How to spend less money and do the most effective advertising?

When we talked about brand planning in the past, for those who value pragmatism above all else, good results and the ability to tell stories did not mean much to us. In the process of APP operation and promotion , there are mainly two aspects involved.

The first one is online promotion . Online promotion mainly includes community interaction, forums, Weibo , WeChat , QQ groups and self-media , including MCU and encyclopedia . In fact, these all belong to online promotion. The characteristics of online promotion are more significant effects, less investment and relatively accurate users. But it takes a long time. However, in advertising promotion , if you are slower, the effect will be lost, which is not very suitable.

Therefore, the second paid advertising is relatively effective. The above picture shows several examples of paid advertising, mainly application products, mobile phone pre-installation, website traffic , and alliance traffic. However, paying also has its own disadvantages and costs money. So how do we need to spend money? The boss is the one who thinks about how to spend money. As for operations , you need to consider how to use the best knife where it is needed most and do your best. Next, I will introduce the characteristics of each channel of paid advertising in detail.

Channel attributes and characteristics of paid advertising

The first is application products, such as Baidu Mobile, 21 Mobile Assistant, App Store , 360, etc. Currently, domestic products also have their own advantages. For example, Huawei and vivo , their business market revenue comes from hardware charges. Because the application products are pre-installed on the mobile phone and cannot be uninstalled, its user base is very large. If you need to promote an APP, you will first enter the user products. Its characteristic is that it has a large number of users, but the amount brought by their users is limited and relatively single.

The second is pre-installed on mobile phones. Many of them are pre-installed on Android phones. Previously, many apps could not be uninstalled, and the retention rate was relatively high. But now the company does not allow this. All apps installed in them must be uninstallable. So many APP developers go to them. Nowadays, users have more APP choices. Considering the user experience of mobile phones, the APPs installed in mobile phones are usually those that are more common and have high user demand. New APPs will not appear in the first batch shipped from the factory. When you buy a new mobile phone, you will find a lot of apps, but in fact, there are only a few apps that we really use. For some things we don't need, we can just uninstall them and don't open them at all. The biggest problem with pre-installed apps on mobile phones is that the retention rate is very low.

The third is the traffic of the Web site. The Web site is the source of information that everyone pays attention to now. For example, novel websites have relatively wide coverage, as well as some news and information websites. They all have aggregation functions, and the websites have their own independent commercial content, but there are also some websites that aggregate content from various industries, and the user traffic is considerable.

The fourth is affiliate traffic. Many large comprehensive websites have DSP . The characteristic of DSP is that the price is very low, much lower than normal channels. But one of the biggest problems they encountered was the lack of transparency in the delivery process. You invest money to put advertisements in it, but where will the advertisements be placed in the end? You don’t know because you can’t see them.

Then there is the information flow module and information flow advertising , which is currently a mainstream channel in the market. I saw an advertising campaign of Facebook a few days ago, and their overall advertising revenue was over 90%, so I found that their share of the entire advertising was also very large. At present, the investment in advertising in China is far less than that in foreign countries. It can be said that there is a relatively long-term potential and a lot can be created. There are already some media outlets that are doing information flow. This shows from the side that this market is still very large. The representatives of domestic information flow are Toutiao , Guangdiantong , Tuia, NetEase, Sohu and Sina .

How to achieve targeted promotion to targeted groups

So, how can we optimize the information flow to achieve a precise layout? Because the attributes of each channel may not be the same, but the internal logic is the same. When a channel is sharing, users can find the corresponding model and it is convenient for users to operate.

How to effectively control costs? The most important thing is to find the right people, that is, to find accurate users and find your core potential users. This is very important. The process is generally divided into three parts.

Judgment based on product attributes

The first part is to make a basic judgment based on the attributes of your own product. For example, if you are a social person, most of the people there are men, and the age range is between 18 and 40 years old. As for campus loans, it is for college students under the age of 24, because they are college students. Although everyone has a basic judgment about their own products, this basic judgment may not be correct. So we often fall into the trap of thinking, what do I think my product should be like? But your feeling is usually wrong. For example, we have been advertising frequently these days, and you may think it is very high-end. But after I advertise it, I find that no one clicks on it, which is a wrong judgment. If the user says: You have released a very low-end product, what is this? and then the number of clicks goes up like crazy. The effect is that the market uses data to test, which is correct.

Utilization of past data

This is our second step, which I think is very, very important, which is to use the data from the Minwang website. If we release the data on each channel, there will be a data report . Then you will also have the overall data of your product. Compared with the data of the same product, their data will be used by you, and the process of your delivery is actually a process of continuous optimization.

