Big data business learning experiments can change the supermarket industry_Data Analyst Exam
Consumption power is insufficient, the profit decline, the erosion of e-commerce, large supermarkets, the road of breakthrough in where? Mastering the five directions of "business experimental learning" based on big data analytics will bring hundreds of millions of dollars of economic benefits to retail enterprises.
Retailers may generate exciting but risky ideas every day. And what will the results be?
Business Learning Experiments with Big Data
Taiwan's Family Convenience Stores has more than 2,000 stores in Taiwan. In their survey, they found that there was significant consumer demand for freshly brewed coffee. For Family, is it better to offer freshly brewed coffee in all stores, or to prioritize the purchase of these coffee machines in certain stores? Will the big sales of freshly brewed coffee crowd out sales of other coffee and beverage products in stores?
In the past, experienced managers formed strategic assumptions through intuition. Now, with APT's big-data analytics-based "Experiment and Learn" software, they can do more careful data validation -- picking a few stores to test different approaches to the new idea in small quantities, learning from the results, and figuring out what impact it will have on revenue, net profit, return on investment, market share, and so on, Learn from the results to find the most successful version for the company in terms of revenue, net profit, return on investment, and market share, and then design an optimal rollout plan and push it forward. This systematic approach is called "experimental learning".
First, APT entered information from more than 2,000 stores into a data platform, and within two weeks converted the transaction records from thousands of Family Mart cash registers over the past three years into a database; and within two months created an attribute profile for each store, including store size, age of the surrounding population, employment status, proximity to competing businesses, location, and other factors related to the business. The results of this process are
After that, APT selects a group of stores and divides them into an experimental group and a reference group. Through Business Experimentation Learning, APT's questions about the addition of coffee machines at Family Mart were quickly answered, as they found out how to prioritize the introduction of coffee machines, and learned which locations didn't need coffee machines at all, and which stores improved their net performance after the introduction of coffee machines.
A similar decision was made at SUBWAY, which in 2008 was planning to launch a foot-long $5 sub in North America in hopes of generating more business from a special offer, but was concerned that the promotion would hurt sales of higher-priced sandwiches.
In order to test the feasibility of the business innovation, they also designed an experiment with APT. After collecting information about all the restaurants, the classification tool selected an experimental group and a reference group that had nearly the same external factors such as the size of the restaurant, geographic location, surrounding demographics, and characteristics of the business district. During the experimental cycle, the overall sales performance of the restaurants, the sales of $5 submarine burgers, and the sales of other sandwiches were compared, respectively. At the end of the experiment, a comprehensive financial calculation tool on the platform can provide results within a few hours, showing that this special submarine burger really brings performance improvement to the business.
APT (Applied Predictive Technologies), the originator of "experimental learning for business," was founded in 1999 and focuses primarily on the mature North American market, as a company that helps companies optimize their strategies through "experimental learning. APT (Applied Predictive Technologies, APT), the originator of APT, was founded in 1999, with its main business focusing on the mature North American market, and is a software company that assists companies in optimizing their strategies through "experimental learning". APT currently serves customers in more than 10 major markets in the Asia Pacific region, including China, Japan, Singapore and Australia.
5 Directions for Learning from Business Experiments
In the current situation of lack of consumer power and rising costs, coupled with the erosion of large supermarkets suffering from e-commerce, large supermarkets have begun to seek a new path of development. "For store-based hypermarkets, the hundreds of distribution networks are a highly valuable source of knowledge and information from which they can learn from their successes and failures." APT vice president and head of Asia Pacific, Lee Chin Hong, said. He said the Business Experiment Learning will help the supermarket industry in the following five ways:
1. Improve O2O ROI.The intent of O2O integration was to increase revenue streams by selling online and attracting new customers who would not have otherwise traveled to these stores. But the fact is the opposite, some stores new customers did not grow at the same time, but also incurred huge operating costs. Li Zhanhong said, if the program before the promotion of a small number of sample stores to carry out "online shopping, door-to-door pick-up" experimental test, it will be possible to reduce the risk of the implementation of these innovative plans. Then according to the data, for the significant profits of the store to lock and accurate promotion, will be more effective to enhance the ROI of the program. supermarkets have a lot of unique advantages, such as those with a short shelf life of food, fresh food, perishable and consumable goods, are the shortcomings of the e-commerce industry, supermarkets may wish to do more to make it stronger.
2. Optimize posters. Poster promotion is the usual supermarket techniques, but it really brings profits? Poster design mainly consider two issues, the poster issued by the crowd and into which products. For the former, the previous poster distribution of the mainstream population is responsible for the family purchasing aunts, they are comparing a number of supermarket posters, buy the cheapest goods; for the latter, may be a special price of beer to obtain a huge growth in sales, the supermarket may think that the promotion is quite successful. In fact, the truth is likely to be that most consumers only bought the special-priced beer and did not buy any other goods along the way, and behind the false prosperity is the supermarket's gross profit is damaged.
Through "experimental learning," we learned that it's important to differentiate between different customer segments when distributing posters, and that attracting more young people into the supermarkets will bring in more profits; and that the supermarkets need to do a lot of transactional data analysis on what to put in the posters in order to design which offers to put in the posters. For example, putting chips or salad next to beer may increase the amount of single sales for customers.
3. Expanding own categories. Private labeling is a major asset of supermarket operations, and in this area, Wal-Mart, CR Vanguard, Watson's, Mannings and other supermarkets and chains have been plowing ahead in recent years. According to the study, the U.S. supermarket industry's private label sales growth of more than 18% in the past three years, more than twice the growth rate of other branded goods.
For retailers, expanding their own categories can't simply be modeled after successful manufacturers' products, which can cause a lot of products to sell poorly. In order for expansion of own product lines to create more benefits, supermarket managers need to strategically optimize their merchandise mix to avoid a decline in transactions that could hurt gross margins, said Li Zhanhong. Supermarket companies can use transaction data analysis to understand which items often appear in higher-value transactions and are often purchased by high-value customers to optimize private label growth strategies.
Taking Sam's Club as an example, Sam's Club's business goal in China is to meet the needs of commercial members in sales, office and distribution of benefits, as well as the needs of high-income individual members. As a result, Sam's Club sells 8-packs of "Member's Premium" soap and full cases of copy paper, and real-world sales figures prove that these revamped products are examples of private label success.
4. Accurate ethnic group promotion. By binding member information and transaction data, supermarkets can carry out group promotions, which can accurately understand the substitution effect of each promotion, the phenomenon of delayed consumption, as well as different consumers in the factors on the variation.
For example, a special beer promotion may cause consumers to buy only a large number of discounted beers, leading to a decline in future sales, which in turn hurts gross margins. Testing on a small percentage of customers will reduce the risk of a significant drop in gross profit and can assist a business in understanding which outlets can generate additional benefits and which outlets are merely giving away to customers who would have purchased the item anyway, ultimately leading to a drop in gross profit.
5. Measuring SNS marketing. As the post-80s are becoming a major customer base for supermarkets, many supermarkets are using microblogging, WeChat and other SNS tools as delivery channels to provide information on promotions and special offers. Companies can conduct experiments to compare the marketing performance of the experimental group with advertising and the control group without advertising to avoid wasting advertising investment and to focus on more effective markets and customer segments.
The above is what I have shared with you about the big data under the business learning experiment can change the supermarket industry related content, more information can be concerned about the Global Green Ivy to share more dry goods