1. Analysis background introduction
This case comes from a leading retail
enterprise, focusing on providing consumers with diversified high-quality goods
through online and offline channels. Since its establishment, the company has
always been committed to optimizing the shopping experience, using advanced
scientific and technological means and data analysis to accurately meet the
needs of customers, hoping to provide important business opportunity insights
to the leadership through shopping basket analysis, and improve customers'
shopping experience to achieve sales performance growth.
2. Statement of key issues
The retail environment is changing rapidly,
and consumers' purchasing behavior and preferences may also change at any time.
Retail enterprises must establish a rapid response mechanism and adjust
marketing strategies and product recommendations in time.
Retail enterprises usually have a variety
of sales channels (such as online e-commerce platforms and offline physical
stores), which leads to data dispersion and difficulty in integration.
When applying the results of shopping
basket analysis in practice, how to effectively implement and monitor the
effect of marketing strategies is also a major challenge, and it is necessary
to systematically collect data monitoring results for a long time.
3. Analyze the plan
3.1 Select key data indicators. Collect
product sales order data.
|
Serial
number
|
Name
of the indicator
|
Paraphrase
|
Analysis
angle
|
|
1
|
Degree
of support
|
Refers
to the proportion of A and B products in all orders (the probability of
purchasing AB goods at the same time): the number of orders purchased by AB
at the same time/the number of all orders
|
The
degree of support reflects the frequency of joint purchases of A and B
commodities. High support means that these two goods are usually purchased
together, which may be because they are complementary products or consumers
tend to buy them together. By analyzing the degree of support, merchants can
understand which product combinations are the most common, and then make
bundled promotions or set up joint recommendations for these combinations.
|
|
2
|
Degree
of confidence
|
It
refers to the proportion of buying B under the premise of purchasing A
(buying A goods and then buyingProbability of buying commodity B): the number
of orders purchased by AB at the same time/the number of all orders including
commodity A.
|
Credibility
can help merchants understand when consumers choose a certain product (such
as product A)After that, the possibility of buying another product (such as
commodity B). High confidence means that buyers of product A are more
inclined to buy product B, and merchants can make personalized
recommendations based on this to improve the opportunity of cross-selling.
For example, if consumers are likely to buy headphones (B product) after
buying a certain brand of mobile phone (product A), then the headphones can
be displayed on the shopping cart page through the recommendation system.
|
|
3
|
Degree
of promotion
|
It
refers to the proportion of buying B under the premise of buying A (the
probability of buying product A and then buying product B is compared with
the natural probability of buying product B): confidence/order ratio of
buying B.
|
The
degree of improvement can help merchants evaluate the intensity of the
relationship between goods. If the degree of promotion is greater than 1, it
means that after purchasing commodity A, the probability of buying commodity
B is higher than the probability of naturally purchasing commodity B,
indicating that there is a strong correlation between commodity A and
commodity B. The higher the degree of promotion, the stronger the connection
between the two products in the shopping cart, and merchants can consider
bundling them together or recommending them to consumers. If the degree of
improvement is close to 1, it means that there is no significant correlation
between the purchase of goods A and B.
|
Explanation: The indicators selected in
this case are common indicators in the analysis. In the analysis work, priority
should be given to the indicators that have the greatest impact on the business
to ensure that the purpose of the analysis is consistent with the business
objectives and key performance.
3.2 Power BI Visualization Scheme
Note: The DEMO page data is simulated data,
which is for reference only to the analysis angle and Power BI function
display, and does not involve any actual business data.
4. Analysis and interpretation
Support: Help to identify the combination
of frequently purchased goods.
Credibility: reveals the probability that
consumers will buy another product when buying a product.
Elevation: Measure the degree of relevance
of a product portfolio to help determine whether they are worth bundling or
recommending.
5. Application effect
Increase sales:Shopping basket analysis can help identify the correlation between
goods, thus prompting consumers to make more cross-purchases. For example,
through analysis, it is found that a brand of shampoo and conditioner are
usually purchased together, and enterprises can set up bundles or joint
promotions for such goods to increase the average value of each shopping
basket.
Optimize inventory management: By using the results of shopping basket analysis, enterprises can
carry out inventory management more accurately. Understanding which goods are
often purchased together can help to formulate a more reasonable inventory
configuration, avoid out of stock or lag, and improve the inventory turnover
rate.
Personalized recommendation:Shopping basket analysis can provide information on consumers'
purchasing habits and preferences, and enterprises can make personalized
recommendations based on these data. On the e-commerce platform, according to
consumers' past purchase records, recommend relevant or complementary products
to improve the purchase conversion rate of customers.
Improve customer satisfaction:By understanding consumers' purchasing behavior, enterprises can
provide product portfolios and purchase suggestions that are more in line with
customer needs in the shopping process, so as to improve the overall shopping
experience of users and improve customer satisfaction and brand loyalty.
Optimize promotional activities:Shopping basket analysis can provide data support for the planning
of promotional activities.By analyzing which product combinations are highly
sold in a specific period of time, enterprises can carry out targeted promotion
design to maximize the marketing effect and consumers' interest in purchasing.
Insight into the market trend:Regular shopping basket analysis will help enterprises track changes
in market trends and consumer preferences. By analyzing the purchase data of
different periods of time, you can find emerging hot-selling products or
categories, quickly respond to the market, and maintain competitive advantages.
Support decision-making:Management can make more scientific and accurate business decisions
based on the data insights of shopping basket analysis. This includes the
formulation of key strategies such as product on-shelf, marketing, channel
selection, etc. to ensure the efficient allocation of resources and the maximum
return.
Summary:Retail
enterprises can identify the commodity portfolio that needs to be analyzed
through models, but the large-scale quantity of commodities and diversified
consumer behavior make it complicated to choose the appropriate commodity
portfolio. Too much combination analysis may lead to "data
overfitting", that is, the model performs well in training data, but its
prediction ability in real scenarios declines. Even if the commodity portfolio
with high support, confidence and promotion is successfully calculated, it is
still a challenge to interpret these results and translate them into practical
business decisions. Retailers need to have an in-depth understanding of the
logic behind these indicators and be able to transform them into feasible
marketing strategies.