Editor's lead: Affected by the big environment, the traffic dividend is gradually receding, and it is increasingly difficult for Internet users to grow. The importance of refined operation is reflected here. The author of this paper shares the basic concepts and reasons of refined operation, and the specific methods of refined operation. Come and have a look if you are interested. In the post-epidemic era, with the disappearance of traffic dividends, it is increasingly concentrated in giants, which not only raises the threshold and cost for ordinary enterprises to acquire customers, but also makes growth difficult. The concept of refined operation has been talked about more and more, and more enterprise operation ideas have also changed from extensive operation to refined operation. As can be seen from the recruitment needs of enterprises, product operation, user operation, community operation, content operation, activity operation and so on. For example, content operations can be divided into more detailed positions such as new media operations, live broadcast operations, and small red book operations. There are some differences between refined operation and the operation we usually understand. You can guess the refined operation from the name alone, with finer granularity and more intensive and meticulous operation direction, which means paying more attention to the effect and the input-output ratio. Let's briefly talk about it from three angles: what is refined operation, why it should be refined, and how to refine operation. 1. What is the era of refined operation of mobile Internet? The core of operation is the user, and everything revolves around the user. Therefore, refined operation is to do more fine-grained operation for users, such as more targeted operation activities and operation analysis for user groups, marketing channels, transformation processes, usage scenarios and user behavior data, in order to achieve operational purposes. To put it bluntly, it is how to play more traffic, how to dig out deeper and more user needs and meet them, improve the conversion rate and improve GMV. Refined operation is undoubtedly a sword in the process of enterprise growth. Refined operation, not only for existing users, but also for all user groups. In other words, it is necessary to refine the addition, loss and recall of users, and have a complete user traffic model strategy. When users pull new ideas, they should not only analyze the number of channel customers, but also analyze which keywords to use to search for more customers. Among the incoming customers, what is the quality of customers in many channels, and whether the payment ratio, repurchase ratio and repurchase times are high (high is only a description, how high, you need to make your own indicators). More refined, with a high proportion of payment and repurchase, is the total GMV they contribute high? If not, is it a quality user? How to improve the customer unit price of this part of users? For another example, the time from the first purchase to the second purchase is longer than the average time to the store, so are these customers quality customers? At this time, it is necessary to analyze the refined operation. Why is the interval between repurchases so long? Is it because the receipt is too slow? Is it because we haven't contacted the customer twice? Is it because our products can't meet his more needs? Second, why should we refine our operations? There are both macro and micro factors. I think there are mainly the following aspects. 1. Traffic is getting more and more expensive. I still remember the article published by 36K, "Can't afford traffic, entrepreneurs are life and death every day". The article pointed out that the cost of obtaining customers is rising, and the cost of obtaining customers in some industries has even reached thousands of dollars. The extensive style of "stupid people have more money" has become a thing of the past. At that time, the cost of acquiring a customer may be a few dollars, which makes a large number of practitioners miss it. Fortunately, there are still some traffic dividends, but the process of cooperation is quite tortuous, so you need professional institutions to help you dock traffic, so as to achieve the effect of growth; Unfortunately, traffic has been gathered in the giant's traffic pool, which is easy to be copied. 2. The disappearance of demographic dividend The number of netizens in China is basically saturated, the demographic dividend disappears, and the incremental market dries up. The pressure of competition in the stock market is not too great, and there are a large number of practitioners in almost all segments. In this case, there are only two ways for enterprises to go, either to find new dividend-paying sites or to dig deep into existing users. Perhaps the sudden cross-border fire in recent years is also the reason. 3. The homogenization of products is serious, especially in the world. For example, in the clothing industry, it has become necessary to search for the same paragraph on the head e-commerce platform. It is true that searching for the same money can increase the sales of high-quality goods, but it also illustrates the problem of product homogeneity from the side. This brings a thought: how to get customers better when everyone's products are the same? I think innovative products and innovative services should be a way to generate sustainable growth. But how to innovate involves that you need to be very clear about the needs of your customers, so that you can have a refined operation idea. 4. User demand segmentation The current customer groups are constantly subdivided, resulting in many kinds of circle cultures such as double income, childless, single, second generation and Z generation. Users in different circles have different needs. Moreover, users in the same circle have different needs in depth and breadth. Only by digging and subdividing can we better understand customers. 5. The vigorous development of mature technical personnel, technical capabilities and technical means of technical resources, especially the maturity of technologies such as big data, cloud computing and AI, has laid a foundation for enterprises to tap customer needs and conduct targeted operations. Paying attention to data has also become a common cognition of enterprises in the digital age. Not clear about DTC mode, civic traffic, interaction, planting grass, etc. Promoted the process of refining, or whether the demand for refining produced them. In any case, the time is ripe for refined operation! Third, refined operation, how to refine? Refined operation has its own ideas, and different industries and operation modes have their own ways of playing. At present, user operation is an effective methodology. Of course, refined operations are definitely not limited to user operations. The following figure shows the common operation ideas and processes of refined operation. From this picture, we can disassemble many points, and the most intuitive user grouping needs user refinement; In addition, there must be process refinement and scene refinement. 1. user refinement in refined operation As mentioned above, the core of operation is the user, so-called "people-oriented", and refined operation should be user-oriented. The premise of user refinement is to be familiar with users, and to have clear user portraits and user stratification. 