Product Analysis -Understanding Product Purchasing Patterns in Retail Brands

Retail brands have an abundance of data about their products and customers. Leveraging this data is essential to retain these customers and sell more products.

The main aim of doing a Product Analysis is to understand how well a brand’s products are doing, their variability and what can be done to improve the KPIs of the products. Understanding the variation in product purchases by different customers is important for creating product-specific campaigns and learning more about the behaviour of the customers and the products they purchase.

Introduction

In this article, I’ll be covering the different sections that are part of a Product Analysis. All these sections tell us some distinct property of the brand’s products. Each section will have a brief introduction about it, what is done to get the data, how it can be used and what kind of visualization would be a good way to display the data. If data is present in tables, it can be extracted for each section using simple SQL and/or Python codes.

What is the use of doing a Product Analysis?

Some important questions related to the brand’s products that need answering to are -

1. Which products are performing the best?

2. Is there consistency in the performance of the product throughout a certain time period?

3. Do customers prefer to buy low or high-cost products more?

4. What products do customers generally buy together?

A product analysis is utilized to address all the above questions and help generate useful insights from the data. We are then able to create product-specific personalized offers for customers.

Components of Product Analysis

Below are the various sections comprising the Product Analysis. Not all would be relevant to every brand and this is also not to be considered as an exhaustive list because there are other aspects also which can be explored.

· Overview of the data -

The first step of any analysis is to see how the overall data looks like. Questions to be addressed under this section would be -

1. What are the top-performing product categories?

2. How do these categories perform in terms of the Key Performance Indicators (KPIs) like Sales, Quantity sold, Customers purchased, and Transactions?

3. How does the trend look at the product categories in terms of Sales and Quantity on a monthly, quarterly, and yearly basis?

4. What is the distribution amongst the product groups and is there any correlation?

All these are important criteria which give us an outline of the data. It is an essential starting point before we can deep dive further into doing the more complex parts of the Product Analysis.

The crucial area here would be how we can show this data. The possible visualizations that can be used are -

§ Tables for the KPIs data. The use of a simple table showing the values of each KPI for the different products would be sufficient here.

§ Line charts for the monthly/quarterly/yearly trends

§ Stacked bar chart to show Quantity or Sales distribution of a product category and its variation over months/quarters/years.

§ Pie chart to show the distribution amongst a product group

Stacked Bar Chart of distribution of Quantity sold

· Top N vs Bottom M (based on Quantity sold or Sales) products comparison -

This section involves comparing products divided into two different groups on the basis of percentiles. Products can be divided either using total quantity sold or total sales or even both combined. Usually, a Top 10 vs Bottom 90 or the Top 20 vs Bottom 80 grouping is done of the products. Both types should be tried and then finalized based on the insights which come up.

To get the aforementioned insights, the following data points should be looked at for the top and bottom groups -quantity contribution, sales distribution, number of customers purchasing and the average price point. From this we would be able to understand what price range do people buy more of and are the highly sought after items more price sensitive or not.

The visualizations here can just be simple bar graphs or pie charts to show the comparison of the values between the Top vs Bottom percentiles.

Average price point and Sales contribution comparison of the 2 percentile groups

· Price Point Segments –

Based on their price points, the products are divided into 4 segments (25 percentile each). This should be done separately for each major product category. On the basis of the segments and KPIs performance, we can determine which price ranges do customers prefer to purchase from and which price range is not performing well and needs to be looked at.

Distribution of the KPIs for the 4 price point segments

· Basket Items Analysis –

This section would help to understand how the number of products and the type of products varies in different basket sizes.

The first part here would be to see how many items do customers buy at one time and what the sales contribution from these is. Different buckets of the number of items bought together should be made which would vary for different brands. For example, a footwear brand could have buckets as: 1 item, 2 items, 3 items, 4 items and 5+ items; while a supermarket brand could have: 1to3 items, 4to6 items, 7to10 items, 11to15 items and 16+ items. A combo chart would be useful here for showing the data with the % of customers on bar chart and the % of sales as a line chart for the different buckets of items.

For the second part we can see the various combination of products in a basket, and how many such % of these baskets has such a combination. This can be used to identify which are the products bought more in smaller basket sizes and which products are bought more in combination for higher basket sizes.

Stacked bar chart for each bucket with % of bills shown for a combination of products.

· Price Sensitivity Analysis –

The last section is to understand the price variation of products. For each product category (similar products), calculate the mean and standard deviation (s.d.) of the prices for that category. Based on these values, each product would lie either under Mid-Market (price in the range of mean +-s.d.) or Price Sensitive (< mean - s.d.) or Upper Market (> mean+s.d.).

Based on the KPI performance of the segments, we can find out what is the preference of customers, what price range is working well in a category and which customers to target to push them to buy higher priced items (Price Sensitive to Mid-Market and Mid-Market to Up Market). A tabular representation of each segment with the KPIs would work.

Bell Curve of the segments and the products coming under them

Final Steps and Conclusion

The last part is to combine all the insights acquired from the analysis done and prepare an engagement strategy for each customer. Such a strategy would include what is the next best offer for that customer and on what products the offer should be given. Customers can be grouped into segments based on similar characteristics using RFM segmentation or some other method and targeted together.