Tencent Social Advertising Getting Started Guide

Recently, many of my friends who have transformed from SEM to information flow optimizers have complained to me that their account performance has not been good and they are very doubtful about their abilities. In fact, the two most important points of information flow advertising are materials and targeting: materials are the way our advertising content is presented, whether it fits the scene in which the advertisement appears, and whether it hits the user's pain points; targeting is whether we have chosen the right people.

You can imagine a scenario: if a person who likes to read comics, after a busy day at work, comes home and wants to entertain himself with his mobile phone, if he receives a new book released by his favorite cartoonist at this time, he will most likely check it out or place an order, but if he receives some advertisements that he is not interested in, then the probability of him clicking or buying will be greatly reduced. Therefore, if the materials and targeting are well prepared, the right people are selected, and the right way of advertising presentation is chosen, then the performance of our account will definitely not be bad. Today, with the help of Tencent's background, I will talk to you about how to make good use of the targeted method of information flow.

Before we talk about it, let’s first take a look at what targeting methods the platform has for us to choose from.

First, let's look at a picture:

From the above figure, we can see that Tencent backend's targeting methods can be roughly divided into five categories:

1. Demographic attributes: We select the target group we want to circle based on geographic location, age, gender, education level, marital status, work status, property status, consumption type, game consumption ability, consumption level, and residential community prices.

2. User behavior: Select the target group we want to circle/exclude through behavioral interest intention, new device users, and excluding converted users.

3. Customized audience: Select the target group we want to circle/exclude by targeting user groups and excluding user groups.

4. Weather targeting: select the target group we want to target based on temperature, UV index, clothing index, makeup index, and weather.

5. Device targeting: We select the target group we want to target based on device brand and model, Internet access scenario, operating system version, networking method, mobile operator, and device price.

After reading these five categories, let's explain in detail the options under each category and the suitable scenarios.

The first is the first category - targeting methods based on demographic attributes:

Geographical location: From the above picture we can see that the geographical location can be selected domestically or abroad, or a certain province, a certain city, a certain district, or a certain business district. You can also select a specific location from the map and how large the surrounding area is covered. In addition, we can also choose whether we want to target people who have been in these locations recently (currently or in the last 1-2 days), people who live in this location permanently (have lived here in the last month), people who have come here on a trip (for tourism or business trip, and have not lived here for the past month), or people who have been here (have been here in the last month).

Age: You can choose according to the age range given by the system, or you can choose your own age range.

Gender: You can choose to target only men or women, or you can choose not to restrict gender.

Educational background: You can choose to apply to PhD, Master, Bachelor, Associate, High School, Junior High School, Primary School, or you can have no restrictions on educational background.

Marital and parenting status: You can choose to target the following groups: single, dating, newly married, married, parenting (0-3 years old), parenting (3-6 years old), parenting (6-12 years old), parenting (12-15 years old), parenting (15-18 years old).

Work status: We can define the people we want to target by occupation type (IT Internet, government officials, scientific researchers and educators, financial workers, etc.), occupational attributes (gold collar, white collar, blue collar, small and medium enterprise managers), occupational status (college students), etc.

Property status: You can choose to target people who own cars or houses.

Consumption type (people who have purchased virtual goods/physical e-commerce goods in the past month): You can choose to target people who have purchased virtual goods or physical e-commerce goods.

Game spending power: You can choose to target people with high or average game paying power.

Consumption level: You can choose to target people with high or low consumption levels.

Residential community price: You can choose to target people whose residential community prices fall within a certain range.

Next is the second category - targeting methods based on user behavior:

Behavior: All online and offline user behavior tracks that can be recorded in the past year. (i.e. what the user has done)

Note: We can choose user behavior scenarios: APP, e-commerce, information. Time limit for behavior: If a certain behavior occurs within seven days, the longest time limit can be selected to be within one year, or there is no time limit. Behavior intensity: unlimited or high intensity.

Interest: In the past year, users have a tendency to take a specific attitude towards a certain thing or activity, and have made a long-term investment of time or money for it. (i.e. what the user likes), up to 250 items can be selected.

Intention: An individual's intention to purchase or participate in a specific product or service in the past month, describing the individual's conditional and feasible choice tendency based on the needs of survival and development. (i.e. what the user intends to do)

Then let’s look at the third category – the targeting method under custom groups:

Targeted group: only target this group of people.

Excluded groups: Do not target this group of people.

Next is the fourth category - orientation under weather orientation:

Temperature: You can select the desired group of people by temperature.

Clothing Index : You can select the desired group of people by the clothing index.

Makeup Index: You can choose the people you want by the makeup index.

Weather: You can select the desired group of people based on weather conditions.

Finally, there is the fifth category - device targeting:

Device brand and model: You can select the desired group of people by device brand and model, for example, only target users whose mobile phones are Xiaomi.

Online surfing scenario: You can select the desired group of people based on the online surfing scenario, for example, only targeting people who surf the Internet at home.

Operating system version: You can select the desired group of people by the operating system version, such as only delivering to iOS or only to Android users.

Connection method: You can select the desired group of people by the connection method, such as only targeting users who access the Internet via WiFi.

Mobile operators: You can select the desired group of people through mobile operators, for example, we only target mobile users or Unicom users.

Device price: You can select the desired group of people based on the device price. You can choose to target users whose device prices are between 3500-4500, or users whose device prices are above 4500, etc.

Now that we have introduced the orientation of information flow, let's talk about the directional application of information flow:

1. Usage ideas: We need to set corresponding marketing goals based on the product - then analyze and study the characteristics of the product's target audience - set corresponding targeted tags based on our conclusions - then place advertisements - perform data analysis - adjust and optimize.

2. Suggestions for use: After we have selected all the targeting tags, there will be an estimate of the exposed population on the right. We should pay attention to this data. If it is too low, it means that we have too many targeting conditions and the probability of the ad being displayed is very low. We need to make some adjustments to the targeting conditions. It is generally recommended that this data be above 10 million.

3. Note: Information flow can be divided into primary targeting and secondary targeting. Primary targeting refers to options such as region, device targeting, and networking methods. Secondary targeting refers to options after primary targeting is expanded. The first-level orientation takes the intersection of options, that is, all the selected conditions must be met at the same time; the second-level orientation takes the union of options, that is, only one of the selected conditions must be met.

Let’s use an example to see how to use information flow targeting correctly: we previously had a client who was engaged in fresh food e-commerce, and the regions where they launched their products were Beijing and Shanghai. After our research, we found that the core group is 24-30 years old, mainly female, and 70% of users have college degrees or above. Now let’s take a look at the estimated coverage population under these conditions.

We can see that under these restrictions, our estimated coverage population is only one million, and the chance of the ad being delivered is very small. After one week of delivery, there is little exposure and the cost is high.

Later, we conducted a detailed study on targeting. Since the client's service scope was limited to these two cities, the regional targeting could not be modified. Since the age range of Internet users was still very large, the age requirement was appropriately relaxed. Although the core users were female, there were also male users, so the gender restriction was removed. For the download of fresh food APP, there is no need to impose special restrictions on educational background, so we chose no restrictions on educational background. Then, we estimated that the coverage population would increase. After rebuilding the advertisement, the advertising growth rate increased significantly, and the overall cost was within the customer's assessment range. The following is a comparison of the data before and after our adjustment.