Unlocking accuracy, efficiency and agility through Continuous Touchless Demand Forecasting

Continuous Touchless Demand Forecasting is defined by Capgemini Invent as a capability that capitalizes on Big Data, Artificial Intelligence, and Machine Learning to “recognize historical patterns, select best-fit statistical models, and draw on a variety of inputs and forward-looking variables, (..) to create more accurate demand forecasts with less manual effort.” In other words, it is an automated, data-led approach to intelligent Demand Forecasting.

Continuous Touchless Demand Forecasting is increasingly part of the modern Supply Chain. Its success rests on the combination of: 1) a centralized analytical organization supported by AI, Machine Learning, Big Data and automation providing statistical forecasting, 2) a thin local organization managing exceptions, all working together in 3) a well-defined process. In this article, we consider why so many leading Supply Chain organizations are adopting Continuous Touchless Demand Forecasting.

“By 2023, at least 50% of large global companies will be using AI, advanced analytics, and IoT in supply chain operations.” (Source: Gartner Predicts 2019 for Supply Chain Operations)”

Traditional and new Supply Chain challenges require a fresh approach to Demand Forecasting. Supply Chains are confronted with increasing operating costs, high working capital, and a lot of manual work in forecasting. Furthermore, expanding product portfolios, channel diversification, and external disruptions increase complexity and a need for Supply Chain resilience and agility. Traditional Demand Forecasting is not designed for this.

Increasing accuracy, efficiency and agility

Continuous Touchless Demand Forecasting realizes higher planning effectiveness and efficiency than traditional Demand Forecasting. In the first instance, the use of Big Data, Artificial Intelligence and Machine Learning can yield a higher forecast accuracy and lower bias. Combined, they allow automatic recognition and extrapolation of patterns in demand, while considering a broad range of external data, leaving exceptions to be handled manually. Here, we can see that automation significantly reduces the amount of manual forecast touches that typically erode forecast accuracy.

Another benefit accrues in the process and organizational efficiencies derived by increased speed and automation. The automation of repetitive manual low-value tasks frees the workforce to focus on managing exceptions and higher-value work.

Finally, due to its increased execution efficiency compared to traditional forecasting methods, Touchless Demand Forecasting allows for a more responsive and continuous weekly forecasting cycle, catering to the need for more resilience and agility. The combined effects drive better planning and a more stable Supply Chain, leading to additional benefits in cost, cash, revenue, and service.

To illustrate this, we use the case of Capgemini’s collaboration with a global consumer goods company. After initial segmentation of the portfolio, we ran a parallel test for three months and compared the outcomes of the existing manual Demand Forecasting process and Continuous Touchless Demand Forecasting. The latter generated positive results across almost all selected SKU types, increased forecast accuracy with 10-15 percentage points, while highly automating the process. We are now in the process of executing the first phase of a multistep rollout plan.

Transforming the planning capability

While every organization’s journey to Continuous Touchless Demand Forecasting will be unique, there are general aspects that can guide an effective transformation of the planning capability and ensure the benefits are reaped. First, the Supply Chain organization should create a vision around the people aspect of Continuous Touchless Demand Forecasting. It is important to have a deep understanding of the current and potential capabilities of the planning teams, to determine the right fit with future needs. Secondly, process mapping should be used to understand current demand forecasting pain points. By assessing the gap between as-is and to-be processes, a future-proof Operating Model can be designed. Thirdly, organizations should be aware of the rapid development in technologies. It is important that the implemented technology is a perfect fit with your planning evolution and can answer your key planning challenges, as well as being a match with your current level of planning maturity. Finding the right technology supplier is critical to the success or failure of implementing new Demand Forecasting capabilities. The above describes the first crucial steps of your journey towards Continuous Touchless Demand Forecasting.

Key takeaways

Traditional and new challenges in Supply Chains require a fresh approach to Demand Forecasting, in the form of Continuous Touchless Demand Forecasting. Compared to Traditional Demand Forecasting, it provides:

1. A higher forecast accuracy and lower bias using Big Data, Artificial Intelligence and Machine Learning, while still manually managing exceptions.
2. Process and organizational efficiencies through automation of manual forecasting tasks.
3. A higher degree of agility & responsiveness through a shorter forecasting cycle.

Combined, this results in benefits in cost, cash, revenue, and service for Supply Chains. To realize this successfully, the following combination is required: a clear vision on the people aspect, a centralized analytical organization providing the statistical forecast & a thin local organization managing exceptions, new technological Touchless Forecasting assets, the right match with a planning organization’s level of maturity & key challenges, and a well-defined Continuous Touchless Demand Forecasting process.

At Capgemini Invent we work with Supply Chain organizations globally to help them embark on this journey and see it through to the end with the ensuing cost, cash, revenue, and service benefits. We have broad experience in facilitating centralized analytical organizations for our clients utilizing Big Data, Artificial Intelligence and Machine Learning to produce statistical AI/ML forecasts as-a-service.

This blog is part 1 of a series on Autonomous Planning. You can follow the series on Capgemini Invent’s LinkedIn or our website. Please contact one of our experts below to start your journey to Continuous Touchless Demand Forecasting.

Authors

Mark Enthoven Senior Management Consultant in Supply Chain Planning

Sven de Jong Senior Management Consultant in Supply Chain Planning

Peter Tacken Managing Consultant in Supply Chain Planning