Move manufacturing
Transition months where you are paying for both, a quality dip that customers can feel, lead times that get worse before they get better, and the unit cost that made the case in the first place.
Move manufacturingFor operators
The change everyone agrees with and nobody will sign
Move the line. Consolidate the sites. Take a layer out. Second source the part that has one supplier. The savings are easy to calculate and the second order effects are the reason the paper has been in draft since March. Counterfirm runs the second order effects: capacity, throughput, the people who leave afterwards, and the customers who notice.
Capacity is people, people cost money, money is cash. The engine carries all three in the same run.
The saving lands in month one. The consequence lands in month four, and it is bigger.
Take fourteen people out and the payroll line moves immediately. Utilisation on everyone left moves too, then morale, then attrition, then lead time, then the accounts that were tolerating your lead time because they liked the account manager who has now gone. A cost case that stops at the payroll line is not wrong, it is half of the sum.
The operating decisions
What is already set up as a scenario
Eliminate a department
The cost is the easy half. The run carries what that function was holding up, and shows it arriving in churn, in delivery or in the sales cycle rather than in the payroll line.
Eliminate a departmentA supplier raises prices
Cost share, concentration, lead time and how much of a rise you can actually pass on. Four numbers that decide whether this is an irritation or a year.
Supplier raises pricesHire into the constraint
Cost out for two quarters before the first new person clears quota or reaches full productivity, with a ramp most plans quietly leave out.
Hire a teamReduce headcount
Severance, the month the saving really starts, morale, the attrition that follows the announcement, and the revenue attached to the people who chose to leave.
Reduce headcountLaunch the thing engineering finished
Attention is a capacity constraint too. A launch takes people off the work that pays today, and the run puts both on the same page.
Launch a productHow capacity is carried
Headcount is not a cost line, it is a constraint
Each twin carries revenue a head can serve per month, monthly cost per head, utilisation, morale, monthly staff attrition, months before a new hire is productive, and the cost to hire one person. When demand rises past what the people can serve, service degrades and the model lets it degrade rather than assuming infinite elasticity of labour.
That is why a growth scenario and a cost scenario can be run against the same twin and disagree usefully. One is constrained by cash, the other by hands, and the point where they collide is normally the real decision.
- Utilisation above the comfortable range feeds morale, and morale feeds attrition
- A new hire is not productive on day one, and the onboarding months are an assumption you can set
- Losing people costs the revenue they were serving, not just the salary you stop paying
Supply
One supplier, one part, one very bad month
Supplier share of cost, supplier concentration, lead time in weeks and the share of a cost rise you can pass on are four separate assumptions. Most cost cases collapse the last one into a single optimistic number, and it is usually the one that decides the year.
The invented worked example is single sourced on a marine grade billet, which is deliberate. It is the shape of risk that never appears in a monthly pack because nothing has gone wrong yet.
- A lead time change shows up as delivery before it shows up as revenue
- Pass through is set by your customers, not by your intentions, so it sits with the customer agents
- The competitors in the run have the same supplier problem, which changes how much you can pass on
Tripwires
The month the plan tells on itself
Every run returns the earliest month at which reality can contradict the model, with the metric to watch and the threshold. If a consolidation was going to hurt lead time, you know which month and which number will say so first.
This is the part that turns a simulation into an operating tool. You are not waiting until the annual review to find out the case was wrong. You have a date, a measure and a line, and if the line is crossed you reopen the decision on evidence instead of on atmosphere.
Honestly
What this will not do for you
It does not model your floor. There is no routing, no queueing, no machine level scheduling, no bill of materials and no shift pattern. Capacity is revenue a head can serve per month, which is a blunt instrument on purpose. If your question is about a bottleneck between two specific cells, this is the wrong tool and a discrete event simulation is the right one.
It does not know your process. A twin built from financial statements and a customer list has no idea that the second site has a forklift problem or that one team has carried the last three launches. That knowledge stays with you, and the model is a way of testing it, not a replacement for having it.
It runs monthly. Anything whose whole story happens inside four weeks is invisible here.
- No shop floor, no scheduling, no bill of materials
- Monthly steps, so a two week disruption is a rounding error
- It will not tell you how your people will feel about the change, only what the model does if morale moves
Month 1
The cost moves
Payroll, severance, transition costs and the new unit price all land where the paper said they would. This is the part the business case already had.
Month 2 to 3
Utilisation and lead time move
The people left absorb the work. Utilisation rises past comfortable, morale drops, and delivery starts to slip in a way that is still deniable.
Month 3 to 5
Attrition follows
Staff attrition rises off the morale change. Replacing those people costs the hire cost and the onboarding months, both of which are in the ledger.
Month 4 to 8
Customers act, one segment at a time
Price led accounts move first. Mainstream accounts weigh the hassle. Anchored accounts complain and stay, which is why the revenue effect looks small until it is not.
Month 6 to 12
The competitor arrives
Competitors react on a lag and match part of your move once. Whether the saving survives that is usually the whole question, and it is the part a payroll calculation cannot reach.
Operator questions
Can I model a specific site or line?
Only as far as your files describe it. Departments, sites and product lines become things in the graph when documents mention them, with their own headcount and cost. There is no facility level detail beyond that, and the graph will show you exactly how thin the representation is.
How does it decide what a department was holding up?
From the relationships in the graph. A department attached to delivery carries capacity. One attached to customers carries service, which feeds satisfaction and then churn. Where the files do not say, it uses a default relationship and marks it as inferred, which you can then correct.
What is the smallest useful upload?
A profit and loss, a headcount list with costs, and a customer list. That is enough to run capacity, cash and churn honestly. Contracts and supplier agreements are the two additions that narrow the most bands.
Run the change before you write the paper
Take the consolidation, the second source or the layer removal, and see what month four looks like. Then write the paper with the bad case already in it.