• Home
  • /
  • IOT
  • /
  • IOT of the future – IOT data in demand forecasting

IOT of the future – IOT data in demand forecasting

Demand forecasting optimises raw material procurement, influence supplier choice and co-ordinate logistics, says Richard Barry, CEO of Polymorph.

Johannesburg, 22 Aug 2019

Annually, manufacturers lose substantial amounts of money to waste. The reasons, as every production or line manager will be able to point out, include “unplanned line downtime, material waste, overproduction, quality-related losses, and poor use of capacity or assets”.

Richard Barry, CEO of Polymorph
Richard Barry, CEO of Polymorph

However, what they fail to consider is the more subtle form of waste, namely, inconsistent or inaccurate data that leads to flawed insights and poor decisions being made. In order to eliminate this type of waste, manufacturers require access to consistent, reliable data that allows opportunities for continuous improvement.

Unfortunately, many manufacturers still collect data manually, which leaves the door wide open for inefficiencies, delays and weak data integrity as a result of differences in human interpretation. The latter can further lead to errors in analysis and ultimately financial losses. With that said, these challenges also bring with them opportunities for advancement, such as leveraging industrial IOT to collect data, forecast demand and remove human error and inconsistencies.

What is demand forecasting in IOT

According to Richard Barry, CEO of Polymorph: “The focus of demand forecasting in industrial IOT turns on predicting demand based on historical data and analogous sensors.” Demand forecasting can be used in various fields and industries such as logistics and manufacturing (for example, predicting the production output of a factory site). Demand forecasting not only assists businesses to plan well, for example, by being able to forecast labour demand, but it is also able to optimise raw material procurement, influence supplier choice and co-ordinate logistics.

Industrial IOT customers can benefit tremendously from demand forecasting as optimising their operations can considerably reduce costs. As companies are increasingly moving towards using IOT and data to track and monitor different elements of their operations, the data can be advanced to enhance demand forecasting. Demand forecasting can be further improved by “customising the model to accommodate domain-specific requirements and to suit the unique business needs of the customer”.

Why forecast demand?

Improved accuracy – demand is a complex interaction between many different variables. For humans to consider all these variables adequately when doing forecasting is difficult, if not impossible. Machine learning models can simplify these complex interactions and forecast demand much more accurately.

Improved tracking – demand forecasting has historically been left to experienced supervisors working in the field for decades. Even so, historically, demand forecasting was more gut-check and less empirical. This meant it was often challenging to ascertain if and how well forecasts matched to demand. A predictive forecasting model and the ability to track forecast versus demand can be used to further improve the model.

Employee turnover – as experienced employees retire or resign, their know-how of the business processes and, specifically, demand forecasting is lost. Having a predictive demand-forecasting model will assist with capturing this knowledge and so equip the new employee to adequately plan for the variations in demand.

Better customer service – a proper demand-forecasting model will see your business not only meet customer demands, but provide exceptional service too.

Benefits of using IOT data in demand forecasting

1. Real-time data for more accurate forecasting optimised inventory levels and customer satisfaction

It is a known fact that a lack of timely data can result in either over- or under-supply. Real-time data can be used for increased forecasting accuracy and, in turn, can create better business outcomes. First, sensors provide “real-time visibility of actual stock levels in operations or machines”. Thereafter, the data is transmitted to a platform that analyses those stock levels and can automatically trigger replenishment orders when required. This results in “increased sales from reduced stock-outs”, optimised inventory levels because of automatic replenishment triggers that do not require human interaction, and a reduction in the costs of logistics as stock is delivered only when required. Therefore, utilising real-time data for accurate forecasting that results in optimised inventory levels that meet customers’ demands and increase their satisfaction will ultimately result in revenue growth for the company.

2. More accurate labour scheduling to meet customer demands

Scheduling the right people at the right place and time sounds easy enough, yet it is not that simple. Factors such as unpredictable demand, seasonality, holidays, etc, must be considered in labour scheduling. Losing sales because of understaffing or incurring increased labour costs as a result of overstaffing can have a substantial impact on the overall business and customer satisfaction. Using IOT data in labour forecasting provides labour planners with accurate projections of sales, transactions and foot traffic over time. Combined with accurate demand forecasting data, accurate schedules can be compiled to avoid over- or understaffing.

3. More efficient maintenance with just-in-time predictive maintenance

Downtime of already expensive machines on the factory floor can be costly. Leveraging IOT intelligence (adding sensors and performing modelling and diagnostics on the data collected), allows you to move from expensive scheduled maintenance plans to just-in-time predictive maintenance to repair, clean or replace parts only when needed.


Customer demands are changing rapidly. Collecting data over time and analysing the patterns in such data enable accurate demand forecasting for both inventory levels and labour and also trigger warnings for intervention in the event of faulty operations or required maintenance. The results are optimised inventory levels, reduced maintenance costs, equipped production, and procurement personnel making better procurement decisions, and forecasts are better and more accurate. Ultimately, customer demands are met and the company’s revenue increases.

If you would like to get in contact with Richard Barry, please visit our Web site or e-mail him at


Editorial contacts