Forecasting Techniques and Sensitivity Checks

Learning objectives

By the end of this chapter you should be able to:

• Construct short-term cash forecasts using trends, averages and seasonal patterns.
• Select suitable forecast drivers and convert them into cash timings (receipts and payments).
• Apply sensitivity and scenario analysis to stress-test a forecast.
• Measure forecast error and use a simple variance review cycle to improve future forecasts.
• Prepare a cash forecast using clear assumptions, transparent workings and sensible checks.

Overview & key concepts

Cash forecasting is a planning tool for managing liquidity. It helps a business anticipate when cash will be received and paid, identify funding needs early, and reduce the risk of unexpected funding gaps.

A strong forecast does not rely on one “perfect” number. It uses a repeatable method, clear assumptions and targeted stress tests to highlight where the plan is most exposed. Sensitivity checks and scenarios are valuable because they reveal which variables matter most and how quickly cash headroom can disappear.

Link to the statement of financial position

Forecasting is mainly about timing. Credit transactions create balances (such as trade receivables) before they become cash.

Assets = Liabilities + Equity

A forecast does not “change” this relationship; it projects how items such as cash and receivables may move over time as sales are made and then settled.

Forecast drivers

Forecast drivers are measurable inputs that explain cash movements. Good drivers are specific, controllable and closely linked to receipts or payments. Examples include:

• Units sold and selling price (cash receipts driver)
• Production volume and material usage (supplier payments driver)
• Headcount and pay rates (payroll driver)
• Credit terms and collection pattern (timing driver)

Drivers matter because a forecast is not just about value; it is about when cash moves. Two businesses with the same projected profit can have very different cash positions depending on credit terms, inventory policy and payment behaviour.

Time series patterns

Time series data is recorded at regular intervals (for example weekly or monthly). It is used to spot patterns that can improve forecast accuracy.

Trend

A trend is the underlying direction over time (upward, downward or flat) after ignoring short-term noise.

Seasonality

Seasonality is a predictable pattern that repeats within a year (for example holiday peaks, summer slowdowns). Forecasts that ignore seasonality often fail at month level—exactly where cash problems occur.

Moving average

A moving average smooths fluctuations by averaging recent periods. It reduces noise and can provide a practical baseline when results are volatile.

Weighted moving average

A weighted moving average places more emphasis on recent periods. It reacts faster when conditions are changing.

Seasonality index

A seasonality index expresses how a period typically compares with an average period. Indices are usually normalised so that the average across a full cycle is approximately 1.00. For example, an index of 1.20 suggests that period is typically 20% above the average level.

Core theory and frameworks

Choosing the right forecasting approach

The best technique depends on:

• Purpose (short-term liquidity control vs longer-term planning)
• Time step (weekly, monthly, quarterly)
• Availability of reliable drivers
• Stability of the environment (stable vs rapidly changing)

A practical approach is to start simple and then add complexity only where it improves decision usefulness.

Common forecasting techniques

1) Naïve forecast (last period repeated)
Useful as a quick baseline for stable series. Avoid when trend or seasonality is strong.

2) Simple average
Useful when results are steady with no clear trend and limited seasonality.

3) Moving average
Useful to smooth volatility and reduce the impact of one-off movements.

4) Weighted moving average
Useful when recent history is more relevant than older data.

5) Trend projection
Extends a consistent growth/decline pattern. Works best when the reason for the trend is understood and likely to continue.

6) Seasonality adjustment
Applies seasonal indices to reflect expected peaks and troughs.

7) Driver-based forecasting (bottom-up)
Builds receipts and payments from operational plans, then applies timing assumptions. This is often the most controllable approach because it is anchored to activity (orders, staffing, production).

Translating sales into cash timing

Sales activity and cash receipts are not the same:

• On a credit sale, cash is received later, based on credit terms and customer settlement behaviour.
• Revenue is recognised when the underlying goods or services are transferred as promised in the contract. This is often earlier than cash settlement when credit terms are offered, but it can also coincide with or follow cash, depending on the contract and payment terms.

