Demand side
One number. Agreed, not negotiated.
Demand planning is where a business decides what it expects to sell. Done well, everything downstream has something firm to stand on. Done as a monthly argument, everything downstream inherits the argument.
The demand side covers three of the fourteen capabilities: Demand Planning, Sales Collaboration and Scenario Planning. Together they produce one agreed demand number the whole business plans against, a structured handshake with commercial teams, and the ability to test a commercial call before it is committed to.
Definition
What demand planning actually is
Demand planning is the process by which a business arrives at a single view of what it expects to sell, by product, by location and by period, and then commits to planning against it. The statistical forecast is an input to that process. It is not the process.
The distinction matters because the two fail in different ways. A statistical forecast fails when the method is wrong for the demand pattern, and that is a modelling problem with modelling answers. A demand planning process fails when there is no agreed owner of the number, no forum in which commercial and supply views are reconciled, and no record of who changed what and why. The second failure is the more common one, and it does not appear anywhere in an accuracy report.
Capabilities on this page
- 01Demand Planning
- 02Sales Collaboration
- 03Scenario Planning
What breaks
What breaks without it
None of these is a symptom of a weak model. Each is a decision that was never assigned, or a disagreement with nowhere to go.
The number is a negotiation
Sales publishes a target, planning publishes a forecast, and the plan is whichever one the last meeting favoured.
Neither side is being unreasonable, because nobody agreed what the number was for. A target is a statement about ambition and a forecast is a statement about expectation. A business that uses one word for both cannot tell when it is missing, or which of the two it is missing against.
Bias is argued about rather than measured
Everyone knows the forecast leans one way. Nobody knows by how much, for which segments, or whether it is getting worse.
Bias is the one property of a forecast that is cheap to measure and immediately actionable, because a consistent lean in one direction is a policy problem rather than a modelling problem. Left unmeasured it is absorbed by inventory, which is the most expensive way to hold a disagreement.
The items with least history get least method
New products, promotions and phase-ins are handled by manual override, and the overrides are never reviewed.
These are the items where the forecast matters most and where statistics help least, so judgement has to carry the weight. Judgement applied without a method and without a record is indistinguishable from guessing, including to the person who applied it.
Scenarios are run in the crisis, not before it
The ability to test a commercial call exists, but it is used to explain what happened rather than to decide what to do.
A scenario run after a commitment is a post-mortem. The capability is the same either way; what changes is whether the planning calendar has a point at which alternatives are expected to be on the table.
What changes
What changes when it works
Not a better model. A number with an owner, a forum, and a memory.
- 01One demand number, owned by a named person
- 02Bias measured by segment and reviewed on a cadence
- 03A structured handshake with commercial teams
- 04Overrides logged with a reason, and reviewed
- 05Scenarios available before a commitment
- 06Disagreement surfaced in the process, not in the results
The engagement
How an engagement touches it
The assessment looks at how the demand number is produced today: who owns it, where it is changed, what happens when commercial and supply disagree, and what the working week looks like on a bad week rather than on the org chart.
If demand is the load-bearing gap, operating model design defines the roles and decision rights around the number and the calendar it runs on, and the build and execute stages stand that up with your planners in the meetings where the trade-offs are actually made. If it is not the load-bearing gap, we say so, because sequencing is most of the value.
Which capability is load-bearing decides the sequencing, and that question comes before this one. The full argument is set out in why a better forecast does not move service levels. The people who would do the work are on the practitioners page.
Questions
Questions we get about Demand Planning
What is the difference between a forecast and a demand plan?
A forecast is a statement of what is expected to sell. A demand plan is what the business commits to planning against, which is the forecast plus the commercial judgement applied to it and the constraints that judgement has to respect. A business can have an accurate forecast and no demand plan, and that is a common place to be.
How much forecast accuracy is good enough?
It depends on what the number is used for, and the more useful question is usually whether accuracy is the binding constraint at all. If buffers are stale, if nobody owns the number, or if the monthly meeting reports rather than decides, improving accuracy will not move service levels.
Do we need a new planning system to fix demand planning?
Not to start. Demand planning is a process with owners, a cadence and decision rights, and all three can be established on what you already have. Instrumentation makes a working process faster and a broken one faster to get wrong, which is why digital enablement sits sixth in the framework rather than first.
The rest of the map
The other three
- Supply and Inventory PlanningThe supply side is where a demand number becomes a commitment. It is also where the difference between a plan and an aspiration stops being a matter of opinion and starts being a matter of cost.
- S&OP and IBPMost businesses have a monthly planning meeting. Fewer have a monthly planning decision. The difference between the two is the whole of S&OP.
- Performance and RiskA plan is a claim about the future. The loop is how a business finds out where the claim was wrong, early enough to do something, and without turning the finding into an argument.
All fourteen capabilities, grouped, are on the services map.
Get Started
Where is your planning organisation today?
Every engagement starts with an honest assessment of current maturity. Let's find out where you are, and what the next capability should be.