, but the conversion rate of this function is very low. At this time, we can analyze it from two aspects: ① The conversion rate of the demand itself is low. At this time, we need to reflect on whether this demand is a strong demand, as well as its breadth, frequency of use, urgency, substitutability, etc.; ② Product design defects, inappropriate or unclear design expression, problems with entry location, color, copywriting, guidance, interaction process, etc.

When it comes to conversion rates, the funnel model is essential. There will be losses at each layer in the funnel, and the funnel model can help us reduce the amount of losses in the entire process. It should be noted that a single funnel model is meaningless for analysis. We cannot evaluate the conversion rate of each step in a key process of a website from a single funnel model, so we must analyze the conversion rate of each step in the process through trend, comparison and segmentation methods:

(1) Trend: Analyze the changes in the time axis, which is suitable for monitoring the effect of improving or optimizing a process or a step in it;

(2) Compare: By comparing the conversion rates of purchase or usage processes between similar products or services, problems in certain products or applications can be discovered;

(3) Segmentation: Segment sources or different customer types based on their conversion rates to identify some high-quality sources or customers. This is usually used to analyze the effectiveness and ROI of a website’s advertising or promotion.

To improve conversion rate, we need to have a data-driven awareness. 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 our user experience analysis and deepen our thinking about our products and user behaviors, we are also making continuous progress in conversion analysis.

Here we first discuss what kind of conversion rate is considered qualified? In fact, the conversion rates of products of different categories and unit prices can vary greatly. There are many factors that affect the conversion rate, such as operating methods, product functions, user experience and interface design, and brand trust. These factors will jointly determine the final conversion rate. Today we mainly share the general steps, and in the next section we will talk about specific cases.

1. Define your core goals

It is very important to clearly define the core goals of anything you do. Unclear goals will lead to unclear behaviors. For example, do we need to increase the repurchase rate or the average order value? Or to increase profit margins or sales? Generally speaking, the definition of goals should be determined according to different product life cycles, because the core goals you should focus on are different at different stages of the product. The core goal can be broken down into relevant indicators. Improving relevant indicators will also improve the core goal.

2. Draw the core process

This step is a critical step, and it is also the step that many people tend to overlook. The funnel model of each product is different. How to build a funnel model depends on the core flow chart of the product. The core flowchart sounds simple, but it is difficult to do because you have to consider a lot of details. However, drawing a product flowchart with better user experience can give you inspiration for process optimization. In addition, with the process in place, clear paths and directions make it easier to control and analyze during execution and review.

Business process diagram of Xianyu

3. List the impact factors (add them)

List all the factors you can think of that might affect user changes in the process and put them under the process. Enumerate all the factors you might think of, such as simulating the entire user process yourself, writing down everything you can think of from the generation of the idea to the end of the process, and then trying to simulate the experience of competing products, write down your ideas, etc.

4. Add influence weight (subtract)

Each factor has a different impact on the results, and you need to express the differences between them in a data-based way. Data analysts can use technology to find the weights of factors, but if we don’t have the technology, we can refer to the following simple and easy-to-use methods to find the weights.

5. Optimize impact factor

After adding the influence weights in the previous step, you can sort the influence levels to see the most influential factors. Then you can optimize them one by one. for example:

(1) Target the right users

Different departments have different understandings of what the correct user is. Market operations are carried out in various channels to attract users through products, services and content, but there are certain differences in the sources of users from different channels. For users in different channels, the market needs to develop different push notifications and different landing pages to improve targeting; product managers should also be concerned about user classification, and develop different operating strategies for different user groups based on user behavior on the product, and optimize different product features; for sales staff, finding the right audience and decision makers based on the results of communication with users is also a process of finding the right users.

(2) Improve product experience and fluency

There are many ways to improve user experience, such as adjusting button positions, interface colors, or making overall revisions, product interactions, etc. Details determine success or failure. However, it should be noted that the product process may be relatively simple, but the user's mental process is relatively complex. At this time, when writing the process, you can also add the user's mental process. The optimization process is an extremely long one with no time limit. Keep a curious mind, learn more, expand your perspective, figure out the cause of the problem, and talk more with users.

(3) Optimize conversion paths

Different industries have different conversion paths. E-commerce, OTA, and O2O industries: the final number of orders is considered a conversion; SaaS and social products: the number of registered users is often used as a conversion point. We need to understand what steps a user goes through and ultimately completes these conversions. The conversion process provides users with sufficient decision-making information. It should be noted here that a conversion path that is too long will lead to user loss.

1. What is a landing page?

A landing page, also known as a landing page, is a landing page where users land after being attracted by an advertisement and taking action. It is a link in the marketing funnel. The advertisements here refer to any form of media that attracts users to the landing page, such as Moments, elevator ads, search engine ads, AppStore introductions, event promotion pictures, etc. Any landing page has one and only one goal, which is to develop users, facilitate transactions, and collect leads.