To model receipts, a common starting point is a collection pattern, for example:

• 70% collected in month 1 after sale
• 30% collected in month 2 after sale

This pattern is applied to forecast sales to produce forecast cash receipts.

Simplified accounting mechanics (for understanding timing only; ignoring tax, discounts, returns and other complexities):

• At credit sale: Dr Trade receivables, Cr Revenue
• At cash collection: Dr Cash, Cr Trade receivables

Mirror note: the same logic applies to cash payments

Receipts are only half the picture. Payments must be forecast using the same timing discipline: supplier credit terms (and any early-payment discounts), payroll payment dates, indirect taxes (such as VAT) timing, loan interest schedules, and planned capital expenditure.

Sensitivity and scenario analysis

Sensitivity analysis

Sensitivity analysis changes one assumption at a time while holding others constant. It answers: “If this single input is wrong, how much does cash change?”

Examples:

• Sales volume down 5%
• Average selling price down 2%
• Collections slower by 10 percentage points (for example 60/40 instead of 70/30)

Scenario analysis

Scenario analysis changes a consistent package of assumptions to create coherent outcomes (for example base / downside / severe downside). It answers: “If conditions change in a realistic way, what happens to cash?”

A scenario typically adjusts:

• Sales level and/or growth rate
• Collection pattern and credit losses
• Supplier payment terms
• Inventory policy and operating costs

Break-even (cash) point

The cash break-even point is the activity level where net cash movement for a period is zero. It is a liquidity concept, not a profit concept.

In practice, cash break-even depends on cash fixed outflows and the cash contribution generated per unit (or per £ of sales). Timing matters: working-capital movements can mean the period’s cash break-even differs from a “steady-state” break-even, especially where inventory builds or customers pay slowly.

Forecast error and variance review

Forecast error is the difference between actual and forecast outcomes.

Forecast error = Actual – Forecast

Error can be measured in money terms and as a percentage. One common approach is:

Absolute error = |Actual – Forecast|

Percentage error = (Actual – Forecast) / Forecast × 100%

A positive error means actual is higher than forecast (under-forecast). A negative error means actual is lower than forecast (over-forecast). Other conventions are also used in practice (for example dividing by actual, or using mean absolute percentage error for a set of forecasts).

Errors should be analysed by type:

• Driver error (sales volume, price, payroll, etc.)
• Timing error (collections or payments delayed)
• One-off items (non-recurring receipts/payments)

A variance review cycle improves future forecasts:

1. Compare forecast to actual by line item and timing.
2. Explain the main variances (driver vs timing vs one-off).
3. Update assumptions and, where needed, improve the driver model.
4. Record changes so future forecasts remain consistent and auditable.

Worked example

Narrative scenario

XYZ Ltd is preparing a monthly cash receipts forecast for April to June to assess whether short-term funding may be needed if customer payment behaviour worsens.

Sales are made on credit. Historically, cash collections follow this pattern:

• 70% collected in the month after sale
• 30% collected in the second month after sale

A significant customer has announced it will extend its payment terms from 30 days to 60 days. For forecasting, assume:

• The customer represents 30% of monthly sales from April onwards.
• For that customer only, each receipt occurs one month later than under the standard pattern (month +1 → month +2; month +2 → month +3).
• Other customers continue to pay under the historic pattern.

Historical sales for the first quarter:

• January: £200,000
• February: £220,000
• March: £242,000

Management expects sales to grow by 10% per month from April onward.

Task (what you are building and why)

Build a three-month receipts forecast to show (i) the base outcome, (ii) the effect of the large customer delaying payment, and (iii) how sensitive cash receipts are to weaker trading and slower collections.

