There are only 3 things management can ever control:

• The arrangement of its own business processes (e.g. sales, production, customer service, etc.)
• The resources (money, people, assets) applied to those business processes, and
• The volume and quality of work carried out within those business processes

However, organisations operate in an uncontrollable and often unpredictable business environment, such as market demand, energy, inflation and exchange rates. As a consequence, the role of planning is to help manage what can be controlled to produce outcomes that will achieve organisational objectives, within an uncontrollable and unknowable external environment.

With this in mind, an analytic model can be used in a number of ways as outlined here:

• To confirm relationships that managers believe exist within an organisation and their impact on performance. These may be between those that are controllable such as resources applied and production capacity, and those outside the control of the organisation, such as the weather and interest rates. Relationships are typically set up as ‘rules’ that use historic base data to generate the results achieved, taking into account any time lag between ‘cause and effect’. By doing this, managers are able to confirm that the organisation operates as expected, and they can assess how this compares with other organisations/industries.
• To explore the boundaries of performance caused by unpredictability in the market. Once a model’s relationships have been confirmed, it can now be used to determine variability in the drivers of the business. For example, if raw materials costs were to increase by 5% and demand reduce by 10%, what would be the impact on profitability and cash flow. By running multiple scenarios, managers can get a better understanding of how a change in one part of the business impacts overall results.
• To assess changes to the way in which the business operates. Business performance can only be improved so much by cutting costs. The most radical changes come when business processes are performed better, more efficiently or differently. Analytic models can simulate these changes, such as outsourcing or the introduction of new initiatives, by making changes to the model’s relationships. Managers can then prioritise how its business operation should evolve over time along with a forecast of the likely costs and impact on performance.
• To predict future results and set interventions.  By using a combination of the above, analytic models are able to predict a range of future results from pessimistic to optimistic. These forecasts should cover the short-term, say 1 month, to the longer term such as 3 to 6 months into the future. Managers can then set thresholds that represent specific levels of individual business process performance at which interventions will take place. For example, if a 3-month forecast shows a 10% under budget variance in sales, then marketing spend will increase to generate more demand.

It's worth emphasising that analytic models can never accurately predict results as it’s impossible to code or even know all the relationships involved. Their true value is in allowing managers to assess the impact of uncontrollable events and the cost of implementing changes to offset them when they arise.

The article was first published in ​prevero Blog