The café manager asks: 'Last year this weekend was busy - what should we expect this year?'
Can We Predict Next Weekend's Sales?
The café manager needs to know how much inventory and staff to plan for - but the future is uncertain
It's Thursday afternoon. The café manager is looking at her records from the past 104 weekends and turns to Sarah: "Last year this weekend, we had $4,850 in sales. Next weekend is the same holiday weekend - what should I expect this year?"
Why This Matters
- Too much inventory = food waste and lost money
- Too little inventory = missed sales and unhappy customers
- Overstaffing = unnecessary labor costs
- Understaffing = slow service and damaged reputation
Sarah realizes this is exactly the kind of question that data can help answer - but she also knows that past patterns don't guarantee future results. Before building any Excel model, she needs to understand what forecasting can and cannot do.
Here's the key tension the café manager faces:
Why Use Historical Data
- Patterns exist - weekends aren't completely random
- Seasonal trends are real - summer is different from winter
- Growth or decline shows direction of the business
- Better than guessing with no information
Why Past Doesn't Guarantee Future
- New competitors could open nearby
- Weather patterns change from year to year
- Staff quality and management can shift
- Economic conditions affect customer spending
Good forecasting isn't about predicting the exact number - it's about understanding the range of reasonable outcomes and planning for uncertainty.
Sarah pulls up a scatter plot showing monthly sales for the past year. She draws a line through the dots that represents the general direction the data is moving - a trend line.
What a Trend Line Reveals
- Direction: Is sales going up, down, or staying flat?
- Speed: How fast is it changing? (The slope)
- Consistency: How tightly do points cluster around the line?
- Pattern: Any clear relationship between time and sales?
What a Trend Line Does NOT Promise
- Exact predictions: The line is an estimate, not a guarantee
- Account for surprises: It can't predict competitor openings, weather, etc.
- Work forever: Patterns that held in the past might not continue
- Be equally accurate: Predictions further from known data are less reliable
1. The café manager looks at last year's data and sees that the first weekend of November had $4,200 in sales. She says, 'This year we'll probably make around $4,200 again.' What's wrong with this prediction?
2. Sarah notices that café sales have been increasing each month: January $12,000, February $12,800, March $13,600. She predicts April will be $14,400. What assumption is Sarah making?
3. A café owner uses a 10-year trend to predict next year's sales. Why might this be less reliable than using a 2-year trend?
4. Sarah shows the café manager a trend line and says, 'We can predict sales for next month.' The manager asks, 'But what if a competitor opens next door?' What should Sarah acknowledge?
Discussion Prompt (3 minutes):
Think about the café manager's question about next weekend's sales:
- What information from past weekends could help her plan?
- What future events could completely change the prediction?
- Why might relying only on last year's number be risky?
Why This Matters
Before building any Excel forecast model, you must understand what that model can and cannot tell you. A manager who treats a forecast as exact will over-order or under-order. A manager who understands the range and uncertainty can plan for multiple scenarios. The goal isn't to predict the future perfectly - it's to make better decisions given reasonable expectations.
In this lesson, you'll learn to:
Interpret Trend Lines
Read what the slope and direction tell you about the business
Understand "Fit"
Know how consistent the pattern is - not if it's good or bad
Recognize Danger Zones
Know when predictions become unreliable
Connect to Excel
Prepare for the Excel forecasting tools in upcoming lessons