We talk about operations every day, so what kind of data does operations need?

1. What exactly is the operation doing?

Before getting into the topic, let’s first discuss a question - what exactly does operations do? The company's operational positions mainly include three aspects: event operations, content operations and community operations. Students who work in event operations may usually organize an event, from early planning to step-by-step execution, to the final summary and review; students who work as content editors may think more about the latest hot topics every day and what kind of content can better stimulate user activity; students who work in community operations may be more focused on community management and guidance on the community's tone.

These three types of operations are essentially doing everything centered around users, using our products as a vehicle to continuously iterate and optimize operational strategies and directions. If we think more deeply, what are these three means (content, activities, and community) doing and what have they done? We will also have a deeper understanding of what we are doing throughout the entire user life cycle.

For operations personnel, the most important thing to pay attention to is the user life cycle. For an unfamiliar user to go from initial understanding to establishing a basic understanding of the product, the content team needs to consider how to tell the user what the product does in a reasonable scenario. If users are impressed by you, they will download your product to experience it. As a novice user, after he uses our product, what kind of things should we do for him? This is also something that operators need to consider.

After the user enters our product, we will guide him so that he can quickly understand the value of our product. After completing the activation of the user, when he understands the function of the product and recognizes its value, the next step is to allow the user to remain active. When the user continues to be active on the platform reading articles, he will bring revenue to the platform (product).

When users have a preliminary understanding of the product and subsequently have a high degree of recognition of the product, they will help you spread emotions. The entire process requires operations (operations in all dimensions) to build an emotional connection between the product and the user.

In other words, when focusing on the user's life cycle, what needs to be done is to make user acquisition more accurate, activation faster, retention more stable, and revenue more diversified. Finally, it can be spread more widely.

In this way, we can extend the entire user life cycle more, better and faster. At this stage, operations need to be considered from the perspective of an enterprise and a strategy. The most important thing is:

• First, improve the user life cycle value;
• Second, reduce customer acquisition costs;
• Third, reduce operating costs.

The above explains what operations should do and what the highest goal is. Next, let’s talk about what kind of data operations actually need.

2. What kind of data is needed for operations?

First, there must be data, and secondly, data indicators: cumulative usage, new users, number of active users, reading volume, repurchase rate, length of stay, UGC conversion and other indicators.

Take UGC conversion as an example. Assuming that the current healthy UGC conversion rate is 20%, by the next stage, the conversion rate may have to be 30% to be considered healthy. For users of different categories and characteristics, UGC conversion may need to be split and defined. In this process, we need to know whether the acquired users are our target audience? Which channels have more accurate users? What are the characteristics of these channel users? How to better activate users?

As a content operator, I found that the number of new users and retention rate have increased, the number of active users is also increasing, but the number of article readings has declined, and there is no improvement after adjusting the content. I don’t understand why users are so active?

3. What is the quality of the content?

For example, a content operator once came to me and said, "Our product has good new user and retention performance, and the number of active users is also increasing. However, the number of article readings is declining, and we haven't found the cause."

After the chat, my first reaction was to see how good the content of the article was. I analyzed the following two indicators (data):

Dimension 1: Look at the users who have read articles in the past 7 days. The proportion of users who have read articles accounts for the total active users in these 7 days. Then I found that there are actually fewer such users. That is, there are 100 active users, 70 of whom used to read articles, but now there are only 60. However, the decline in this proportion can only mean that he doesn't read it. However, whether it is good or not may need to be measured by the average number of readers per capita. That is to say, people who have read an article can read a few articles on average. This indicator is actually increasing. When I saw these two data, I had an idea in my mind. That is to say, in terms of the quality of the article content, the content quality is up to standard. Because once users read an article, they will read more content, so the average reading volume per person actually increases. But the biggest problem now is that a wave of users come in but do not read. Following this logic, we can find the behavior of these users and analyze what they did after entering the product.

Dimension 2: Group users based on their behavior. This is a case from the beginning of this year. From January 5th to January 10th, I divided the users who read the article into a user group. Then, we divide the users who have not read the article during this period into a user group. Through such a split, we can explore what the users who do not read the article are doing through the characteristics of other user behaviors, user portraits and conversion rates. By comparing the differences between the two user groups, we can eventually find the root of the problem in this dimension of information (channel, source).

Generally speaking, the users brought by a stable channel are basically stable. If some investment and activities are carried out on a channel, it may result in low content quality. Under normal circumstances, a certain channel should be stable, and it would not be the case that more people do not read articles on this channel than on other channels. So, when I saw the difference in channels, I wondered if something happened in the channel.

At that time, there was a guess that this series of changes might be related to the users brought by Channel 1. From the perspective of traffic, we need to consider the characteristics of the number of new users brought by Channel 1, the retention of new users, and going deeper, how many of the current active users are downloading the product for the first time, and how many of the first downloads of the product come from Channel 1. This directly determines the retention and overall activity. We will then look at the average daily duration of stay in channel 1 and the average duration of each visit, which are judged based on basic health quality. Then, if we delve deeper from the perspective of user portrait, if we can separate the users of this channel, we can clearly understand the user behavior characteristics.

