Podcast: Where Industries Fall Short in Demand Forecasting
Next in our podcast series, LLamasoft’s Industry Principal, Matt Tichon, talks with me about when it comes to new product launches, seasonal offerings and balancing inventory to meet customer demand, businesses still depend heavily on spreadsheets to chart the course ahead. But these manual processes often fall short, leading to stock outs, lost sales or excess waste throughout the supply chain.
Listen to the podcast:
Tichon has over 20 years’ experience as an executive-level change agent in supply chain planning. He has implemented and sold forecasting systems, managed demand planners and built forecasting departments from the ground-up. In 2004, he was hired by a leading quick service restaurant with ~3,500 locations to build out a limited-time offer forecasting process and department. What he found was a siloed organization that was not sharing data or collaborating in a way that was meaningful to stakeholders.
“They were missing fill rates with the customers, stocking out of a lot of ingredients. You had distributors that were throwing away ingredients that were expiring,” he shares. He witnessed the marketing department review point-of-sale data, and then summarize the sales data based on market offerings into an email to send to the supply chain purchasing department and their suppliers.
“For growth, many companies are merely adding an overall factor to their forecast, for example 5-8%, with no real understanding of what products will drive the growth and why they will drive it,” Tichon observes.
In this specific case, he explains, “It became an issue as the annual dues the stores were paying were actually less than what each restaurant was paying every year to write-off expired food to the distributors. The inventory that was rotting and getting thrown out at these distribution centers and never even made it to my restaurant cost me more than to be part of the purchasing function.”
Tichon introduced Microsoft Excel and Access to manage the supply chain. At the time, he says they were usable software tools that created less errors.
However, it posed massive challenges. “When someone new onboarded with the company, they got this spreadsheet with no instructions or user manual. They didn’t even know which set of data to input, where the formulas were located and how the data impacted the other sheets.”
He added that it wasn’t a scalable process. “You put data in, you ran some macros and you got data out. But it wasn’t repeatable.” Reviewing previous months’ data was impossible because the workbook was updated. Further panic was caused by corrupt spreadsheets due to macro errors.
During his consultations, he helped companies transition from Excel, which lacked advanced statistics, and move toward statistical packages that scaled and created consensus forecasting.
“Companies create a non-biased forecast but then they pass it over to marketing and sales to add their expertise, which becomes bias,” Tichon declares. “You’re generating a non-biased forecast, but it’s based on your bias.”
He says it comes down to looking at the data to make decisions. “From a data-driven standpoint, that’s really where it’s at – getting back to the numbers and taking out the emotional bias. For forecasting, that baseline is critical, that you can justify it and it’s based in reality.”
“It starts with exposure,” he explains. Companies should educate themselves on how to move the needle forward in the business and ask, “What if we looked at this a different way, applied some tools to it, automated a lot of this, made data-driven decisions, and freed people up to do more value added things in the supply chain?”
He advises, “Look at examples and case studies. ‘Hey, a company just like us in the same marketplace is having these kinds of results.’ I think that’s what really makes a difference and starts those conversations – just natural curiosity.” Tichon encourages diversity, “It’s crucial to just have a different opinion and an outside influence that can bring those ideas and spark the change process.”
Learn more about LLamasoft’s demand modeling solutions and how you can better predict the future for savings, efficiency and growth.