I analyzed 1 million retail transactions — 29.6% of…
0 reactions
I analyzed 1 million retail transactions — 29.6% of customers generate 80.8% of revenue
Performed RFM (Recency, Frequency, Monetary) customer
segmentation on 1,067,371 real transactions from the
UCI Online Retail II dataset.
After cleaning — removed cancelled orders, missing
customer IDs, and invalid prices — analyzed 805,549
transactions across 5,878 unique customers.
Key findings:
1. Champions (29.6% of customers) generate 80.8% of
total revenue. Classic Pareto distribution confirmed.
2. Champions bought on average 38 days ago, made 15
purchases, and spent GBP 8,244 each.
3. 1,784 At Risk customers represent GBP 829,843 in
potentially recoverable annual spend. These customers
bought 307 days ago and are drifting toward Lost.
4. Lost customers (9.8%) bought 548 days ago with only
1 purchase each. Recovery ROI is extremely low.
RFM scores each customer 1-4 on Recency, Frequency,
and Monetary using quartiles. Total score 3-12 maps
to segments.
Tools: Python, Pandas, Matplotlib, Seaborn
Data: UCI Online Retail II (public dataset)
Full project with code:
github.com/surendrasinghdata/rfm-customer-segmentation http://github.com/surendrasinghdata/rfm-customer-segmentation
#education
source
Performed RFM (Recency, Frequency, Monetary) customer
segmentation on 1,067,371 real transactions from the
UCI Online Retail II dataset.
After cleaning — removed cancelled orders, missing
customer IDs, and invalid prices — analyzed 805,549
transactions across 5,878 unique customers.
Key findings:
1. Champions (29.6% of customers) generate 80.8% of
total revenue. Classic Pareto distribution confirmed.
2. Champions bought on average 38 days ago, made 15
purchases, and spent GBP 8,244 each.
3. 1,784 At Risk customers represent GBP 829,843 in
potentially recoverable annual spend. These customers
bought 307 days ago and are drifting toward Lost.
4. Lost customers (9.8%) bought 548 days ago with only
1 purchase each. Recovery ROI is extremely low.
RFM scores each customer 1-4 on Recency, Frequency,
and Monetary using quartiles. Total score 3-12 maps
to segments.
Tools: Python, Pandas, Matplotlib, Seaborn
Data: UCI Online Retail II (public dataset)
Full project with code:
github.com/surendrasinghdata/rfm-customer-segmentation http://github.com/surendrasinghdata/rfm-customer-segmentation
#education
source
| 1,000 mlx | |
| 2 | |
| 0 |