Bank marketing campaign performance report

#Brief

A Portuguese bank ran phone campaigns selling term deposits. 5,289 of 45,211 calls ended in a subscription (11.7%). The marketing manager wants to know which customers and channels to prioritise in the next campaign.

#What to build

1. Load bank-full.csv (semicolon separated) into Excel. Build pivot tables of subscription rate by job, age band, contact channel, month, and previous campaign outcome (poutcome).

2. Always show the number of calls next to each rate; hide or group segments with fewer than 100 calls.

3. Test at least one difference for significance and give 95% confidence intervals (Excel CONFIDENCE.NORM or Python if you prefer).

4. Note why duration must not be used to decide who to call (it is known only after the call ends; the dataset's own documentation warns about this).

#Deliverables

• The workbook

• A one-page memo: recommendation first, then three supporting numbers, then caveats

• The statistics appendix

#How it is graded

The reviewer checks your rates against the data, then reads the memo as the marketing manager would. A clear, specific recommendation that respects uncertainty scores highest.

#Stretch

Use the bank-additional-full.csv version with economic indicators and see whether results hold in different economic periods.