AI4Business pre event
21/02/2018

Demand forecasting in S&OP takes the stage at AI 4 Business Summit

The AI 4 Business Summit is just around the corner. On 27 February, experts and business leaders from a range of Belgian tech companies will speak on AI-related topics in four key areas. delaware's own Maarten Herthoge and Tom Bergmans will be giving a presentation on their area of expertise: demand forecasting with Microsoft Azure.

Fact: almost any business uncertainty can be modeled and predicted using a set of machine learning algorithms. Put simply, demand forecasting is the science of using historic data about customers, events and trends to predict customer response to your product (or idea).

Rely on data, not gut feelings

Once a company grows beyond a certain size, the S&OP processes and systems – or Excel spreadsheets – that it has relied on lose their efficiency. Growth implies increasing complexity, making it more difficult to understand what truly drives sales and incorporate it effectively in your company’s day-to-day operations. There are plenty of insights lying dormant in your (valuable) data, but they can no longer be revealed by a set of simple calculations. Enter demand forecasting.

“A company of any size can accomplish a lot with the right model,” explains Tom. “Demand forecasting can, for example, enable you to give customers exactly what they are looking for in that particular moment. Determine how much you should spend on a campaign to reach an audience with maximum return on investment. Proactively boost sales where forecasts predict a dip. Get a clearer understanding of when you will need resources. The possibilities are endless.”

 

 

“Demand forecasting is a mature approach, but it has to be done correctly, and the solution must grow with the company.”

Bridging past, present and future – and multiple data sources

On the sales and marketing side, a 360-degree customer overview is an invaluable tool that can be used to predict demand for a product, or customer reaction to different brand profiles targeting specific customer subgroups.

“Even if you have limited information about a potential customer, an appropriate model uses many data sources to fill in some of the blanks, enabling you to approach them in a way that matches their preferences,” Maarten asserts. “Companies can also predict how social media campaigns or targeted ad campaigns will influence a product introduction. After the introduction, they can score its effectiveness, including collateral effects.”
On the operations planning side, companies can use demand forecasts to proactively determine operations, logistics and warehousing needs with precision. “That means no expired products or packed warehouses – or consumers who are forced to buy the product of your competitor because they can’t find yours on the shelf.”

Quantifiable results, more transparency

In addition to S&OP, machine learning also has an impact within an organization. When companies trust their tools and see firsthand that a well-designed, well-trained model leads to quantifiable results, stakeholder confidence and internal flexibility get a huge boost.

“Thanks to the clear pictures of the future generated by predictive analysis, business leaders can guarantee a smoother cashflow, anticipate risks and take proactive measures to avoid those risks,” Tom explains. “Even more, when the business has complete control over the inputs, this transparency avoids that suspicious ‘black box’ feeling. Add a familiar frontend, and users have an interactive tool that they understand and trust.”
Interested in a hands-on demand forecasting demo with Microsoft Azure? Attend Maarten and Tom’s exclusive presentation (16:30h, PM Track: hands-on labs) and get the full scoop on other disruptive AI applications at the AI 4 Business Summit.
You can connect with Tom Bergmans and Maarten Herthoge on LinkedIn.