Stop 05 · Data Analysis for Business AnalystsTopics · 9/10

Forecasting & Trend Analysis

In short

'How many support agents will we need in December?' Every business asks questions about the future. Simple trend analysis gives a reasonable answer, as long as you're honest about how uncertain it is.

Projecting future values based on historical patterns — from a simple trend line to more structured methods, always paired with an honest sense of the forecast's uncertainty.

The purple blob holds an umbrella and looks at dark rain clouds rolling in.

Think of it like this

The monsoon forecast

The weather department doesn't promise rain at 3 pm on a date. It says the monsoon will likely arrive around early June, give or take a few days, based on past patterns. A good business forecast is like that: a likely range, not a promise.

Watch it happen

Smoothing monthly orders with a 3-month moving average

120

132

128

141

150

149

1/5Six months of orders, in hundreds. The numbers bounce around, so the trend isn't obvious.

The big ideas

What you need to know

Trend

The general direction over time, up, down or flat, once you look past the month-to-month bumps.

Seasonality

Patterns that repeat, like festive season peaks or exam month dips. Compare December with last December, not with November.

Simple methods first

Moving averages and a straight trend line often do the job. Complex models need more data and expertise.

Always give a range

Say 'between 14,000 and 16,000 orders' rather than '15,212'. Fake precision makes people trust a forecast too much.

See it in code

A tiny example

excel

1=AVERAGE(B2:B4)2=FORECAST.LINEAR(7, B2:B7, A2:A7)3=TREND(B2:B7, A2:A7, 7)

Column A holds month numbers 1 to 6, column B the orders. Line 1 is one moving-average value. Lines 2 and 3 project month 7 using a straight trend line.

Save yourself the pain

Common mistakes

Oops

Drawing a straight line from three months of data into next year.

Do this instead

Use as much history as you have, check for seasonality, and keep forecasts short-range.

Oops

Presenting one exact number as 'the forecast'.

Do this instead

Give a likely range and state your assumptions, like 'assumes no price change'.

Your mission

Forecast your own spending

~30 min

Quick check

An online store wants to forecast November orders. Which comparison is most useful?

Go deeper

1Build a simple trend line forecast for a sample time series in Excel.
2Learn why every forecast should come with a confidence range, not a single precise number.
3Practice explaining forecast assumptions clearly to a non-technical stakeholder.