Every budget starts with a guess about the future: how many units will we sell, what will materials cost, when will customers pay? A forecast turns that guess into something structured, based on evidence and stated assumptions. It is not a promise, and it is not a target. It is the best available estimate, produced in a way that others can check.
This topic covers what forecasting is, the types of forecast, why forecasting matters, the qualities of a good forecast, the steps in the forecasting process, how forecasting differs from budgeting, its limitations, and the main qualitative and quantitative methods. The next two topics then work through the calculations of time series, moving averages, trend lines and seasonal variations.
In 30 seconds
Forecasting uses historical data, current trends and analytical models to estimate future results. A forecast is the output. Forecasts may be short term (up to a year), medium term (one to three years) or long term (over three years); by function (sales, expenditure, cash flow, profit, economic); and by method (qualitative: judgement, market research, Delphi; quantitative: time series, moving averages, exponential smoothing, ARIMA, causal models, simulation).
A good forecast is accurate, reliable, timely, relevant, simple enough and cost-effective. The process runs from setting the objective to monitoring and revising. A budget is a plan and a target. A forecast is a prediction that feeds the budget.
What forecasting is
Forecasting is the process of using historical data, current trends and analytical models to estimate future outcomes. In management accounting it is applied to sales, revenue, costs, cash flow and demand, in order to guide decisions, budgeting and strategic planning. It is not guesswork: it rests on logical assumptions, statistical tools and models.
A forecast is the actual result of the process, a quantitative or qualitative prediction of a future event or measure. A sales forecast might say that the company will sell 10,000 units next quarter, and a cash flow forecast might estimate the inflows and outflows of next month. Forecasts can be numerical (revenue, units, expenses) or descriptive (market trends, the economic outlook).
Types of forecast
| Basis | Type | Description |
|---|---|---|
| Time span | Short-term | A few weeks to a year. Used for operations, inventory control, staffing and cash flow |
| Medium-term | One to three years. Supports budgeting, pricing and resource planning | |
| Long-term | Beyond three years. Used for strategy, investment and capital budgeting | |
| Function | Sales forecast | Predicts future sales volume or revenue. Drives production and marketing plans |
| Expenditure forecast | Estimates fixed and variable spending | |
| Cash flow forecast | Predicts the timing and amount of cash inflows and outflows | |
| Profit forecast | Combines revenue and cost forecasts to project profit | |
| Economic forecast | Looks at inflation, interest rates, exchange rates and other variables that affect the business | |
| Method | Qualitative | Based on judgement, intuition or market research. Common when historical data is limited |
| Quantitative | Uses historical data and statistical methods. More objective and data-driven |
Why forecasting matters
- Strategic planning. It helps the organisation see trends coming, prepare for market change and set realistic long-term goals.
- Budgeting and financial planning. Forecasts are the basis of budgets. Accurate ones keep budgets realistic and prevent over- or under-estimating revenues and costs.
- Resource allocation. It shows where labour, materials, capital and equipment will be needed.
- Risk management. It highlights risks and uncertainties so that contingency plans can be made.
- Performance evaluation. Forecasts are benchmarks against which actual results are measured, so variances can be explained.
- Investment and financing decisions. Forecast data shows whether a new project, purchase or funding is feasible and profitable.
- Communication and coordination. A shared forecast means all departments plan on the same expectations.
Characteristics of a good forecast
| Quality | Meaning |
|---|---|
| Accuracy | It matches actual outcomes closely |
| Reliability | It gives consistent results under similar conditions |
| Timeliness | It is made early enough to influence decisions |
| Relevance | It relates directly to the decision being made |
| Simplicity | The model is as simple as possible while still capturing the essential patterns |
| Cost-effectiveness | The benefits outweigh the cost of preparing it |
The forecasting process
The process is a structured sequence that combines quantitative and qualitative elements, so that decisions rest on data and also on expert judgement.
- Identify the objective. Decide what is to be forecast (sales, costs, demand, cash) and why. The objective guides every later step.
- Collect relevant data. Gather historical records, financial data, operating measures, market research and economic indicators. The forecast is only as good as the data, which must be complete, accurate and relevant.
- Analyse and prepare the data. Clean it, identify trends, remove outliers and one-off events, and adjust for anomalies. Organise it by period, category or department.
- Select a method. The choice depends on the type of forecast, the data available and the accuracy needed. It may be qualitative or quantitative.
- Apply the technique. Run the data through the model, such as a moving average, regression or time series model, to produce the forecast figures.
