How Accurate Should a Forecast Be?

Owners often abandon forecasting after one bad miss. The fundamental misunderstanding: a forecast doesn't need to be precise to be useful - it needs to be calibrated. Knowing what accuracy is realistic at each horizon turns forecasting from a frustrating exercise into a reliable planning tool.

Realistic accuracy targets by horizon

These are revenue accuracy targets - some line items (recurring revenue, contracted) are easier to predict; some (one-off deals, marketing-driven sales) are harder. Set tighter targets for predictable line items, looser for the volatile ones.

Why accuracy degrades with distance

Three reasons forecasts get fuzzier the farther out you go:

• Compounding uncertainty. A 5% miss at month 1 compounds with another 5% miss at month 2. By month 6, the ranges spread significantly.
• Decision freedom.Six months from now, you might have made decisions (a new hire, a price change, a new market) that your current forecast doesn't reflect.
• Market drift.Customer behavior, competitor actions, and economic conditions all change over time, and each one's direction is hard to predict.

You can't outwork these forces with better models. The right response is to forecast tighter at short horizons and treat long horizons as directional.

Accuracy comes from updating

The biggest misconception about forecast accuracy: that it comes from building a better initial model. It doesn't. It comes from revising assumptions as new information arrives.

A forecast built once and not updated becomes increasingly wrong every month. A forecast revised monthly with actuals stays approximately right because the assumptions that were wrong get corrected.

Variance analysis: diagnosing misses

When the forecast misses, the question isn't whether it was wrong - it's which assumption was wrong. That's what variance analysis answers.

For each material variance:

• Quantify - dollar amount and percentage
• Decompose - was it volume? Price? Timing? Mix?
• Identify the assumption - what specifically did we assume that turned out wrong?
• Update going forward - revise the forecast for the corrected assumption

Done monthly, variance analysis turns the forecast into a learning loop. Each month's actuals improve next month's assumptions.

Watch for systematic bias

A forecast that misses in the same direction repeatedly has a structural problem, not a precision problem.

• Consistently too optimistic on revenue - the input forecast is biased. Tighten the assumption.
• Consistently too low on expenses - the expense base has hidden items. Audit and adjust.
• Consistently late on timing - sales cycle is longer than forecast. Lengthen the build.

Three consecutive misses in the same direction is a signal to recalibrate, not just to revise.

Common mistakes

1. Treating accuracy as the goal

The goal of forecasting is better decisions, not precise predictions. A 90% accurate forecast that doesn't change any decisions is less valuable than a 70% accurate one that catches a cash problem early.

2. Adding precision where it doesn't exist

Forecasting revenue at $1,234,567 implies false precision. Round to meaningful figures. "$1.2M ± 15%" is more honest and more useful.

3. Hiding misses

Optimistic owners often revise the forecast downward to match underperformance, masking the original miss. The original miss is the signal worth investigating.

Related concepts

• What Is Financial Forecasting - the foundational concept.
• Revenue Forecasting Methods - the methods that affect achievable accuracy.
• Scenario Planning Explained - using ranges instead of single numbers.
• Leading vs Lagging Indicators - what to watch to catch forecast misses early.
• Financial Forecasting for Small Businesses - the broader guide.