Analysis of Internet promotion channels!

Our marketing channels are located where there are many users.

With the fragmentation of mobile Internet, users' attention is attracted to various vertical apps, search traffic and effects are declining, and more advertisers are shifting their budgets to information flow. Due to the precise label targeting, native content, high interactive transmission rate of information flow advertising, and the fact that it is not restricted by advertising space, media advertising inventory has increased, and the overall market size continues to grow.

1. Principles of Information Flow Advertising

Information flow advertising is also bidding advertising, but it is different from the traditional search competitors’ ranking competition, in that it competes for every user’s exposure rights. Whether or not you can get the exposure right depends on the ECPM value, eCPM (pre-allocated exposure) = bid * quality (CTR).

Of course, ECPM has corresponding conversions for the current various bidding methods (CPM, CPC, CPA, oCPM, oCPC).

The most important thing here is the estimated CTR and estimated CVR. The CTR and CVR of each time will be estimated based on comprehensive factors such as population characteristics, industry characteristics, historical CTR, CVR, and advertising and page materials.

2. Information Stream Advertising Platform

At present, information flow platforms are mainly divided into social, information, tools and video categories . The first camp is still monopolized by BATT traffic giants. Because of the large number of users and precise algorithms, it is also a must-fight place for advertisers.

3. How to operate information flow

Advertising = right time + right scene (media) + right advertisement (material) + right crowd (push). The important links of information flow advertising are targeting, material, page and bidding. That is, how to accurately label and target the right people, how to create corresponding materials to attract and convert these different targeted groups of people, and what purchase price should be adopted for different groups of people to ensure a reasonable ROI.

1. Product understanding

Before information flow or any marketing, we must first understand the characteristics of our own products. Only by refining the selling points of our own products can we meet the needs and pain points of the targeted groups of people in order to produce corresponding materials. At the same time, when understanding the product, we will also find that some people are not our own users because the product cannot meet their needs.

Take K12 offline children's English as an example:

Products in different industries should be decomposed into structured dimensions based on their own characteristics. The more detailed the understanding of the products, the more accurate the extraction of selling points. It is recommended to communicate more with the company's customer service, sales and service personnel.

2. User understanding

The user part is mainly analyzed around user portraits, user needs and pain points. The clearer the user portrait is, the easier it is for later targeted operations. It includes age, gender, region, education, occupational income, interests and hobbies, Internet habits, consumption attributes, etc.

In addition, we must dig deeper into users’ real needs and pain points to understand why users choose this product and what their purpose is. There are so many suppliers of this product, how can I choose you? So where is the need for users to make judgments and decisions?

In addition to communicating with customer service and sales, you can also build your own community to conduct user surveys, check out user comments in communities, forums, and e-commerce service review sites where users frequent, etc.

3. Platform selection

1) Daily activity, monthly activity, duration, and number of launches

You can check the relevant data of the APP through platforms such as iResearch or Analysys. The larger the daily active users, the more users, the longer the usage time and the more times the APP is started, the higher the user quality and the greater the stickiness.

2) Platform user analysis

The premise for determining whether a large amount of exposure can be obtained is that the platform has a large number of users that meet its own needs, so it is necessary to analyze the user population distribution of the platform in order to match it with the user population of its own products.

3) Data dimension algorithm orientation

Information flow advertising targets the target audience based on user tags. The accuracy of the target audience is determined by the authenticity and diversity of the platform's data dimensions, the algorithm's ability to recognize tags, and the targeting dimensions.

For example, in terms of data accuracy, WeChat is one step further because the data is real (binding ID card to open payment wallet), so the basic targeting is more accurate.

From the data dimension , WeChat has social data, payment data, reading data, map data, etc. Combined with Tencent's many matrix products, the ability to obtain user tags is more accurate when the data is connected.

However, accurate data still requires algorithm recognition . We know that most information flow platforms are targeted by interests, but interests are strong and time-sensitive. How to target a large number of dynamic user tags is a test for the platform algorithm.

For example, if a male reads an article related to sports or football, then this user will be given a sports tag, but what is the coefficient of this sports tag? 10% or 90%? Is it a long-term interest? How long does it take to read an article and follow it?

4) Crowd targeting

After we understand the products, users and platforms, we need to establish our own user population targeting. Whether it is information flow or search advertising, it is nothing more than spending money to buy, but the population is divided into core, target and potential population. For search, it is divided into brand words, product words, general words and population words according to the strength of the keyword intent. The core population is brand words, the target population is product and general words, and the potential population is industry words.

The traffic volume increases from small to large, but the conversion rate decreases from high to low. We need to comprehensively calculate the CVR of each group of people and give a reasonable price and budget. The same principle applies to information flow. Divide the core target potential population and determine which targeted combinations of each group of people on the platform can meet the target. Of course, information flow platforms can adopt a posteriori approach.

Currently, most platforms are divided into basic, interest and behavioral targeting. Among them, basic targeting (age, gender, region, education, marriage and love) tends to be general, interest targeting tends to be targeted, and behavioral targeting (search, APP, e-commerce, social interaction) tends to be more precise. Therefore, when the effect is the main factor, tests are generally carried out with various cross-targeting combinations such as basic + interest, basic + behavior, etc.

In addition, mature information flow platforms use DMP data analysis, and advertisers can upload converted data population packages to expand the population through lookalike, which is also a kind of targeting method.

5) Material landing page

To target different groups of people, different combinations of materials (selling points + styles) are needed to attract them.

For example, for the target population, users have many demands, including teachers, teaching materials, prices, teaching effects, etc. Each demand point can be refined into different selling points to meet them.

Even for the demand point of price, the copy can be full discount, 0 yuan, free, trial, gift, discount deadline, etc., so a variety of price copy can be edited. The pictures can be divided into large pictures, three pictures, groups of pictures, four-grid, nine-grid, splicing and other styles. There are many results of combining two. Of course, some platforms such as Toutiao can already program creativity, and the machine automatically combines the uploaded creativity for testing and optimization.

In addition to diversity and relevance, another important point about materials is freshness. For example, if you want to gain exposure using the CPC bidding model (or ocpc, ocpm), the first prerequisite is to increase CTR to obtain a higher ECPM value.

Before a new ad is launched, the system will give an estimated CTR for the ad based on user demographics, historical CTR and other factors. When the CTR is higher, it can get more exposure. However, when the actual CTR of the ad appears, if it is low, the ECPM value will be reduced over a period of time. Even if the CTR is high, it will decrease over time. This is why information flow advertising platforms continue to add new ads.

Another thing you must pay attention to is that different information flow platforms must describe the copy in accordance with the user characteristics of the platform. For example, for Zhihu, it can be a question-and-answer format, which will increase CTR more naturally.

In addition to some platform carriers, you can try various methods for landing pages, such as WeChat official accounts, mini programs, and H5 pages.

If the official account is used as the carrier, the direct conversion path will be longer, but it will benefit advertisers who are good at content and user operations. This needs to be tested.

The mini program can make the conversion path smoother. Pages for information flow are generally highlighted with a short version of the bananer, with the conversion entrance placed at the front and another entrance added at the end.

6) Bidding budget

Currently, there are various bidding methods on information flow platforms, including scheduled advertising, CPM, CPC, CPA, Ocpm, Ocpc, etc. The starting bid should be higher than the system recommendation or the historical stable CPA method. It is best not to limit the budget too much. The data is only reliable when the exposure volume is greater than 5,000. And it is necessary to divide the bids and budgets according to different circles of people.

Summary: Funnel Principle

Author: Aiqi Online

Source: Aichi Online