Editor's Note: After completing the introduction to the traffic allocation strategy, we move on to the second part of the pre-investment strategy - the targeted strategy. "Targeting" can be said to be a high-frequency word in the field of commercial advertising, whether it is discussing advertising technology or commercial product design or advertising techniques, it is impossible to avoid the topic of "targeting". Through the analysis of advertising in the Internet automotive industry, this article considers the targeting strategy in a simple and in-depth manner. Let's take a look!
In many ad serving courses, even the targeting setting techniques summarized over the years of advertising are considered to be advertising strategies. This has led to the fact that when people search for content related to the advertising strategy on the Internet, they are mostly teaching us how to set up various settings in the ad delivery background. But in fact, the advertising strategy is far more than just targeting, and many people's understanding of the targeting strategy has also been greatly biased for a long time.
Ad targeting is not essentially a strategy but a product feature used to filter the audience for your ads. This function can filter the target population for different delivery purposes through the combination of tags, so the targeting strategy is essentially designing different tag combinations, but it also does not involve the core decision of advertising delivery and the reach of the crowd. Then the one-sided targeting strategy is considered to be the commercial product manager of the advertising strategy is obviously unqualified.
From the perspective of ad optimizers, because they don't involve the underlying logic of the ad system, their delivery strategies are often based on account, targeting, and budget settings. However, from the perspective of the commercial product manager, it is not only necessary to clearly understand the position of the orientation in the entire system and the implementation principle, but also to consider which delivery strategies the optimizers can implement based on this on the basis of the functional implementation, and support a variety of intelligent delivery strategies (machine automatic delivery).
This section is demanding on the personal experience of commercial product managers and an understanding of the entire advertising system.
First, the principle of advertising targeting and the direction of optimization
In many articles, the principle of advertising targeting is mysterious, and what is interesting to recommend ads of interest to users in real time according to their current browsing behavior. As soon as this kind of rhetoric comes out, it is known that it is "fooling" the layman, and the reason is simple, because the vast majority of the current targeting functions of the advertising platform rely on the label system to achieve. Although it is true that the user's browsing behavior will be obtained in real time, friends who have actually done user behavior analysis will know how chaotic and disorderly user behavior is.
We can't exhaust all the user behavior and naturally can't make different ad delivery decisions for each user behavior, so it is more economical to classify user behavior as limited labels and then make advertising decisions for different combinations of labels.
Next, let's talk about labels, there are many types of labels, common ones are: region, user attributes, context, behavior, preferences, etc., without considering the accuracy of labels, these are the standard of major platforms. But multiple types of labels also bring another problem, too many labels do not know how to choose?
At this time, DMP (data management platform) is needed to provide the user portrait of their target users for the advertisement, for example, the target group is a post-95 male, then the target is set according to this conclusion when targeting. Theoretically, the more accurate the label, the better the performance of the ad.
The example we gave above is the simplest and most intuitive scenario, but in real life, each industry, each type of product, the user's behavior, and the decision cycle are different. This requires us to deal with it in different categories.
Back to the automotive industry we are familiar with, the user's car purchase decision cycle is much higher than other e-commerce products, so the user behavior during the period will become extremely complex, how to find the law of these behaviors by stripping away the cocoon, forming a label system for automotive industry users, and then forming different directional strategies in advertising has become a problem for us internet automotive industry commercial product managers to think about.
Second, UVN-BI labeling system
Here we share a set of car purchase user labeling system that we have explored after years of practice - UVB-BI label system.
The design of the entire system draws on the famous RFM model. The RFM model is based on the three indicators of the most recent consumption (Recency), consumption frequency (Frequency), and consumption amount (Monetary) as the dimension to establish a coordinate system can divide customers into up to 125 categories.
Based on rfM, we bring this approach to the group of car buyers. UVN-BI user grouping is to characterize and cluster a variety of basic attributes of users and map these attributes with the user's car purchase stage and the user's interest in the car to form a user group model with the car system as the dimension, before showing the model, we first define the dimensions involved in it in detail:
U( User) user basic attributes: mainly use the city level and intergenerational city level (first to sixth tier cities) intergenerational (post-00s, post-90s, post-80s, post-70s, pre-70s)
V( Value) user annual income level: less than 100,000, 100,000-200,000, 20-300,000, 300,000-500,000, more than 500,000;
N (Need) user group preferences: less than 80,000 models, 80,000-100,000 SUVs, 80,000-100,000 cars, 100,000-150,000 SUVs, 100,000-150,000 cars, more than 50w SUVs, MPVs, sports cars, etc.;
B( Behavior) User car purchase stage: attention, rough selection, interest, preference, intention;
I (Interest) User interests: space, power, handling, fuel consumption, comfort, exterior, interior, cost performance.
The construction of user group preferences is mainly based on the upward abstraction of the car series that users care about. Intersecting the price segment of the car series with the type of car series, a user may pay attention to multiple cars at the same time, but in most cases, the multiple cars they are concerned about will be under the same type and price range. The characteristic of ethnic preference can portray this situation very well, and such a feature has better generalization ability than the characteristics of the car system.
In addition, because there are at least three dimension features of the intersection U, V, N these three characteristics are in the form of intervals to avoid too many crossover dimensions. After a total of 480 of the three types of features are crossed in various combinations, 480 blocks are formed in three-dimensional space, and the number of people of a car can be counted according to these 480 blocks, and the user distribution can be obtained. At most, the following users of a car series can be divided into 1738 categories, and the users of a car series will be mainly concentrated in 160-210 categories. As shown in the following figure:
The entire stereo coordinate system can be rotated arbitrarily by 360, of which each small category can be clicked on individually, and the specific information of the category will be displayed on the right side of the page. In addition to the three indicators of UTV, each sub-category will also be hung with its corresponding B label and I label, that is, the car purchase stage and the purchase point preference. The characteristics of this display method are more novel and more analytical than the traditional DMP to show advertisers the age, gender, geographical distribution and other commonly used charts of users, which is out of the initial stage of direct display indicators.
In the next article, we will use this labeling system to design a targeted strategy based on different business needs.
This article was originally published by @Everything Needs Jingsheng on Everyone is a Product Manager. Reproduction without permission is prohibited
The title image is from Unsplash, based on the CC0 protocol