End-to-End Supply Chain Leadership & Transformation
Demand & Forecasting
Forecasting doesn't predict the future.
It provides a structured view of what might happen — so the business can make better decisions about what to do next.
HMS helps manufacturing businesses develop practical demand and forecasting processes that support planning, purchasing, production, inventory and wider business decisions.
The focus isn't simply on producing a forecast.
It's on making the forecast useful.
The forecast will be wrong.
That's probably the first thing worth saying.
No forecast can predict exactly what customers will buy.
Demand changes.
Customers change their plans.
Markets change.
Products change.
And sometimes something happens that nobody could reasonably have predicted.
So the objective shouldn't be to create a forecast that claims to predict the future.
The objective is to create the best useful view of future demand available, understand the uncertainty around it and use that information to make better decisions.
That's very different.
What is the forecast actually for?
A forecast isn't an end in itself.
It should help the business answer questions such as:
-
What might customers need?
-
How much capacity could we require?
-
What materials might we need to purchase?
-
What inventory could be required?
-
Where might supply constraints appear?
-
How much production capacity could be needed?
-
When do we need to make decisions?
-
What could happen if demand is higher or lower than expected?
A forecast becomes valuable when it changes a decision.
If nobody uses the forecast to make a decision, why are we forecasting?
Orders aren't necessarily demand.
One of the challenges I've seen repeatedly is treating customer orders as though they represent the complete picture of demand.
They don't always do that.
An order may arrive today for delivery tomorrow.
By that point, the opportunity to plan for it may already have passed.
The business may already need to have considered:
-
Capacity
-
Materials
-
Labour
-
Inventory
-
Supplier lead times
-
Production requirements
-
Customer commitments
That's why effective demand planning needs to look further forward than the order book.
The purpose is to give the supply chain time to respond.
Forecasting and planning need to work together.
A forecast sitting in a spreadsheet doesn't improve a supply chain.
It needs to connect with the wider planning process.
Demand → Supply → Capacity → Production → Materials → Inventory
That connection is central to effective S&OP and supply planning.
My experience includes developing and implementing integrated demand forecasting and supply planning solutions, including linking forecasting outputs with MRP and wider S&OP processes.
A forecast that says demand will increase is useful.
A forecast that allows the business to understand what that increase means for production, materials, capacity and inventory is much more useful.
How much detail do you really need?
Forecasting can become unnecessarily complicated.
I've worked with forecasting environments ranging from individual customer and product forecasts through to solutions managing 20,000+ to 60,000+ forecasting points.
The answer isn't always more detail.
The right level of detail depends on what the forecast is being used for.
A forecast for commercial planning may need a different level of aggregation from a forecast supporting material purchasing.
The important question is:
What decision is this forecast intended to support?
Start there.
Then determine the appropriate level of detail.
Forecast accuracy isn't the whole story
Forecast accuracy is useful.
But it can also become a distraction.
A business can spend considerable effort trying to improve a forecast accuracy percentage without asking whether the forecast is actually helping the business make better decisions.
For example:
A forecast might be statistically accurate at an aggregated level but provide little useful information for production planning.
Or a forecast might be inaccurate at individual item level while still providing valuable forward visibility of overall demand.
The measure needs to reflect the purpose of the forecast.
Otherwise, the business can end up improving the number rather than improving the decision.
Building a useful forecasting process.
Effective forecasting isn't simply about selecting an algorithm.
It involves understanding:
The data
What history is available?
Is it reliable?
Are there gaps, unusual events or structural changes that need to be understood?
The demand
What actually drives demand?
Are there seasonal patterns, customer behaviours, product changes or other factors that need to be considered?
The process
Who owns the forecast?
How is it reviewed?
Where does commercial knowledge enter the process?
How are exceptions handled?
The planning horizon
How far ahead does the business need visibility to make meaningful decisions?
The output
How does the forecast feed into S&OP, supply planning, MRP, purchasing, production and inventory?
A good forecasting process connects all of these.
Forecasting supported by practical experience.
I've worked with both established forecasting systems and practical solutions developed around the needs of the business.
Examples include:
60-week forward forecasting
A recent solution used three years of customer and item sales history to generate forward item-level forecasts, which were loaded into MRP to provide material demand requirements over a 60-week horizon and support purchasing and contract negotiations.
60,000+ forecasting points
A specialist demand forecasting solution was scoped, sourced and implemented for a business requiring management of more than 60,000 forecasting points, alongside a structured monthly S&OP process.
20,000+ forecasting points
A SaaS-based, ERP-integrated demand forecasting solution was sourced for an environment managing more than 20,000 forecasting points, with the forecasting process supporting the wider S&OP cycle.
45,000 tonnes across 1,450 customer listings
Earlier in my consultancy career, I designed and implemented planning solutions and sourced a specialist demand forecasting solution to forecast 45,000 tonnes of product across 1,450 customer listings.
These are different businesses and different requirements.
The common principle is the same:
The forecasting solution needs to serve the planning process — not become the planning process.
Technology can help. It isn't the starting point.
I've worked with specialist forecasting software, ERP-integrated solutions and Excel-based forecasting and supply-planning models.
The technology available today is considerable.
But technology doesn't answer some of the most important questions:
-
What are we actually trying to forecast?
-
Why do we need the forecast?
-
Who will use it?
-
What decisions will it support?
-
How far ahead do we need visibility?
-
How much detail is useful?
-
What information needs human judgement?
Those questions need answering first.
Then choose the appropriate technology.
When might your forecasting process need attention?
You may recognise some of these:
-
Forecast accuracy is discussed constantly but doesn't seem to improve decisions.
-
The forecast is produced but isn't trusted by operations.
-
Sales and supply chain have different views of future demand.
-
The forecast is heavily influenced by the latest customer orders.
-
Planners spend more time adjusting the forecast than using it.
-
Forecasts don't extend far enough to support purchasing or capacity decisions.
-
Forecasting is disconnected from S&OP.
-
MRP receives forecasts but the resulting requirements aren't trusted.
-
The business has a forecasting system but limited confidence in the output.
-
Forecasting has become unnecessarily complicated.
The question isn't always:
“How do we get a more accurate forecast?”
Sometimes it's:
“What do we need the forecast to help us decide?”
From forecast to decision.
The purpose of demand planning isn't to predict the future perfectly.
It's to give the business enough forward visibility to make better decisions.
Understand the demand.
Understand the uncertainty.
Connect it to supply.
Make the decisions.
Need better forward visibility?
Whether you're reviewing an existing forecasting process, implementing a new solution or trying to connect demand forecasting with S&OP and supply planning, the first step is understanding what the business actually needs from its forecast.
Let's have a conversation.
Get in Touch