Marketing-Insights-For-E-Commerce-Company-Case-Study

Business Context: One of the leading E-Commerce Company would like to get marketing insights from the data to define marketing strategies going forward. Also, expecting to build an analytical dashboard to monitor various KPI’s & business metrics. Business Objective: The e-commerce company is expecting below analysis using the data

1. Calculate Invoice amount or sale_amount or revenue for each transaction and item level  Invoice Value =(( QuantityAvg_price)(1-Dicount_pct)*(1+GST))+Delivery_Charges

2. Perform Detailed exploratory analysis  Understanding how many customers acquired every month  Understand the retention of customers on month on month basis  How the revenues from existing/new customers on month on month basis  How the discounts playing role in the revenues?  Analyse KPI’s like Revenue, number of orders, average order value, number of customers (existing/new), quantity, by category, by month, by week, by day etc…  Understand the trends/seasonality of sales by category, location, month etc…  How number order varies and sales with different days?  Calculate the Revenue, Marketing spend, percentage of marketing spend out of revenue, Tax, percentage of delivery charges by month.  How marketing spend is impacting on revenue?  Which product was appeared in the transactions?  Which product was purchased mostly based on the quantity?

3. Performing Customer Segmentation  Heuristic (Value based, RFM) – Divide the customers into Premium, Gold, Silver, Standard customers and define strategy on the same.  Scientific (Using K-Means) & Understand the profiles. Define strategy for each segment.

4. Predicting Customer Lifetime Value (Low Value/Medium Value/High Value)  First define dependent variable with categories low value, medium value, high value using customer revenue.  Then perform Classification model

5. Cross-Selling (Which products are selling together)  You can perform exploratory analysis & market basket analysis to understand which of items can be bundled together.

6. Predicting Next Purchase Day(How soon each customer can visit the store (0-30 days, 30-60 days, 60-90 days, 90+ days)  For this, we need create dependent variable at customer level (average days per one transaction for only repeat customers and divide into groups 0-30 days, 30-60 days, 60-90 days and 90+ days) then build classification model to predict next purchase of given customer.

7. Perform cohort analysis by defining below cohorts  Customers who started in each month and understand their behaviour  Which Month cohort has maximum retention?

Pl download following files:

1. Data files in Data folder
2. Code in .ipynb format
3. Factor loadings in .csv
4. Models as pickle objects