JD.com product manager shares shopping cart marketing methods based on big data

If artificial intelligence algorithm models are added to shopping carts, what new marketing methods will there be?

The concept of online shopping carts originated from the physical entities of offline supermarkets. The main function of the shopping cart is to facilitate consumers to shop on the website, make it easy for product settlement and selection of intended products. The shopping cart serves as a transfer station for commodity transactions. Hundreds of millions of users on the entire network add products of their choice to the shopping cart every day. In an instant, sales of over 100 million can be generated.

Faced with such a large flow, all major manufacturers are thinking about this gold mine. In the past, the main product forms of shopping cart marketing based on big data were Guess You Like and Recommend for You, both of which revolved around the user's shopping behavior, user product hobbies and user portrait attributes. After big data analysis, the system intelligently Recommend products that meet the user's taste. However, this marketing method is based on the products in the shopping cart or other products recommended by the user's portrait. It is not a marketing strategy for the products in the shopping cart. This method is a bit putting the cart before the horse.

The following article combines the author’s work experience to describe how to use AI technology to design a shopping cart marketing product based on the products in the shopping cart.

1. Marketing process

The merchant checks the additional purchase data, such as the number of additional purchases and the number of additional purchases. The system automatically analyzes the portrait data of the additional purchasers, and the groups can be labeled.

The merchant can create marketing for different labeled groups according to its own needs. For example, it can choose new customers, old customers, and user groups aged 15 to 25, and provide services with a price reduction of 40 yuan

After creating an event, it will be reached to the corresponding covered groups.

The next day, the merchant can view the corresponding marketing data. At the same time, it can compare the natural conversion rate and the conversion rate after promotion

2. Merchant-side insight into shopping cart data

The shopping cart carries all product information, including product name, price, Shops, promotions, orders and coupons, etc. When performing big data analysis, it is necessary to disassemble and clean the data to extract valuable parts. Each item in the shopping cart can be regarded as an entity. Some people may have added the same item to the shopping cart at different places and at different times. This shows that these groups are interested in this product and may place an order, but they are not ready for it. There are also some people who have added products to their shopping carts early, but have never placed an order. Using big data technology, you can label additional purchasers and carry out precise marketing strategies for people with different labels, which can improve shopping cart conversions to a certain extent.

How to proceed? Follow the following steps:

Inventory of merchant purchase data

Products need to consider both the merchant side and the user side. First of all, merchants need to understand their own product status, sales, additional purchase data, etc., so that they can make targeted marketing strategies.

The merchant can see the number of people who have purchased additional products in their store, calculate in real time how many people have purchased a certain product in their shopping carts, the total number of additional purchases in real time, and the real-time inventory. You can also check the natural conversion rate of these products without intervention (yesterday's conversion rate of consumers who purchased the product in the past 15 days).

The products in the list are sorted from high to low by the number of people who have added to the purchase. The more people who have added to the purchase, the more popular the product is. Marketing intervention for products with a large number of additional purchases will have better results. Of course, some products that have been removed from the shelves and have expired will be automatically removed.

The portrait section summarizes the account information of all users, and provides five dimensions of portrait latitude, new customers, gender, consumption level, Taobao level, and region. Portraits label users, and using these labels, different marketing actions can be performed on them. For specific grouping strategies, please see my previous article "How to play member task marketing based on big data?"

Merchants can market each product separately and deliver it to customers based on their own brand conditions. Target specific groups of people and conduct low-price, promotional interventions.

Based on the selection of tags, the system will predict in advance the conversion ratio of people who have added purchases based on user behavior data on the website. Through machine learning, it can automatically filter out the user groups with low conversion probability.

The calculation rule here is based on whether the user has purchased the same product before, or whether the user added it to the shopping cart for price comparison.

Promotional effect analysis

Through user grouping, you can understand the characteristics of your customer groups, what kind of people have purchased your products or are interested in your products. Precision marketing can Hold these customers firmly in your hands and use means to intervene with them. For merchants, performance analysis data is also needed.

Encircled number of people: the group of people covered by the activity. The system can calculate the number of people that can be reached according to the event tags and promotional prices

Number of transactions: the number of people who submitted orders after the event started

Number of people reached: the final number of people reached through push and message center Number of people reached

Transaction amount: total amount of transaction orders

3. The logic of reaching consumers

Of course, all activities organized by merchants are It needs to finally reach the consumer end. Based on shopping cart marketing, the optimal solution for his approach is to reach users on the items in the shopping cart that participate in the event, but only the users covered will be reached. The contact methods are divided into:

Shopping cart icon contact

The shopping cart displays "limited time" icon reminder and real-time promotion countdown reminder. Time reminders can enhance consumers' sense of urgency in shopping, and improve bustling conversions through promotions and a sense of time.

Price reduction reminder, the specific price reduction amount is displayed in red letters, and is an emphatically reminded.

Message center reach

When the event is started, push marketing content will be received in the message center, and the content will be sent to the covered people in real time. Clicking on the message content will jump to the shopping cart. However, this push approach is not very effective and the click-through rate is low. For specific contact methods, you can also read my previous article "Member Task Marketing Based on Big Data, How to Play?"

Conclusion

There are many ways to play the shopping cart , you should evaluate what improvements need to be made at the current stage based on your own products and R&D capabilities. But the core goal is the same, to convert as many shopping cart items as possible into orders and bring actual benefits.