Business Forecasting And Modelling – Keys To Unlocking Business Growth

Business Forecasting and Modelling - Keys To Unlocking Business Growth.



Leading Tax Advice

Call Nicholas Kilpatrick




At all stages of the company’s existence, short– and long-range plans should be in place that dictate its actions.  Company management and external stakeholders, such as investors, banks, and creditors want to know that growth prospects for the company have been forecasted to the most reasonable extent possible.  The strength and validity of these forecasts will determine to a large extent the capital available to the company to continue and expand operations.

In addition, the company needs to understand all the variables both within and outside its control that will influence cash inflows and outflows.  Extensive study on economic and consumer environments is important to forecast sales levels, which can then be used to determine other future financial components, such as costs of goods sold,  capital requirements, and, most importantly, net income.

Therefore, one of the most important strategic exercises that the company should complete on a regular basis is the forecasting of sales.  To do this, a complex  range of economic factors needs to be analyzed, including:

-Existing competition and potential competitive threats


-Sources of revenue

Preparing usable forecasts that the company can base decisions on means building an understanding of the many different strands that influence the relative strengths and weaknesses of the economy in which the company operates.

When undertaking a forecast of business components, due diligence takes on primary importance in order to ensure that, when forecasting, all possible variables affecting the forecast of sales an ancillary components affecting those sales have been considered.  Necessary to exhaust diligence efforts include a study of :


  1. Company analysis
  2. Industry analysis
  3. Domestic economic analysis
  4. International economic analysis


Doing a Company analysis involves assessing the strength of the company by calculating and assessing the results of financial ratios.  If the ratios expose any problems within operations, they need to be corrected to make sure that the result of the forecast can to be extent possible represent what will happen in reality.  If, for example, there are ongoing cases of fraud or wastage, then a forecast of increased sales will not result in the anticipated forecast of cash inflows; the theft of cash will not allow it.  All anomalies arising from the study of financial ratios, therefore, need to be addressed and corrections made.

Industry analysis involves a SWOT (Strengths, Weaknesses, Opportunities, Threats) to show attributes that can progress and hinder the Company’s growth.  An understanding of these areas will enable the company’s leader to allocate resources so that growth and sustainability are maximized.  A SWOT analysis combined with a constructive discussion on future economic and business trends which the company will be influenced by will help to position the company for future growth, because time has been spent determining what will happen in the future and how the company should be positioned and built to capitalize on those future friends.

An understanding of local and, if applicable, international economic trends is necessary to optimize the accuracy of the forecast.  This is done by influencing the eventual forecast equation with realistic economic variables that can give a practical picture of what the environment in which the business operates is like.

Once these considerations have been assessed and exhaustive research undertaken to create a reasonable and accurate landscape upon which the forecast will be built, we proceed on to the exercise of actually building the forecasting model.  Building a forecasting model usually involves the following steps:


  1. Data pattern recognition
  2. Analysis and recognition of model influencers
  3. Assess and confirm Forecasting method
  4. Build forecast model
  5. Apply model to historical data and test accuracy
  6. Refine model for accuracy
  7. Apply the forecasting model to results obtained through analysis of 3-and 5 years projections.
  8. Refine through iterations as necessary


Data pattern recognition involves compiling historical data which can be used to refine economic and industry forecasts.  A simple time series plot can be employed here.  When data is placed on a plot diagram, we can see the behaviour of the data points (such as sales, and variable costs), and be able to co-relate economic conditions, market influences, etc.. to those data points to ascertain what influences the data and what constitutes drivers for company sales.  It’s essential to be able to identify drivers for sales because they constitute influencers of that which we’re trying to forecast – if we can identify and hopefully influence the driver, we can influence the component we’re forecasting, which in this case is sales.



Act Now To Get Our Value Pricing Adjustment on Audit and Review Engagements For Your Business


Drivers which cannot be influences are dealt with by determining how we can minimize the exposure of the forecasting component (sales) to that driver.  Management may go to the extent of assessing alternative product offerings and/ or production processes which mitigate exposure to that driver.

Once the data have revealed the behaviours of sales, we can look for model influencers  in the data.  Model influencers are items like auto-correlation, seasonality or cycles, and must be identified and confirmed as influencers so that the eventual forecasting model can be refined to take into consideration these effects.  Also of benefit to do in order to identify influencers is the decomposition of historical data.  Decomposition of time series data is based on the concept that underlying patterns in data can be distinguished from randomness.  Decomposition essentially breaks down data into its’ component parts and can be beneficial in building a model which can accurately forecast business components.

Based on the data and influencers identified, a forecasting method can be chosen which provides the right strength and flexibility to contain all identified variables influencing the forecasted component (ie: component drivers).  The most popular methods, listed from simplest to complex, are as follows:


  1. Time Series
  2. Single, Multiple and Stepwise Regression
  3. Exponential Smoothing
  4. Additive and Multiplicative Decomposition (ARMA and ARIMA)


The type of forecasting model used depends on the complexity of the business and the type of industry it operates within.  Regression methods are best suited for business that contain multiple drivers influencing the forecast variable, such as price elasticity, cost of equipment operation, or staffing costs to name a few.  Exponential Smoothing methods allow for smoothing techniques to find commonalities in historical data with wide variations, and also provide consideration for influencers such as seasonality and cyclicality.  ARMA and ARIMA models are usually employed when moving average trends and auto-regressive trends are detected in the historical data and need to be forecasted forward.

Once the model has been refined, future estimated data of the driver variables can then be inserted into the model to derive the forecasted variable, sales in this case.  Once sales are forecasted, then other models can be used to forecast variables themselves influenced by sales.  If these components have a direct causal relationship with sales. Then estimating their respective costs and requirements becomes a simple mathematical process.

The process of initially designing and refining the forecasting model to maximize the level of accuracy in the historical data can take time, but the ultimate benefits are obvious,  the ability to forecast sales, can consequently cost requirements, enables company’s to adhere to strict budgets and strategic initiatives, thereby saving executive time and money.


Nicholas Kilpatrick is a partner at the accounting firm of Burgess Kilpatrick.  He leads the firm’s consulting and strategy practice and works with companies to enhance their Analytics, Forecasting , and Data Optimization functions.  The practice’s focus includes quantitative forecasting, corporate and unit strategy and planning.  Please visit our website at or on Facebook at for more information on our firm.

Leave a comment

Your email address will not be published. Required fields are marked *

Want Great Insights, Tax Planning, and Business Videos?

Subscribe To Our Newsletter!

You have Successfully Subscribed!