- Interpret and evaluate. Ask whether the results make sense, what assumptions were made and whether they fit current conditions. Assess risks and limits, and do not accept the numbers at face value.
- Monitor and revise. Compare the forecast with actual results. If they differ, update the model or the assumptions, so the forecast remains a living tool.
Forecasting and budgeting
Forecasting is the foundation of budgeting. Predicting sales, production, costs and cash flows gives the estimates on which budgets are built. A sales forecast, for example, determines production levels, resource needs and expected revenue. In budgeting, forecasts serve as:
- basis for planning: they supply the data for budget estimates;
- risk reduction: anticipating conditions reduces the chance of large variances;
- resource allocation: they show where resources are needed most;
- performance benchmarks: they set expectations against which actual performance is judged in budgetary control; and
- flexibility: accurate forecasts let the organisation revise budgets quickly when conditions change.
| Aspect | Budgeting | Forecasting |
|---|---|---|
| Definition | A detailed financial plan for a specific future period | An estimate or prediction of future financial outcomes |
| Purpose | Sets financial targets and guides operations | Predicts likely future outcomes from current trends |
| Time frame | Usually a fixed period, such as a year | Short, medium or long term |
| Flexibility | Generally rigid once approved | More flexible and updated regularly |
| Formality | Part of a formal planning process | May be informal and used mainly internally for decisions |
| Focus | Financial control and resource allocation | Predicting trends and informing decisions |
| Use in control | A standard for performance evaluation (variance analysis) | Helps anticipate deviations and adjust plans |
| Relationship | Depends on forecasts to set targets | Supplies the input and basis for budgets |
Limitations of forecasting
| Limitation | Explanation |
|---|---|
| Reliance on historical data | Forecasts assume past trends continue, which may not happen |
| Uncertainty and external shocks | Recessions, policy changes and pandemics can make a forecast wrong |
| Model assumptions | Models simplify, and may not reflect real-world complexity |
| Bias and subjectivity | Qualitative forecasts especially can reflect personal bias, over-optimism or pressure from managers |
| Inflexibility to rapid change | Long-term forecasts adapt slowly to sudden shifts |
| Cost and resources | Some methods need much time, data, skill and technology |
| Over-reliance | Treating a forecast as a certainty ignores its risks |
| Data and technology limits | Poor or incomplete data, and over-dependence on tools, weaken results |
Methods of forecasting
Qualitative methods
These are used mainly when historical data is scarce, for example for a new product or a changed market.
- Delphi method. A panel of experts, inside or outside the firm, is consulted, usually in rounds, to build a forecast from their collective expertise. A firm launching a new product line might ask marketing specialists, analysts and industry experts to estimate revenue and cost.
- Market research. Surveys, focus groups and customer feedback to estimate demand, preferences and sales.
- Executive judgement. Senior managers’ forecasts based on their experience and understanding of the market.
Quantitative methods
These use historical data and statistics for more objective forecasts, when reliable data exists.
| Method | What it does | Best for |
|---|---|---|
| Time series analysis | Examines past data for recurring patterns and trends and extends them forward | Sales, production costs, cash flows |
| Moving averages | Smooths fluctuations to show the underlying trend | Short-term forecasts, such as monthly sales |
| Exponential smoothing | Gives more weight to recent data | Short-term forecasts |
| ARIMA models | Statistical models for data with complex patterns, including trend, seasonality and irregular movements | Longer-term forecasts of revenue, expenses, cash flow |
| Causal methods (e.g. regression) | Link the forecast variable to outside factors such as advertising spend, production level or gross domestic product (GDP) | Longer-term forecasts. Understanding relationships |
| Simulation (e.g. Monte Carlo) | Uses probabilities to produce a range of possible outcomes, not a single figure | Risk management, investment under uncertainty |
For short-term forecasts, moving averages and exponential smoothing are common. For long-term forecasts, causal methods and ARIMA are more suitable, because they can build in wider economic factors and long-term trends.
The ideas in more depth
Exponential smoothing: a short illustration
Background: no formula for exponential smoothing is given in the main notes. The usual form is: next forecast = α × latest actual + (1 − α) × latest forecast, where α (between 0 and 1) is the weight given to the latest actual. Suppose the forecast for month 1 was 100 units, α = 0.3, and actual sales were 110. The forecast for month 2 = 0.3 × 110 + 0.7 × 100 = 33 + 70 = 103. If month 2 actual was 120, the forecast for month 3 = 0.3 × 120 + 0.7 × 103 = 36 + 72.1 = 108.1. A larger α reacts faster to recent change.
