Common Mistakes Beginners Make Learning Business Analytics

Analysis brief (fill before touching data)

Business question:
  Why did repeat purchases fall in the last quarter?
Decision it supports:
  Whether to change the loyalty offer
Audience:
  Marketing manager (non-technical)
Success metric:
  % of customers with 2+ orders within 90 days
Data needed:
  orders, customers, offers
Deadline:
  Friday
Out of scope:
  New customer acquisition

Metric definition sheet

Metric: Active customer
Definition: Placed at least one order in the last 30 days
Source table: orders
Exclusions: test accounts, cancelled orders
Owner: Sales analytics

Metric: Repeat purchase rate
Definition: Customers with 2+ orders in 90 days
            / customers with 1+ order in 90 days
Notes: Agreed with marketing on [date]

-- 1. Is customer_id unique in customers? (should return nothing)
SELECT customer_id, COUNT(*)
FROM customers
GROUP BY customer_id
HAVING COUNT(*) > 1;

-- 2. Row counts before and after joining
SELECT COUNT(*) FROM orders;
SELECT COUNT(*)
FROM orders o
JOIN customers c ON o.customer_id = c.customer_id;

-- 3. Missing values in key columns
SELECT
  SUM(CASE WHEN amount IS NULL THEN 1 ELSE 0 END)      AS null_amount,
  SUM(CASE WHEN order_date IS NULL THEN 1 ELSE 0 END)  AS null_date
FROM orders;

Quick Excel data checks

Duplicates in an ID column (in a helper column):
=COUNTIF($A:$A, A2)>1

Count blanks in a column:
=COUNTBLANK(C2:C5000)

Find numbers stored as text:
=ISTEXT(D2)     -> TRUE means it needs converting

Reconcile a total with the source report:
=SUM(D2:D5000) - [source total]   -> should be 0

Use Remove Duplicates only after you know why they exist.

import pandas as pd

orders = pd.read_csv("orders.csv")

# The overall average can hide very different groups
print("Overall average:", round(orders["amount"].mean(), 1))

summary = (orders.groupby("customer_type")["amount"]
                 .agg(["count", "mean", "median"])
                 .round(1))
print(summary)

# Compare with the previous period, not in isolation
orders["month"] = pd.to_datetime(orders["order_date"]).dt.to_period("M")
monthly = orders.groupby("month")["amount"].sum()
print(monthly.pct_change().round(3).tail(3))

One-slide story template

HEADLINE (the answer):
  Repeat purchases fell mainly among customers who joined
  through the festive sale.

EVIDENCE (2 charts max):
  1. Repeat rate by acquisition channel
  2. Repeat rate trend for festive-sale customers

RECOMMENDATION:
  Test a follow-up offer for festive-sale customers at day 30.

NEXT STEP:
  Measure repeat rate for the test group vs a control group.

Before you start

[ ] Business question written in one sentence
[ ] Decision and audience identified
[ ] Metrics defined and agreed
[ ] Data sources listed
[ ] Time period and filters agreed
[ ] Out-of-scope items noted

While analysing

[ ] Data types checked
[ ] Nulls counted and handled consistently
[ ] Duplicates investigated, not just deleted
[ ] Row counts compared before and after joins
[ ] Totals reconciled with a known source
[ ] Segments and distributions checked, not only averages
[ ] Results compared with a previous period or target

Before you share

[ ] Headline states the answer
[ ] Two or three charts, each with a clear title
[ ] Axis labels and units correct
[ ] Correlation not presented as causation
[ ] Recommendation and next step included
[ ] Assumptions and limitations listed
[ ] A colleague has reviewed the logic