Forget forecasts, businesses have an inventory problem
By Jon Taylor on August 14, 2023How many conversations do you have in a month about forecasts? Whether it be for demand, on the supply side, for SKUs, region, category, or fulfillment… that ever elusive, exceptionally accurate forecast isn’t going to happen.
So why do we still spend hundreds of thousands of hours tweaking, iterating and obsessing over our forecasts?
“This is something of a pet peeve of mine,” says Simon Spavound, Peak’s Head of Data Science Ops USA — who has a PhD on using machine learning (ML) for forecasting and was formerly a university lecturer on the subject.
“A forecast is only as good as the process it supports. If it’s part of an effective planning function that successfully optimizes inventory, determines reorder points and reduces movements, then a forecast with 85% accuracy could deliver higher service levels than one with 90% accuracy.”
In other words, even the most accurate forecast on its own is not enough: it’s what you do with it that counts.
A forecast is only as good as the process it supports.
Simon Spavound
Head of Data Science Ops USA at Peak
A new approach to forecasting
Take it right back to basics, and demand fulfillment — the ultimate goal of a forecast — is about having the right stock in the right place at the right time.
“The temptation has always been to get the demand forecast as accurate as possible,” explains Peak’s Product Marketing Director, Ira Dubinsky, who has spent the past 15 years leading teams that generated demand — including his former role as Innovation & Marketing Director for KFC across multiple markets.
of consumer goods companies say demand forecasting is a top area of focus (CGT)
“But in today’s climate, maximizing availability and margin and meeting OTIF targets is about how well you can manage volatility across the entire supply chain. Squeezing an extra 2% accuracy out of a forecast doesn’t offer a meaningful competitive advantage — a forecast shouldn’t be the priority over everything else, it’s one piece of a bigger puzzle.”
Simon agrees, explaining that “if your end goal is to maximize on-shelf availability of stock (and keep costs down while doing it), then the focus should be on your end-to-end supply chain — the forecasting process is just one part of this.”
A forecast shouldn’t be the priority over everything else, it’s one piece of a bigger puzzle.
Ira Dubinsky
Product Marketing Director at Peak
Linking demand and fulfillment
Supply chains typically run on aggregated data, which gives a ‘big picture’ overview across the supply chain. By contrast, customer operations — where demand is generated — is more granular. The focus here is on understanding the consumer, through demographic and other types of segmentation and category planning.
“Consumer data and the intricacies of working with consumers aren’t often considered in the supply chain,” says Ira. “But there is real power in using this data to fuel supply chain operations and supply chain data to fuel customer operations.”
Linking fulfillment and demand data represents a huge opportunity for demand planners. Location data can help to plan where to place inventory and optimize transportation and warehouse locations. Fulfillment times and time horizon to restocking can be used to manage customer expectations and allocate orders. Trends in customer behavior can help companies to anticipate promotions, or disparities in final order quantities. These are all traditionally objectives of an accurate forecast.
of CPGs don't believe their business is data-driven (Peak)
“Ensuring you have enough stock to keep service levels high without holding too much should logically be solved by an accurate forecast,” says Ira. “But in reality it’s a function of the entire supply chain, so the focus should be on inventory rather than forecasting.”
This means you need a broader overview of what is happening at every stage of supply, including demand generation.
“Peak’s approach has been to include forecasts in each of our Inventory Intelligence apps, enabling customers to capture different demands,” explains Simon. “But the forecast is a starting point. Our Dynamic Inventory application, for example, uses a forecast to predict demand and then applies safety stock models on top that consider a broad range of data from both customer and supply operations. It optimizes for volatility and then determines inventory levels based on that uncertainty.”
of CPGs report consumer insights as an analytics focus (CGT)
Leveraging data from the full breadth of the supply chain provides insight that is increasingly valuable, particularly in an environment of volatile demand, goods shortages and transportation challenges.
Businesses don’t have a forecasting problem, they have an inventory problem. And aligning data from across the supply chain is how they solve it.
Three easy steps to linking demand and fulfillment
- Prioritize cross-functional collaboration: reach out to colleagues and build relationships in other departments
- Collect consumer data whenever you can. The future is about 1:1 relationships with consumers and a consumer centric way of operating. Even if you don’t have a direct to consumer model, use every marketing channel available to you to drive signups and then provide rich experiences to nurture that relationship
- Break down internal silos to bridge the divide between demand generation and demand fulfillment. Build marketing campaigns that are fed by stock data and supply chain programs that fulfill a market of one.
See for yourself!
Find out how Peak can help you connect demand and fulfillment — book a demo here.
Get the AI for inventory guide
Right stock, right place, right time. Transform your forecasting to drive demand and supply efficiencies with AI.