A causal forecast in numbers
A retailer finds, using regression on past data, that monthly sales (₦m) = 40 + 2.5 × advertising spend (₦m). If the planned spend next month is ₦8m, forecast sales = 40 + 2.5 × 8 = ₦60m. This is a causal forecast. It is only as good as the strength of the relationship (the correlation coefficient) and the assumption that it continues.
Choosing a method
| Situation | Suitable method |
|---|---|
| New product with no sales history | Qualitative: Delphi, market research, executive judgement |
| Monthly sales with a stable pattern and enough history | Time series: moving averages with seasonal adjustments |
| Sales that depend on a known driver such as advertising | Causal: regression |
| A decision with several uncertain variables | Simulation, which gives a range of outcomes |
Forecast error and review
Forecasts are always wrong by some amount. What matters is whether the error is small, unbiased and understood. If actual sales are consistently below forecast, the model or the assumptions are biased and should be revised. Comparing forecast with actual each period is the discipline that keeps forecasting honest.
Forecast versus plan
A forecast says what is likely to happen if present trends continue. A plan or budget says what management intends to happen. Management may set a budget above or below the forecast, for example to stretch staff, but it must then explain the gap and the actions that will close it.
Worked example: matching method to situation
Scenario. Aso Textiles Ltd needs (a) a sales forecast for a new range of fabrics it has never sold, (b) a forecast of next month’s sales of its main product, which has five years of monthly data with a clear seasonal pattern, and (c) a forecast of how sales will respond if advertising is doubled. Suggest a method for each and a limitation.
- (a) New range: a qualitative method, such as Delphi with marketing and industry experts, supported by market research. Limitation: it is subjective and may be biased by over-optimism.
- (b) Main product: time series analysis using moving averages and seasonal variations. Limitation: it assumes that past patterns continue, and cannot allow for sudden changes such as a new competitor.
- (c) Advertising response: a causal forecast using regression of sales on advertising spend. Limitation: doubling spend lies outside the range of past data, so the relationship may not hold, and correlation does not prove cause.
Simple extrapolation: a growth rate forecast
The simplest quantitative forecast projects past growth forward. Sales over four years were ₦80m, ₦88m, ₦99m and ₦110m. Year-on-year growth was 10.0%, 12.5% and 11.1%. The compound average growth rate is (110 ÷ 80)^(1/3) − 1 = 1.375^0.3333 − 1 = 11.2%. A forecast for the next year is ₦110m × 1.112 = ₦122.3m. This method is quick and gives a sensible starting point, but it ignores seasonality, turning points and any change in conditions. It works best for short-term forecasts in a steady environment.
Scenario and probability forecasts
A single-figure forecast hides uncertainty. A manager may prepare three: an optimistic, a most likely and a pessimistic scenario, each with an assessed probability. Suppose sales next year could be ₦140m (probability 0.2), ₦120m (probability 0.5) or ₦90m (probability 0.3). The expected value is 0.2 × 140 + 0.5 × 120 + 0.3 × 90 = 28 + 60 + 27 = ₦115m. The range (₦90m to ₦140m) shows the risk.
Budgets can then be prepared on the most likely figure, with plans ready for the pessimistic case, such as cost cuts that can be triggered quickly.
Measuring forecast accuracy
The characteristics of a good forecast include accuracy, so it must be measured. Four months of actual and forecast sales (₦m) are compared.
| Month | Actual | Forecast | Error (actual − forecast) | Absolute error | Absolute error as % of actual |
|---|---|---|---|---|---|
| 1 | 100 | 105 | −5 | 5 | 5.00% |
| 2 | 110 | 108 | +2 | 2 | 1.82% |
| 3 | 120 | 125 | −5 | 5 | 4.17% |
| 4 | 115 | 110 | +5 | 5 | 4.35% |
| Average | −0.75 | 4.25 | 3.83% |
- The mean absolute deviation (MAD) is 4.25: on average the forecast is out by ₦4.25m, whichever direction.
- The mean absolute percentage error (MAPE) is 3.83%, a measure that can be compared across products of different sizes.
- The mean error (bias) is −0.75, a small tendency to over-forecast. Averaging the signed errors shows whether the forecast is consistently too high or too low. A large persistent bias means the method or the assumptions need revising.
