This repository consists of the three tasks completed by me during this virtual experience program.
Quantium has had a data partnership with a large supermarket brand for the last few years who provide transactional and customer data. You are an analyst within the Quantium analytics team and are responsible for delivering highly valued data analytics and insights to help the business make strategic decisions.
Conduct analysis on your client's transaction dataset and identify customer purchasing behaviours to generate insights and provide commercial recommendations.
We need to present a strategic recommendation to Julia that is supported by data which she can then use for the upcoming category review however to do so we need to analyse the data to understand the current purchasing trends and behaviours. The client is particularly interested in customer segments and their chip purchasing behaviour. Consider what metrics would help describe the customers’ purchasing behaviour.
Main focus area:
1. Data Cleaning:
- Explore both the transaction and purchase behaviour data file.
- Create and interpret high level summaries of the data.
- Find outliers/anomalies and missing data and deal with it.
- Check data formats and correct if needed.
2. Exploratory Data Analysis:
- Derive extra features such as pack size and brand name from the data and define metrics of interest to enable you to draw insights on who spends on chips and what drives spends for each customer segment.
Transaction data
-
Sales are increasing in the lead up to christmas, but there are no sales on christmas day itself as the shops are closed.
-
Kettle is the brand with the most chips sold
-
170g is the pack size with most chips sold
Purchase Behaviour
-
Top 3 segments (Lifestages + customer segment) with the most sales are:
-
Old Family / Budget Customer
-
Young singles/couples / Mainstream Customer
-
Retirees / Mainstream Customer
-
Frequency of purchase is high for all top 3 segments
-
Older Family/Budget segment has low unique cutomers but the total sale per customer is high leading to overall highest total sale
-
In Young singles/couple Mainstream and Retirees/Mainstream segment, the number of unique customer is high but the total sales per customer is low, leading to overall high total sales.
-
The top brand sold in all 3 segments is Kettle.
-
Top pack size sold in all 3 segments is 175g.
Recommendations
-
Older Families (Focus : Budget Customers)
-
As the frequency of purchase and the sales per customer is high in this segment, to increase sales we can promotion that encourages the customers to buy even more chips per visit.
-
For Mainstream and Premium customers, as the sales per customer is almost same as budget segment but the frequency of customers is less. To increase more frequent visits, discounts can be offered on the top brand and pack size in these segments to increase sales.
-
As the overall number of unique buyers are less in Older families, to retain customers that keep buying, a loyalty program can be started that benefits the buyers.
-
Young singles/couples (Focus : Mainstream Customers)
-
The frequency of purchase and the unique number of buyers are high, but the sales per customer is low. To increase this, discounts and be promoted on top brands and pack size to increase sales.
-
Retirees (Focus : Mainstream Customers)
-
The frequency of purchase and the unique number of buyers are high, but the sales per customer is low. Again to increase this, discounts and be promoted on top brands and pack size to increase sales. Others
-
Kettle is the brand and 170g is the pack size with most sales, so promotion and discounts can be offered on these to increase sales.
Extend your analysis from Task 1 to help you identify benchmark stores that allow you to test the impact of the trial store layouts on customer sales.
Julia has asked us to evaluate the performance of a store trial which was performed in stores 77, 86 and 88.
Main focus area:
- Break down the monthly sales experience for each store by:
- total sales revenue
- total number of customers
- average number of transactions per customer
- Find the best control stores for the above three trial stores using different metrics such as Pearson's correlation or magnitude distance.
- Once you have selected your control stores, compare each trial and control pair during the trial period. You want to test if total sales are significantly different in the trial period and if so, check if the driver of change is more purchasing customers or more purchases per customers etc.
Use your analytics and insights from Task 1 and 2 to prepare a report for your client, the Category Manager.