How to use data to drive operational growth?

With the advent of the data age, the previous extensive management is no longer in line with the trend. We need to carry out refined management, especially the C-end driven operation model. Every detail of the operation is inseparable from the support of data. The Internet finance industry is no exception. Major banks and Internet finance giants are also scrambling to seize the user data market.

First of all, we need to build a relatively complete data indicator system. In fact, building a data indicator system is to sort out our analysis ideas. Many people often don’t know where to start when doing data analysis , and the content and indicators of the analysis are also relatively scattered, so people will question whether the analysis results are correct. Therefore, it is very necessary to build a complete data operation indicator system, which can help us sort out our ideas, ensure the systematization of the data analysis structure, the integrity of the data analysis dimensions, and provide guidance for the subsequent data analysis.

Indicators are the link between problems and data. Only by choosing appropriate indicators can we fully reflect the problems. A good indicator should be quantifiable and easy to observe. So how to build a data indicator system? We can design a data indicator system through some marketing management models, such as 5W2H analysis method, 4P analysis theory, user life cycle , logic tree analysis method, etc.

Of course, the corresponding analysis model must be combined with the actual business model and analysis purpose. Data analysis without business logic will not produce any value.

For example, the data indicator system in the Internet finance industry can be built based on the user life cycle.

After the data indicator system is designed, we can design data collection plans based on user events in different stages and scenarios. This is actually a process of driving indicator design through business and then driving data collection.

3. Data-driven operational growth

After obtaining user data, how do we apply the data and make it generate value? We will mainly describe it through the following three aspects.

01. Use data to optimize operational strategies

After collecting user behavior data, we can know the conversion rate of users browsing, registering, downloading, binding cards and investing in operational activities, the browsing time and number of views of each product page, the number of first-time investors, the investment amount, etc.; but the data will only be meaningful if it is combined with business scenarios and summarized, compared and analyzed, otherwise it will just be numbers.

For example, our most common funnel analysis method, when we find that the user's investment conversion rate is 30%, it seems that the conversion rate is quite high. But if we compare it with other similar products, or with different user segments in the same link, we find that the conversion rate of other similar products or other user segments is 40%. Only then do we realize that there is still a lot of room for optimization in this link.

02. Use data to validate operational strategies

In the operation of Internet products , we often encounter multiple choices of product design and operation plans . Even the color of interface buttons or the difference in copywriting can cause controversy. Although this is just a matter of detail compared to the entire operation plan, for C-end users, details often determine everything. In this era of information overload, what you often strive for is whether you can enter the user's heart at the first moment.

At this time, we can conduct A/B testing . Under the premise that all conditions are the same, only one variable is different. We can use data to tell us which plan is more feasible, let the data verify whether the operation strategy is correct, and reduce the cost of trial and error. Of course, when conducting A/B testing, it is best not to have too low data volume and data density, and to have enough time for testing, otherwise it will be difficult to obtain statistical results.

For example, the Ant Fortune app uses a progress bar to guide new users to become first-time investors. The main purpose is to create a sense of urgency. Although it uses user psychological factors, there are many ways to display it.

Of the two progress bar designs above, the first one takes advantage of users’ anxiety and panic, making them feel that if they don’t rush to buy, the product will be gone; the second one takes advantage of the mass psychology of users: since so many people are rushing to buy, why not give it a try? These two designs each have their own considerations, and it is difficult to say which one is better. At this time, you can use A/B testing to verify the data.

03. Use data to guide operational strategies

There is a connection between data. If you don’t know, it’s just that you haven’t discovered the connection between them. The most classic data analysis case is Walmart’s beer and diapers. I believe everyone has heard of this case. When a business goal is related to multiple behaviors, portraits and other information, we can use data mining to perform data modeling to predict the user’s next behavior and propose targeted operational solutions.

For example, regarding the serious problem of new user churn , we can use clustering algorithms to build a user churn prediction model . Through data, we can characterize the portrait information of churned users, including their attribute characteristics, behavioral characteristics, and how long the churn cycle is. In this way, we can more accurately capture users with potential churn tendencies.

In the Internet finance industry, for example, we can build a user churn prediction model based on the user’s investment behavior, investment amount, and churn cycle.

From the above, we can see that the behavioral tendencies of the expected churn users are: users who have no investment behavior in the recent period and users who have invested funds but want to withdraw cash. For these users, we need to adopt some retention and activation strategies.

Finally, data analysis can provide us with effective information and guide marketing decisions, but we should not be superstitious about data. We should think from the perspective of others so that the data can play its true value.

Source: ITFINLAB