Measuring Errors for Quality Forecasting


It is perhaps part of human nature to react to errors as soon as they see it. Being able to identify them is crucial in just about anything especially in instances where minor discrepancies would impact the expected results a lot. Forecasting business flow is necessary in making important decisions. This involves extensive data gathering and reference of past actuals to prevent forecasting from becoming mere guesswork.

The irony is, the future is unpredictable and this is what many businesses miss. Forecast errors occur when there is no follow-up of the applicability of the predefined target in a certain time in the future. While looking at the past records of your business would give tremendous help in planning, managing forecasting would help build adaptive management in the business. Alongside forecasting are forecasting errors which are equally necessary when creating effective forecasting. This means that if you want to spot the errors in your forecasting, you should be able to create a good one.

Putting Together an Adaptive Organization

Adaptive organizations are the ones who will succeed over the changing tides of the market. Thus, every business needs an adaptive management process capable of detecting forecasting errors that could change the course of planning.

The IBM Software business analytics white paper, “Embracing Error: the first step to quality forecasting,” conveniently pointed out that learning how to manage errors is the key to enhancing financial performance management.

Experts, Steve Morlidge of Satori Partners and Steve Player of The Player Group, argued that there are 2 ways on how a company can implement adaptive management. First is to develop organizational reflexes which can be manipulated for changes without spending too much time on planning. Major companies have been using this type of organizational reflex, however this may not be realistic for many.

The second approach focuses on the ability of the organization to change course when there is a need. It’s either:

forecasting is extended further into the future to allow management convenience in case of changes
plans are purposely made dynamic as anticipation for the changes that might require reallocation of resources while pursuing the plan.

Quick response to reallocation is, of course, agreeable to any organization. The goal is to save as much time as needed whenever the situation calls for some change of plans. In order to genuinely make a move for adaptability, enhancing forecasting is crucial. Unfortunately, many organizations fall short in providing adequate forecasting that may be useful for creating new targets. The reason is that they are not able to create a forecast good enough to identify and eliminate errors.

In putting together an effective forecasting, you should be able to answer these questions:

What is there to forecast?
When is the the right time to measure the factors needed for forecasting?
What’s the frequency of measuring them?
What are you going to do with data gathered?

What to do?

Forecasting gives you a mind map of what is likely to happen in the future. In building a better perspective of how your business can fully adapt to a changing market, there are things that you should avoid doing.

Anchoring bias kills your chance of having an objective forecasting. A person’s tendency to believe what he hears first could become dangerous to future negotiations and adjustments. Neglect all kinds of bias to prevent bad decision-making. Along with this is openness to variations. To a certain degree, variations could be your key to a realistic forecasting.

Sometimes, managers tend to believe that business has a linear movement and what happened five years ago has a hundred percent chance of happening again. This fallacy tends to mislead decision makers in providing the same remedy to an existing problem. Committing errors and not realizing it could potentially be the most dangerous thing that could happen in business.

These are some steps you can follow when identifying and eliminating errors:

1. Recognize the errors in the forecast and be transparent as much as possible. Don’t worry about how the forecast will make CFOs and CEOs feel. It’s better to be completely honest about the shortfalls and surprises of your forecasting in order to build an action plan to remedy the problem.

2. Measure the data errors by finding the difference between the forecast and the real data results. Businesses create forecasts way into the future, but in measurement, comparing long term actuals to short-term forecasts may be less convenient. Ideally, forecasts should be compared to actuals with a similar time frame.

3. After gathering and measuring data, it’s time to make meaning out of the error measurements. Upon measuring the data, you’ll notice that error is relatively constant in forecasting.

4. Once error is spotted, appropriate action should be taken and adjustments should be made in planning. In this step, remember to eliminate bias in creating solutions that vary at some level. Define the action that should be done with the errors so it suits what the forecast lacks.

A good forecasting involves a lot of editing and follow-up. Eliminating biases and allowing variations in the process make it more adaptive to various performance management techniques. The practice of good forecasting will eventually make an organization realize that the forecasting process is not predefined with set of rules. Instead, you play with the unpredictability of circumstances and watch the plan grow from there.


About Benjamin Goewey

Founder, President, and Principal Consultant of Datamensional, LLC.

LinkedIn Twitter Google+

Ben has been passionate about providing actionable, relevant, and timely information for decision makers to help them make decisions more quickly and effectively and helping transform organizations as a result. He has been a Business Analytics and data guru for nearly 11 years and enjoys seeing the positive impact an effective solution can have on an organization.

Ben currently leads a team to bring the best practices of Business Analytics as well as hands on implementation to companies big and small.

Specialties: Performance Management, Analytics Strategy, System Architecture, Sales Management, Sales Analytics, Marketing Analytics, Business Intelligence, Reporting, Data Integration, IT Management, Data Warehousing, Management Decision Support, Dashboards, KPIs, Automation.