Operations are not based on intuition, but are driven by data.

Many college students who are about to graduate will face the important decision of which industry to join. After all, "men are afraid of choosing the wrong profession, and women are afraid of marrying the wrong man." In fact, when we don’t know what industry we want to work in, we can first look at which industry most of the outstanding talents are gathered in. Industries with many talented people will often be the hottest industries with the most concentrated resources in the next few years.

There is no doubt that the Internet will be the hottest industry in recent years and even for a long time to come. Because the Internet more efficiently connects people, businesses, services, and even things, a large amount of valuable data is generated during this period, and these data can be collected and analyzed and then fed back to actual application scenarios. It can be said that data is the guideline for the iteration and update of Internet products. This is no exception in our daily operations . We must always keep in mind data-driven operations.

1. What is data?

Since we always say that we want "data-driven operations", what exactly is data? Today we will not discuss the concept of big data, because most operators find it difficult to come into contact with so-called big data in their daily work. Since big data is called big data, its premise is to have a database of sufficient scale, which many small companies do not have.

This time we will mainly define data from an operational perspective: data is a description of users. It can be a description of user identity attributes, a representation of user preferences, or a record of user behavior.

Generally speaking, the data sources of a product can be divided into two categories: one is the collection of user identity information, and the other is the recording of user behavior.

1. User identity attributes, for example, most products require users to fill in birthday, gender, contact information, etc. during registration. Some dating products even require financial status, family status, and marital status, while recruitment products can collect users’ past resumes, salary status, strengths, and specialties.
2. User behavior can be divided into user product behavior and user preference behavior.

Product behavior refers to the number of times users click on product pages and buttons. Product and operation personnel can use these user behaviors to optimize product functions or pages.

Preference behavior refers to the user's stay on a certain page, or the login status of a certain page. For example , Toutiao provides personalized and accurate recommendations based on user preferences.

2. Why should we use data to drive operations?

As for why we emphasize data-driven operations, the reason is actually very simple. I wonder if you still remember the three major principles of operations that I have always emphasized: customer first, results-oriented, and efficiency first.

Without data support, how can we know what users’ needs are? Do users really need the features of our product?

Without the support of data, we keep saying that we are result-oriented, but how can we determine which path leads to the results?

Without data support, how can we choose the most effective path and make good use of one of the most important functions of the Internet - efficiency?

Therefore, if we want to effectively implement the three principles of operation, we must use data as a guiding light. Otherwise, all decisions and plans can only be made on a whim. Even if it is a purposeful trial and error through MVP, if there is no data feedback in the end, how can we know where we went wrong and how to improve.

3. How data drives operations

Regarding the question of "how data drives operations", since I have given too many examples of personalized recommendations for content-based products such as "Toutiao" in the past, this time we will change the topic and use e-commerce as an example to talk about "retargeting ads" that track consumers based on data.

Everyone can recall whether you have had a similar experience: when you visited e-commerce website A, you browsed a certain product. Although you were quite interested, it was not enough to stimulate you to place an order. A few days later, when you visited another website B, you saw the product again. At that time, did you have a feeling that it was God’s will that you should buy it, providing you with a "plausible" reason to place an order? Even if your rationality still prevailed this time and you reluctantly closed the advertisement, but if you saw this advertisement on website C or website D next time, and were repeatedly tempted by it and aroused the urge to consume, you would most likely give in.

For Internet practitioners like you and me who have experienced the baptism of routines, it is not difficult to find that these so-called "God-given" advertising exposures cannot be pure coincidences. In fact, these companies just used an online advertising technology called "redirection."

So what exactly are these “lingering” “redirect ads” that keep haunting us?

In fact, it is a kind of advertising method that remarkets to people who have already browsed the website. The realization of this re-marketing must be based on the user’s last browsing behavior data. Generally speaking, these companies will join the redirect advertising alliance to obtain the qualification to read the user identity and product information recorded when the user browses other websites in the alliance, and then replace the dynamic advertising space in the current website browsed by the user with advertisements for products that the user is interested in.

This is a product-based "data-driven operation" case. The implementation of "retargeting ads" requires inserting a tracking code into the website.

In addition to the situations described in the cases and the recommendation mechanism of content products based on user preference data, our other daily operations also rely on data-driven work.

For example, when a product needs to add a new function, we need to conduct A/B testing to prove the feasibility of our new function through data;

When we want to improve user retention rate , we need to use data to discover the characteristics of our lost users and the characteristics of our retained users in order to formulate our retention plan.

For example, Facebook found that when new users add friends after registration, their retention rate is often higher, so the operation staff designed a mechanism to recommend friends to new users. In the early days, Sina Weibo also found that users who set up avatars tend to have higher activity and retention rates than users without avatars. Based on this data, the operation staff guided users to set up avatars.

When we plan an event, we also need to refer to the user participation and active user group attributes of previous similar events (including our own company and competitors) to decide whether it is necessary to organize an event of this type.

As such, it can be said that the daily work of operators cannot exist without data. Any of our operational plans need to be based on facts and be targeted. Data is the basis for us to find this "right". Operations need to be driven by data in order to reach the deepest needs of users in the most efficient and accurate way.

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

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