How it is changing the demand planning game?
Demand planning plays a crucial role in supply chain management as it has a pervasive impact on all three levels of decision making: strategic, tactical and operational decisions.
To ensure sustainability in an ever-changing and competitive domain, companies must constantly adapt and readapt supply chain technology and processes. And a natural starting point to this is demand planning.
Data analytics technology has developed exponentially over the past few years due to the phenomenal increase of computer capabilities, algorithms and the emergence of easy to use, simple and user-friendly BI tools. Many companies have accumulated large volumes of data coming from a varied number of sources. A key success factor for companies lies in the capability to leverage such data efficiently to the supply chain advantage.
Data analytics is widely explained in three approaches. Understanding and analyzing historical data (descriptive analytics), forecasting the future (Predictive Analytics), and thirdly decision-support (Prescriptive Analytics). How will Analytics change the way we forecast today?
Descriptive Analytics – What has happened?
It is the interpretation and visualization of historical data (sales history for example), from aggregated levels to the most detailed. One of the major benefits of Descriptive Analytics is the ability to work based on exceptions rather than having the need to wade through astronomical amounts of data.
This, in turn, generates time that can be used in adding value to the process, like detecting a bullwhip effect or identifying seasonal variations in sales. This is the first step towards setting up continuously improving processes.
In summary, Descriptive Analytics is focused on diagnostics analysis. This supports companies to analyze, understand and visualize historical data. This is a critical step towards identifying problems and improving the overall supply chain performance.
Predictive Analytics – What will happen?
Based on historical data and Descriptive Analytics, Predictive Analytics tries to define a picture of the future by forecasting various data series. The aim is to augment the decision-making process.
Methods such as machine learning or data mining are often labeled predictive models. The main idea of these techniques is to search, discover and analyze patterns, relationships, and correlations between variables. This includes history sales, item attributes or external data like POS or CRM.
In order to make forecasts coherent with the historical data, it is restricted within a Likelihood corridor. This is where Descriptive and Predictive analytics work hand-in-hand by looking at what was predicted and what actually happened. This is done by comparing an archive of sales forecast for a particular period with known sales values.
Prescriptive Analytics – How to make it happen?
By studying this data: History, forecasts & results, Prescriptive Analytics has the capability to build several what-if scenarios in order to give the user, flexibility of choosing the best possible outcome for the business.
As an example, they could analyze what was sold well in the past, why and when, and then suggest a recommendation like a promotional quantity in order to gain a market share. Systems could also recommend changes on plans duly taking into account extraneous factors such as weather, macro-economic or political issues.
In both cases, Prescriptive Analytics offers recommendations, opportunities, outcomes and possible solutions to issues.
There are many possibilities and at the outset, it might even be intimidating. But this the impact this could have on the decision-making process outweighs the complexities and investing in such technologies generate a fast and significant return on investment.
Descriptive, Predictive, and Prescriptive Analytics are not just buzzwords. They are gaining maturity and they are becoming the norm in a lot of companies, helping them improve their decision-making process on an everyday basis. We at QAD Dynasys R&D develop innovative functionalities leveraging Analytics and Machine Learning techniques to support planners and managers make fact-based decisions and overcome supply chain challenges.