Case 1: I previously made an APP called Tao Travel, which is an APP for air tickets . The departure point of most air tickets is Shanghai, so he would typically think that his customers must be in Shanghai. Because his starting point is Shanghai. If you are too far away from here, even if it is originally a cheap air ticket, if it is accompanied by advertising, the investment will be relatively large. But later we felt that there were too few such people, or we might miss potential users. Later, we made an adjustment to target the vast rear audience. Later data showed that 47% of the orders came from this advertisement. That is to say, he has a large number of uncertain users.

Case 2: When we think of a lipstick, the most intuitive feeling we get is labels like feminine, beauty, and fashion. But after we ran that ad, all subsequent displays showed that the top-ranked tag for that ad was not these words, but "dog". From this we can see that the compatibility between women and dogs is very high. If we target people who are interested in dogs and place advertisements, the return rate will be very high. This is an example, so it makes me think why there is the term "single dog"? You may be single because you don't have a dog.

I think casting is a very detailed process. During the delivery process, your chatting or your targeting will have an impact on the delivery process and will be useful, so this is actually a process that requires frequent debugging. For example, Guangzhou and Shenzhen, which one is better? You say Shenzhen is better, but he says not necessarily. Then you have to test it for a long time and find that Shenzhen is good, then it is certain.

Application of intelligent delivery tools

However, there are many disadvantages in the process of manual debugging, which is too cumbersome and has limitations. So how to solve this problem?

Using intelligent tools means replacing human beings as intelligent users with machines. As far as I know, Tencent and Baidu are all doing this, and this is a major trend. The Go competition between Alphago and Ke Jie in the past two days has attracted a lot of attention, which also illustrates this point from the side.

Some of the tools we currently have have different platform names, but the internal logic is actually the same.

DMP: Customizable audiences, rich and flexible targeting functions

DMP is a customized population, which is a concept of potential users. For example, in the Guangzhou case, there is a woman who is always interested in dogs, so you can artificially adjust the part of people who are interested in dogs to expand the scope of your delivery. This group of people, while being interested in dogs, will also be interested in other things. This is what we got . So how does the DMP tool work? For example, if you already have technology for 10,000 users in your backend, these 10,000 users must have many tags or many attributes. You import these 10,000 data into the background, and it will automatically identify and then automatically look for an activity similar to that of these 10,000 people. What you have collected currently is only 10,000 or 20,000 yuan, but in the background there may be 10 million yuan. Then it plays automatically, so that you can intelligently find your potential users.

PMP: Premium advertising space combined with the advantages of private DMP

The second one is PMP, and its full English name is “Portable Media Player”. Translated into Chinese, it means a small vegetable market. If I tell you that there is now an opportunity to play, then the users have so many. The user's needs are conveyed to the DSB platform, and the DSB platform will then tell its own database about this need. For example, if you want to buy cucumbers, the database will immediately go to the vegetable market to look for them, and whatever it finds is what I have. DSB will automatically search for such a material. Then the advertisements will be delivered, and DSB will be further displayed to the media background, and then displayed to users, thus realizing such a virtuous circle.

In the later stage, through the DSB behavior, continuous optimization and then distribution, such a project is our main intelligent tool.

oCPC: Intelligent conversion bidding, bidding strategy to maximize conversion

The third one is oCPC, which can only convert. The internal logic is the same and it also helps you find accurate users. The difference between DSB and oCPC is that DSB is the backend data, the more important one. For example, if a user is interested in e-commerce or the gaming industry, DSB will find a corresponding scenario. But there are also many tool-based ones. For example, for tools, I only need to consider my active people. What I need is a front-end customer, an activation, and COPC is generated based on such a demand or need from the front end.

Originally, CPC was the purchase price, and charges were made for each click. This is optimized through human behavior, and the cost will be reduced to ten dollars or less. But now it is not necessary. You can use oCPC, which will be divided into two stages. When you place advertisements, you also use oCPC at the beginning, use it to place advertisements, and then sell them to users. If the users collect the advertisements at the end and there are 100 users, the first stage is over. Analyze all the attributes and labels of these 100 users and then draw a conclusion. Then we enter the second stage. In the second stage, you only need to pay 10 yuan. The machine will automatically search for this group of people, and pick up those who have activated within 10 yuan.

So you will find that now is the machine age, you don’t need to do manual analysis, the machine will automatically find this group of people. This is also a very good platform that we are currently using.

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

This article was compiled and published by @操作鹤手 (Qinggua Media) by the author @操作鹤手. Please indicate the author information and source when reprinting! Site Map