Solution

Step 1: Forecast sales (10% monthly growth)

April sales
£242,000 × 1.10 = £266,200

May sales
£266,200 × 1.10 = £292,820

June sales
£292,820 × 1.10 = £322,102

Step 2: Cash receipts (base case: historic 70% / 30%)

April receipts
70% of March + 30% of February
= (0.70 × £242,000) + (0.30 × £220,000)
= £169,400 + £66,000
= £235,400

May receipts
70% of April + 30% of March
= (0.70 × £266,200) + (0.30 × £242,000)
= £186,340 + £72,600
= £258,940

June receipts
70% of May + 30% of April
= (0.70 × £292,820) + (0.30 × £266,200)
= £204,974 + £79,860
= £284,834

Step 3: Impact of the customer payment-term change (30% of sales delayed by one month)

We split receipts into:

• Other customers (70% of sales): unchanged 70% in month +1 and 30% in month +2.
• Large customer (30% of sales): each receipt occurs one month later than under the standard pattern (month +1 → month +2; month +2 → month +3).

Under the normal pattern, receipts in May include:

• 70% of April sales (month +1), and
• 30% of March sales (month +2).

For the large customer only, both of these cash elements shift one month later.

Impact on May receipts

Reduction in May
= (0.70 × 0.30 × April sales) + (0.30 × 0.30 × March sales)
= (0.70 × 0.30 × £266,200) + (0.30 × 0.30 × £242,000)
= £55,902 + £21,780
= £77,682

Adjusted May receipts
£258,940 − £77,682 = £181,258

Impact on June receipts

June base receipts include:

• 70% of May sales (month +1), and
• 30% of April sales (month +2).

With the delay, three cash-flow effects occur for June:

1. Large customer’s 70% of May shifts out of June into July (reduces June).
2. Large customer’s 70% of April shifts into June (it slipped from May to June).
3. Large customer’s 30% of March shifts into June (it slipped from May to June).

Net June adjustment
= –(0.70 × 0.30 × May sales) + (0.70 × 0.30 × April sales) + (0.30 × 0.30 × March sales)
= –(0.70 × 0.30 × £292,820) + (0.70 × 0.30 × £266,200) + (0.30 × 0.30 × £242,000)
= –£61,492.20 + £55,902.00 + £21,780.00
= +£16,189.80

Adjusted June receipts
£284,834 + £16,189.80 = £301,023.80
= £301,024 (rounded)

Interpretation
May takes the largest hit because cash that would normally arrive in May from both April (month +1) and March (month +2) is pushed into later months. June can rise above the base case because it receives delayed cash from earlier sales, even though some May cash is pushed beyond the forecast horizon into July. The timing delay increases forecast receivables at May and June month-ends, reducing cash headroom.

Step 4: Sensitivity analysis (sales 5% lower; collections remain 70% / 30%)

Revised sales (5% lower than forecast)

April sales
£266,200 × 0.95 = £252,890

May sales
£292,820 × 0.95 = £278,179

June sales
£322,102 × 0.95 = £305,996.90
= £305,997 (rounded)

Revised cash receipts (historic pattern unchanged)

April receipts
Unchanged (driven by February and March actual sales): £235,400

May receipts
70% of revised April + 30% of March
= (0.70 × £252,890) + (0.30 × £242,000)
= £177,023 + £72,600
= £249,623

June receipts
70% of revised May + 30% of revised April
= (0.70 × £278,179) + (0.30 × £252,890)
= £194,725.30 + £75,867
= £270,592.30
= £270,592 (rounded)

Step 5: Downside scenario (sales 8% lower; collections 65% / 35%)

Apply the 65% / 35% collection split to cash collected from April onwards (i.e., April receipts use 65% on March and 35% on February).

Sales reduced by 8% from the base forecast

April sales
£266,200 × 0.92 = £244,904

May sales
£292,820 × 0.92 = £269,394.40
= £269,394 (rounded)

June sales
£322,102 × 0.92 = £296,333.84
= £296,334 (rounded)

Collections worsen to 65% in month +1 and 35% in month +2.