1) Let's talk about the user portrait first. The establishment of user portrait is the first step that users are familiar with. It's difficult and complicated, but exploring users is interesting and boring. There is a lot of knowledge about creating user portraits in the market, and everyone will find a lot when searching. Here, I probably won't elaborate. User portrait can be understood as labeling user information, that is, artificially labeling customers with an established labeling system. Label system, I can divide it into three types: basic information label, behavior label and customized label. The basic information is easy to understand. It is similar to age, gender, region and so on. Commonly used content is as follows. Different industries can add different content according to their own needs. Such as operating habits, equipment model, system type, common IP, etc. Followed by behavioral tags. Behavior labels, as the name implies, are labels generated by customers' behaviors. The picture below is for reference only. Different industries add their own labels such as recharge, cash withdrawal, coupons and points. Custom tags are generated based on certain behavior characteristics, and can also be attributed to predictive tags in behavior tags. This kind of label alone is quite special. What is special is that it represents a kind of people. For example, the second element, basketball fans, equipment party; For example, pets, homeowners, car owners, cosmetics, games and so on. Jane talked about tags, and then how to build a tag library, both simple and complex. Generally speaking, according to the user's information statistics and behavior analysis, combined with different dimensions of data types, and given a certain weight, a certain calculation method can be used to generate a user tag library and portrait, and apply it to the operation. Explain separately here that the process of data cleaning is indispensable and helps to deliver business logic more accurately. Then there are other user behavior statistics, which can be increased according to operational needs, such as financial behavior and business behavior. The role of tags must be applied to operations, such as content operations, event operations, advertising or other optimization. Operators can operate according to their own operational objectives. Refer to the refined thinking above &; In order to facilitate the use and management of tags and refine the data used by user tags in operation, the process will not be discussed in detail here. 2) After the user portrait is finished, there are many ways to stratify users, such as AAARR model, RFM model, pyramid model and stratification based on user life cycle. There are many tutorials about this kind, and you can choose to use them according to your own needs without discussion. The way I often use is user stratification based on the first picture. User groups are divided into four categories: new customers, active customers (user pool), lost customers and user recalls. RFM model will also be used to subdivide the current situation of users. 2. Process refinement Process refinement, analyzing funnels of different paths; Traffic map, analyze the obstacles of the core path, and refine the process. To put it bluntly, it is to analyze the user path. Analyze different user paths and view funnel data under different paths. There are many paths to analyze, but the core path must be clear. Because sometimes the core path you think is not necessarily the most real core path, and it is not necessarily the greatest expectation of users who use your products. So sometimes you can't fix the core path. Maybe your idea is wrong? In the process of guiding users, through the analysis of different paths, we can find the obstacles under different path contacts, at which node users are lost, and what strategies should be done here to broaden the funnel. There needs to be a buried point here. There are a wealth of tutorials about burying point teaching on the Internet, which you can choose according to your own needs. These path analysis, also need not be too frequent, every week, every half month, every month. According to the nature of the product and the value of the demand. For example, if you are an e-commerce platform, when analyzing the user's path from login-search for goods-buy now-place an order, you can see what is the final probability of placing an order directly, and what is the proportion of customers from buying immediately to placing an order? We can analyze the reasons why customers give up buying according to the length of stay on the order page and the portraits of large users. If a customer gives up buying because he chooses the wrong product; Whether there are customers who have placed orders for a long time but have not placed orders; Whether a customer clicks on the coupon and finds it unusable; Whether there is a new address, if you want to fill it in again, don't buy customers for the time being; Whether there is an activity in the store recently, customers who place orders during the activity after joining the shopping cart will do more path analysis and analyze the current reasons, and the judgment of customers' purchasing attributes will become more and more accurate. For another example, is it very common? It's almost time for dinner, so I hurriedly opened the take-out APP to order take-out. As a result, I forgot to use the red envelope on the next page, and then canceled the re-ordering. Analyze the proportion of customers who cancel orders for this reason along this path. Guess, can you forget the red envelope page again and tell him that you have useless red envelopes? The purpose of process refinement is not to generate data, but to find the strategies you need to adjust and improve. First, there must be product design, then product verification, and finally re-optimization. In the process of product verification, user feedback and your deeper understanding of the product will push you to optimize the product design, which may not only reduce the user operation path, but also improve the core of user transformation. Just like more and more convenient registration and login links, complex options are constantly being eliminated and split; Eliminate redundant items and split some unimportant and low-value options into other more reasonable paths. You also need to make a traffic chart to analyze the flow direction of users and the distribution of stay pages in your product, which is more convenient for you to find the core functions that users value most. Moreover, the traffic map is also liked by leaders, and its value is quite high. 3. Finally, talk about the refinement of the scene. Refined scenes are common in recent years, such as VR house viewing of Keke Holding Company, novice mode of Baidu navigation, and high-definition road navigation on many navigations. There are also some transit tickets used when buying air tickets and train tickets, which are made after digging deep into the needs of users and solving the pain points of some customers in a certain scene. We are used to other things, such as express delivery information, you can clearly see where the express delivery is and when it will be delivered. This is similar to the current takeaway. Where is the rider now and how long will it take? There are also kitchen monitoring in some hotel restaurants, which can see the working situation of the kitchen at any time and solve some food safety problems that dining customers are worried about. From these cases, we clearly know what is the refinement of the scene. To sum up, it is to solve some problems of some customer groups in the process of using products after digging deeply into user needs. More accurately, it is to understand customers and find that their specific needs in using products are difficult to meet. If it is, it is not only icing on the cake, but also solves his core pain point. Generally speaking, these customer groups are either of high value or large volume, or have a high effect on customer transformation or secondary transformation, or have a good promotion effect on user activity. Their similarities are obvious, but they need to dig deeper, be familiar with users and be able to listen to their voices. When the user is active, the polishing function of the idle fish APP solves the pain point of the user's activity. Of course, he didn't just solve this pain point, and active users didn't just rely on this solution. The operation work is very boring and imaginative. Although there are some fixed routines to play, more imaginative schemes and verification may bring you more surprises.