After several years of operation, I took stock of the gains and found that it can be summarized in two words: customers & data. The previous two articles talked about customer-related content. This time I will talk about how to use data for marketing.
Many times when chatting with others, as soon as I mentioned that I was engaged in data marketing, the other party immediately looked at me like a mathematician, and then started to think of "cluster analysis, regression analysis, group preference" in his mind. "And so on. As we chatted, and found that I didn’t mention those things at all, they would start asking me: “What exactly do you mean by data marketing?”
I won’t go into details about theory. Recently, I have been discussing Taobao Supermarket a lot with my Taobao brothers. I will use an online supermarket as an example to see what data marketing does.
The first level of data-based marketing: Thousands of people have the same face - Thousands of people have the same face
The second layer of data marketing: customer life cycle management
The third level of data marketing: cultivating old customers
The first level of data-based marketing: Thousands of people have the same face - Thousands of people have the same face
The first major role of data-based marketing is that it can segment target customers more finely and accurately, making our content in the promotion process more relevant to buyers, and changing operations from one-size-fits-all to one-size-fits-all. It is one of the main goals of digital marketing.
I previously received a promotional email from Store No. 1, as follows:
I think everyone receives and sends out a lot of emails like this every day. We won’t discuss the email itself today. Let’s take a look at how to make this email better through data marketing.
Assume that this email is to be sent to 1 million members. If we want to get the highest purchase rate, the best way is to send 1 million personalized emails to 1 million people and promote different products to each one. But this method is obviously not feasible in actual operation. Therefore, data-based marketing is about finding an operable marketing method that allows us to get the highest purchase rate.
How can the above email sent to 1 million people have a higher conversion rate?
Let’s look at a slightly better method first:
We divided the customers into 4 groups according to age and gender, and then selected some recommended products based on the characteristics of this group, and then made 4 pages and delivered them to the 4 customer groups respectively.
Typical copywriting of typical products for age and gender
15-25 Men's Coke, Coke is 18 yuan/box, 5 yuan cheaper than in supermarkets. You no longer have to carry it yourself. It will be delivered to your door within half a day.
15-25 women various snacks, shampoo
25-35 men beer,
25-35 Women’s shampoo, paper towels, salad oil
A major prerequisite for data marketing is data accumulation. The degree of data accumulation determines how sophisticated data marketing can be. There is very little information we need to accumulate here. Age and gender are just two attributes. (If we add more customer attributes here, we can divide the entire customer base into smaller groups, such as "income", "education level", "occupation", etc. The bank will make customers based on this information. CRM management and risk management. The advantage of segmentation is that it can be more precise, but the disadvantage is that the cost of promotion is higher. At the same time, as the customer base is divided into smaller segments, the marginal benefit of promotion efficiency decreases, so the segmentation is more or less fine. Okay.)
Then, let's do a little better than this:
In each group, we add two more fields: "Browse Category" and "Purchase Category".
Top 3 browsing categories by age and gender Top 3 purchasing categories Typical copywriting
15-25 Men's drinks, paper products, imported food and beverages, imported food, paper products. Coke is 18 yuan/box, which is 5 yuan cheaper than the supermarket. You no longer have to carry it yourself, and it will be delivered to your door within half a day.
15-25 female
25-35 male
25-35 female
Then we can see that "the three most purchased categories by men aged 15-25 are food - beverages, daily chemicals - paper products, food - imported food", and then we can have several choices: Simple Click to select one promotional product from each of "beverages, paper products, and imported food" to make a promotional package and promote it to the entire group.
The data accumulation here is more complicated. First, we need to accumulate the purchase records and browsing records of each member. These two fields alone require a huge database.
Could it be any better?
After we see everyone’s transaction records and browsing records, the business has several ideas that can be further refined:
1. What are the things that customers often look at but don’t buy?
2. What B is the customer most likely to buy after buying A?
To do 1, we need to associate browsing records and purchase records, and set the standard of "category views > N and no purchases" through data analysis.
Doing 2 is more complicated. The most common way is to analyze the purchase records. By analyzing the entire customer base, establish an algorithm. Assume that there are 10,000 people who purchased product A. Analyze the purchase records of these 10,000 people and find out the other products purchased by these people. The most purchased products B, C, and D are recommended on the page of A (this is basically what Amazon does)
Today I will write about the first level first. If you are interested, we will talk about the second level: customer life cycle management.
Source of article: Paidai.com