April receipts
65% of March + 35% of February
= (0.65 × £242,000) + (0.35 × £220,000)
= £157,300 + £77,000
= £234,300

May receipts
65% of April + 35% of March
= (0.65 × £244,904) + (0.35 × £242,000)
= £159,187.60 + £84,700
= £243,887.60
= £243,888 (rounded)

June receipts
65% of May + 35% of April
= (0.65 × £269,394.40) + (0.35 × £244,904)
= £175,106.36 + £85,716.40
= £260,822.76
= £260,823 (rounded)

Common pitfalls and misunderstandings

• Treating sales as cash: Credit sales create receivables first; cash arrives later.
• Ignoring timing shifts: A one-month delay for a large customer can create a major funding gap.
• Using one technique everywhere: Stable series may suit averages; changing markets may need weighted or driver-based methods.
• Forgetting seasonality: Forecasts that ignore seasonal peaks and troughs often fail at month level.
• No audit trail of assumptions: Forecasts are hard to improve if assumptions are not recorded and reviewed.
• Weak stress testing: Failing to test key drivers (sales, collections, payroll, supplier terms) can create false confidence.
• Not separating one-offs: Non-recurring items should be identified so they do not distort trend-based forecasts.

Summary

Forecasting is about method, assumptions and timing. Cash forecasting requires converting operational activity into receipts and payments based on realistic patterns, not simply projecting profit.

Trend-based and average-based techniques can provide a starting point, but the strongest forecasts use drivers and cash timing that reflect how the business operates. Sensitivity analysis highlights the variables that matter most. Scenario analysis shows how combined pressures can affect liquidity.

Regular variance reviews—comparing forecast to actual and updating assumptions—create a practical improvement cycle and steadily increase forecast reliability.

FAQ

What is the difference between sensitivity analysis and scenario analysis?

Sensitivity analysis changes one input at a time to show exposure to that single assumption. Scenario analysis changes a linked set of assumptions together (for example lower sales plus slower collections) to model coherent outcomes.

How does seasonality affect cash forecasting?

Seasonality changes the timing and scale of receipts and payments within the year. Forecasts should reflect predictable peaks and troughs and the likely lag between sales peaks and cash collections.

Why should forecast errors be analysed rather than ignored?

Forecasts improve when errors are explained and categorised (driver, timing, one-off). Without a variance review, the same mistakes repeat and the forecast remains a poor decision tool.

What are the benefits of driver-based forecasting?

Driver-based forecasting ties cash movement to operational plans (orders, staffing, production). It improves transparency, supports accountability and allows quicker updates when business activity changes.

How can a business improve its cash forecasting process?

Use clear drivers, apply realistic timing assumptions, stress-test key variables and run a routine variance review that updates assumptions based on actual outcomes.

Summary (Recap)

This chapter covered practical methods for building cash forecasts using trends, averages and seasonal adjustments, and showed how operational drivers are converted into cash timing. It introduced sensitivity analysis and scenario analysis as tools for stress testing, and explained how forecast errors can be analysed and used to improve future forecasts through a simple review cycle. A worked example illustrated how a customer’s delayed payment can materially shift receipts even when sales growth remains strong.

Glossary

Forecast driver

A measurable input that explains cash movement (for example units sold, pay rates, headcount, supplier terms, collection pattern).

Time series

Data recorded at regular intervals (such as weekly or monthly) used to identify patterns over time.

Trend

The underlying direction of a series over time after ignoring short-term fluctuations.

Seasonality

A repeating within-year pattern that causes predictable peaks and troughs.

Moving average

An average of the most recent periods used to smooth volatility and highlight underlying movement.

Weighted moving average

A moving average that places greater emphasis on more recent periods to reflect changing conditions.

Seasonality index

A factor used to adjust forecasts for seasonal effects, typically normalised so the average over a full cycle is approximately 1.00.

Sensitivity analysis

A stress test that changes one assumption at a time to measure the forecast’s exposure to that variable.

Scenario analysis

A stress test that changes multiple linked assumptions together to model coherent outcomes.

Break-even (cash) point

The activity level where net cash movement for a period is zero.

Forecast error

The difference between actual results and the forecast, measured in money terms or as a percentage.

Variance review

A routine process that explains forecast vs actual differences and updates assumptions to improve future forecasts.