If the user doesn’t read further or interact, what is he doing? Based on business assumptions, we continue to segment users by the first visit time (5th to 10th) and the first-time users coming from channel 1.

Only through this flexible process of splitting users can we truly discover the differences between this group of users and other users. In the process of splitting this user group, all users in this stage are also grouped as a whole. By comparing them with another full set of users, we compare the differences in traffic, quality, new retention, behavioral characteristics, and key funnel conversions between the two user groups.

By analyzing and comparing the differences in key behavior conversions, we can draw some clear conclusions: For users coming from channel 1, the reading conversion rate is lower than the overall average, and is nearly 20% lower. This established the basic understanding: when making any product, whether it is a tool , community, e-commerce or other categories of products, there must be a goal. This goal guides the core value of the entire product. If the new users brought by a place do not meet expectations in the most core conversion, it can be directly judged that the user quality of this channel is relatively low.

Going further, I want to know what characteristics the users coming from this channel or the lost users have.

After looking at the user portraits through Zhuge io and clicking on each user to view the user's behavioral history in the product, we found that many users did not perform any operations after opening the app. After opening the app, the user only triggered one startup event, and the time of each visit was very regular, with more than 80 startups every day, which is obviously illogical.

Therefore, this involves the possibility that the channel may be inflating the volume for profit. The crux of the problem was found. The friends who were responsible for content creation found the students in channel operations and explained the situation. Because channel operators often don’t pay attention to the subsequent key behavioral conversions after users come, but focus more on retention to determine the channel. It may take at least 3 to 7 days to determine whether a channel can continue to be invested. In this process, the channel focuses on retention. Due to frequent channel refreshes, there are more than 80 launches every day, and the same launch is done every day. With this behavior, there is no noticeable difference in retention. Therefore, it is easy to cause some customer acquisition costs to be too high, and the users acquired may be fake users.

Therefore, here is a suggestion: Colleagues in charge of content, activities, and community operations should not just focus on their own assessment indicators. If they only focus on the data they want to monitor, they may not be able to truly discover the problem, and it will be difficult to find their real value point.

4. What kind of data is needed for operations?

Operations require data that can be analyzed. Understand users emotionally through rational analysis. By completely tracking the user's entire behavior, we can know how the user uses our products. Only then can we deeply understand the user and truly restore the user's usage scenario.

Good operations practitioners can predict possible problems before they occur, and then quickly find ways to give users a better experience and avoid existing risks. Rather than solving problems after they arise.

5. Operations + Data = Growth Hacking

Take Airbnb as an example. In the early days, it was just a small online platform. By constantly studying user behavior characteristics and how users choose a house, it discovered through data that more beautiful photos and more exquisite houses would attract more people to pay attention to and live in them, and its revenue would increase accordingly.

After that, Airbnb dispatched a high-quality photography team to help hosts take beautiful photos and upload them to the page. In the case of this service, the income for the month doubles. Although Airbnb has incurred relatively high costs, it has satisfied more users.

From the Airbnb case, we found that by understanding users and using data to find strategies, we can tap into the value of users.

6. How to develop a data-driven strategy?

Step 1: Set growth goals .

In our operations, there are three dimensions that are closely related to our goal setting.

• The core value of the product.
• The company's strategic direction.
• The life cycle of a product.

A product itself has different stages, and its biggest demand in the initial stage of product launch is to acquire users. Then during its growth period, we may have other, more considerations about its retention. If we enter a stable or declining period, our operational goals will be different, and may even include exploring new growth methods or new product directions.

Step 2: Collect and refine data as needed.

I personally think that the biggest difference between general operations and advanced operations is the ability to measure the final results based on data. There are three points to consider here:

• Refine functional events,
• Core Behavior
• Business information.

In fact, we should pay more attention to the business information behind it. Business information means that any action of yours can only be triggered by obtaining some information behind you. That is to say, what really affects users may not be your processes or experiences, but more likely their business information. For example, if I collect information about an event in which a user views a product, I want to collect more information about the product name and category behind the product, and whether the product is on promotion. How strong is the promotion of this product? These dimensions are what truly influence user behavior.

Step 3: Measure results.

When our new products, features, and operational activities are launched, we must measure their effectiveness in the shortest possible time. If there are any problems, especially in operations, we have to make quick adjustments. Focusing on user behavior, we work around four dimensions: traffic, retention, users, and conversion.

Step 4: Cross-analysis to discover value

Step 5: Understand user optimization strategies

Finally, I would like to share with you a passage from "Lean Data Analysis ": "If you have the opportunity to observe user behavior in depth, you will find that only a small number of visitors become loyal users of the product, while most of them are just passers-by." I would like to share this passage with you as the closing remark for today, and I hope that your operation will get better and better.

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

The author of this article @Zhugeio Compiled and published by (APP Top Promotion), please indicate the author information and source when reprinting!