2. How to improve the conversion rate of landing pages

(1) Value proposition optimization

Optimizing the value proposition is to make the content of the landing page more consistent with the user’s perception, making the user more willing to respond to your call to action. For example, a home improvement platform found that the conversion rate of the advertising landing page had not been able to be improved. In order to obtain more user registrations from search engines and increase the click-through rate and registration conversion rate on the landing page, it developed an optimization plan and conducted a 15-day trial with the help of A/B testing. The initial traffic started at 10%, and the traffic distribution was gradually adjusted based on user data feedback.

In the later stage of the test, we obtained data on the impact of different versions on the registration conversion rate. Compared with the original version ("Are you afraid of tricks in home decoration? Learn the first three steps of home decoration to save you money"), the number of registered people increased by 27% and the registration conversion rate of the page increased by nearly 30%, thus finding a better solution to promote sales.

(2) Page clarity optimization

Content and design are the two primary aspects that affect clarity, so we need to optimize the landing page information hierarchy, visual flow, pictures, copywriting, CTA (user call to action), etc. For example, after analyzing the landing page of a certain Internet finance platform, it was found that the original version had many problems with lack of clarity, such as the CTA was not easy to read, and the defects in the graphic layout interrupted the visual flow.

The optimization solution improves the visual flow of potential users on the page. The left-right structure of the product is adjusted to an upper-lower structure, the product name is weakened, the "yield" that users are most attracted to when purchasing financial products is strengthened, and the CTA button is enlarged and highlighted to attract user attention. In the end, the conversion rate of the optimized version was significantly higher than the former one.

Image from: ew design consulting

(3) Avoid user distraction

When we design, we are often asked a question: Why do you use this element in the design? Sometimes, in order to make the picture fuller, designers use some elements for no apparent reason, which makes the picture even more cumbersome. In the design industry, it is easier to add than to subtract. Some designs may look simple but they can actually reflect a designer's skills.

Therefore, we should simplify the content and design, and focus on user experience: do not burden users with operations, reduce the steps that make users wait and think, especially for some important and complex functions, such as filling out registration information forms, payment processes, etc. Simplification is one of the effective ways to reduce user churn.

(4) Reduce anxiety

What issues will users worry about during the conversion process? Are they afraid of privacy leakage? The product is not easy to use? Website loading slow? Lots of steps to complete? Don’t trust brands and payments? For example, an airline’s SMS link landing page. After you book a ticket, you will receive a ticket issuance notification SMS. You can see your flight information on the landing page and choose to add value-added service products, such as purchasing baggage allowance or meals. Before you purchase value-added services, you need to go through two steps to verify your information. If the system can identify the mobile phone number of the device that opens the short link as the mobile phone number of the person who made the reservation, these two steps can be reduced.

On the one hand, this reduces the effort required by users to complete the entire operation, and on the other hand, it eliminates the user's doubts about whether they already have an account. Therefore, we need to improve the usability and ease of use of the product, reduce user effort, and increase credit endorsement to gain user trust.

Image from: ew design consulting

(5) Create a sense of urgency

The sense of urgency is reflected in creating timeliness and scarcity, and is manifested in the form of limited time or limited quantity. Generally speaking, limited quantity is better than limited time. Research has found that consumers will have stronger impulsive purchases when faced with quantity restrictions, because when consumers find themselves competing with others for the same product, they will have stronger behavioral motivations, thereby giving the product a higher purchase value and forming a stronger impulsive willingness. Moreover, quantity restrictions mean that the number of products is limited and only a certain number of people can own them. Quantity restrictions can make consumers want to be different from others. They will have a stronger psychological ownership of quantity restrictions for the sake of the "feeling of being different."

Event countdown, through content design, we can enhance the sense of urgency of potential users. For example, limited offers, seasonal reminders, and even the tone of your ad copy can influence their sense of urgency. The above analysis combines the six optimization rules in the LIFT model. However, the rules are not independent of each other, but complement each other. If you can make good use of these rules in optimizing landing pages, you will get twice the result with half the effort.

(6) User matching

The landing page is the starting point for users to come into contact with the product. Users from different channels enter the landing page, and user conversion rates vary due to different channels and user characteristics. For example: users from Baidu and 360 have different conversion rates after entering the official website, because the characteristics and needs of users from different channels are different. Therefore, we often make corresponding plans for users from different backgrounds, such as different advertising media, different advertising sources, different cities or regions, so as to provide different landing pages for each segmented group of people.

We can also use big data to label and group users according to their behavior and have an understanding of the characteristics of each user group. Sort out the core behaviors of users, set weights for each page visited by users in the process, establish user purchase intention indicators, distinguish user values, understand the purchasing preferences of users at different levels, and conduct group operations, thereby continuously promoting the improvement of conversion rate.

The specific methods to improve conversion rates for different products in different industries are different, such as how to optimize the product conversion rate of e-commerce stores, how to improve the App registration and download conversion rate, how to improve the conversion rate of microcredit products... But the core is still the understanding of the business itself and the users. Only by understanding users in a targeted manner can we develop different optimization plans for different businesses. Secondly, by proposing hypotheses through data analysis, conducting iterative experiments in small steps, testing the hypotheses and conducting review and feedback, the conversion rate can be quickly improved.

Source: DataHunter (DataHunter01)