Comparing these measures for two methods on the same history shows which has been more accurate. The method with the lower MAD or MAPE has fitted the past better, although this is no guarantee for the future.
Combining statistics with judgement
Quantitative forecasts show what will happen if the past pattern continues. Managers then adjust for what they know about the future. A sales forecast based on the past three years might be raised by 5% for a new branch opening, or lowered because a major customer is moving to a competitor.
Good practice is to record these adjustments and the reasons, so that they can be reviewed later when actual figures arrive.
A common Delphi procedure illustrates how judgement can be structured:
- experts give independent estimates (say ₦100m, ₦130m, ₦150m, ₦120m)
- the coordinator summarises the range and median (₦125m) and sends it back with the reasons given
- in a second round the experts revise (to ₦115m, ₦128m, ₦135m, ₦122m)
- the range has narrowed, and the process stops when the estimates converge
Rolling forecasts
A rolling forecast always looks a fixed number of periods ahead (for example twelve months) and is updated each month: the month just ended is replaced with actual figures and a new month is added at the end. It keeps the plan up to date in a changing environment and avoids the rigid, once-a-year budget.
The cost is the effort of regular revision, so firms often limit the amount of detail.
Forecasting in an unstable economy
In an economy with high inflation, a volatile exchange rate and changes in fuel and power costs, forecasts of cost and cash flow need extra care. Practical steps include forecasting in real and nominal terms, using shorter horizons, updating forecasts monthly, preparing exchange-rate scenarios for imported inputs, and tracking the key external indicators that affect the business (inflation, interest rates, the naira rate, commodity prices).
A forecast that was accurate at the start of the quarter may be out of date by its end, which makes regular review essential.
Common mistakes in the exam
- Treating a forecast and a budget as the same thing. A budget is a plan and a target. A forecast is a prediction.
- Naming methods without saying when each is suitable.
- Listing the steps of the process without explaining them.
- Saying quantitative forecasts are always better. They need reliable data and a stable environment.
- Forgetting the limitations, especially reliance on past data.
- Mixing up qualitative methods (Delphi, market research, judgement) with quantitative ones.
Quick revision
- Forecast types: short, medium, long term; sales, expenditure, cash, profit, economic; qualitative or quantitative.
- Good forecast: accurate, reliable, timely, relevant, simple, cost-effective.
- Process: objective, data, analyse data, choose method, apply, interpret, monitor and revise.
- Budget = plan and control standard. Forecast = prediction that feeds the budget.
- Methods: Delphi, market research, executive judgement (qualitative); time series, moving average, exponential smoothing, ARIMA, causal, simulation (quantitative).
Practice questions
These are original questions written in the style of the Institute of Chartered Accountants of Nigeria (ICAN) exam. Try each one before you open the answer.
Section A style: multiple choice
Question 1. A forecast is:
- A plan that management must achieve
- A prediction of a future event or measure
- A record of past results
- A variance
Show the answer
B. It is the output of the forecasting process.
Question 2. A long-term forecast typically extends:
- Up to three months
- Beyond three years
- Exactly one year
- One week
Show the answer
B. Used for strategy and capital budgeting.
Question 3. The Delphi method is a:
- Quantitative method using regression
- Qualitative method consulting a panel of experts
- Simulation model
- Moving average
Show the answer
B. It draws on collective expert judgement.
Question 4. Which is a characteristic of a good forecast?
- Complexity at all costs
- Cost-effectiveness
- Delay
- Irrelevance to the decision
Show the answer
B. Benefits should outweigh costs.
Question 5. The first step of the forecasting process is to:
- Apply the technique
- Identify the objective of the forecast
- Monitor and revise
- Interpret results
Show the answer
B. The objective guides every other step.
Question 6. Compared with a budget, a forecast is generally:
- More rigid
- More flexible and updated regularly
- A target that must be achieved
- Only for external use
Show the answer
B. Budgets are rigid once approved.
Question 7. A main limitation of forecasting is:
- It assumes that past trends will continue
- It uses data
- It is always correct
- It is always cheap
Show the answer
A. Historical reliance is a key weakness.
Question 8. Linking sales to advertising spend by regression is an example of a:
- Causal method
- Qualitative method
- Executive judgement
- Delphi method
Show the answer
A. Sales depend on an outside factor.
Question 9. Which method gives a range of possible outcomes using probabilities?
- Moving averages
- Monte Carlo simulation
- Executive judgement
- Market research
Show the answer
B. Simulation models uncertainty.
Question 10. Sales = 40 + 2.5 × advertising. Planned advertising is ₦8m. Forecast sales (₦m) are:
- 48
- 60
- 80
- 20
Show the answer
B. 40 + 2.5 × 8 = 60.
Section B style: written questions
Question 11 (10 marks). Explain the forecasting process and the qualities of a good forecast.
Question 12 (20 marks). Aso Textiles Ltd uses forecasts in its planning.
(a) Distinguish between a forecast and a budget. (6 marks)
(b) Explain the role of forecasting in budgeting. (4 marks)
(c) Describe three qualitative and three quantitative forecasting methods. (6 marks)
(d) State four limitations of forecasting. (4 marks)
A further written question
Question 13 (10 marks). (a) The forecast for month 1 was 200 units. Actual sales were 220. Using exponential smoothing with α = 0.4, calculate the forecast for month 2. If month 2 actual sales were 210, forecast month 3. (4 marks) (b) A regression gives sales (₦m) = 30 + 4 × advertising (₦m). Forecast sales for advertising of ₦5m. (2 marks) (c) State four types of forecast by function. (4 marks)
Model answers
Show the model answer to Question 11
The process. (1) Identify the objective: decide what is to be forecast and why. (2) Collect relevant data: historical, financial, operating, market and economic. (3) Analyse and prepare it: clean it, find trends, remove outliers. (4) Select a method, qualitative or quantitative. (5) Apply the technique to produce the figures. (6) Interpret and evaluate the results, checking assumptions, risks and limitations. (7) Monitor and revise the forecast against actual results.
Qualities. Accuracy (close to actual outcomes); reliability (consistent results in similar conditions); timeliness (early enough to affect decisions); relevance (linked to the decision); simplicity (as simple as possible while capturing the patterns); and cost-effectiveness (benefits exceed costs).
Show the model answer to Question 12
(a) A budget is a detailed financial plan for a specific future period. It sets targets, guides operations, is generally fixed once approved, and is used as a standard for control through variance analysis. A forecast is an estimate of likely future outcomes based on current trends. It can cover any time span, is flexible and revised regularly, and may be informal.
A budget is built from forecasts, and a forecast can warn of deviations from the budget.
(b) Forecasts are the basis for planning, because a sales forecast determines production, resource needs and expected revenue. They reduce risk by anticipating conditions, help allocate resources, set the benchmarks against which actual performance is judged, and allow budgets to be revised quickly when conditions change.
(c) Qualitative: Delphi method (a panel of experts consulted in rounds); market research (surveys and customer feedback); executive judgement (managers’ experience). Quantitative: time series analysis with moving averages (smooths fluctuations to show the trend); causal regression (links sales to outside factors); simulation or ARIMA (for uncertainty or complex patterns).
(d) Limitations (any four):
- reliance on historical data
- uncertainty and external shocks
- model assumptions that simplify reality
- bias and subjectivity
- inflexibility to rapid change
- cost and resources
- over-reliance on forecasts
- poor data and technology limits
Show the model answer to Question 13
(a)
- Month 2 forecast = 0.4 × 220 + 0.6 × 200 = 88 + 120 = 208.
- Month 3 forecast = 0.4 × 210 + 0.6 × 208 = 84 + 124.8 = 208.8.
(b) Sales = 30 + 4 × 5 = ₦50m.
(c) Sales forecast, expenditure forecast, cash flow forecast, profit forecast and economic forecast (any four).
Pathfinder on Forecasting types, process and methods
These are the ICAN Pathfinder questions that test this topic. ICAN owns the questions and the suggested solutions, so we point you to them and do not copy them. Open your Pathfinder, attempt each question under exam conditions, then check your answer against the notes above.
From the previous syllabus. Asks for the difference between a budget and a forecast: a budget is a plan of where a business wants to go, while a forecast is a prediction.
Revise first: Forecasting and budgeting
From the previous syllabus. An item mentions co-ordination of business forecasts among the functions of budgeting, linking forecasting to budgets.
Revise first: Why forecasting matters
Multiple-choice items can test types of forecast, qualitative and quantitative methods, and the difference between a forecast and a budget. Check Section A of every diet you practise.
Revise first: Methods of forecasting
Source: ICAN Pathfinder, Management Accounting. Diets covered: May Diet, 2018 and May Diet, 2016 (previous syllabus). May Diet, 2025 and November Diet, 2024 are not covered here.
Spotted a mistake? Tell us and we will check it.
