A scenario
Margin first, then the decision about whether to pass it on, then the customer answer to a decision you only made because somebody else made one. This is the only scenario on the list that starts with a letter you did not send.
Run this on your own numbersAll scenarios
Ships as a scenario: one lever, supplier_cost_pct at 15, landing in month one.
Mechanics
The lever moves one number, the supplier index, and that number multiplies your cost of goods every month from then on. Everything else in this scenario is the company reacting to it.
The supplier index is multiplied by one plus the percentage times the share of your cost of goods that comes from suppliers. It does not know that only one supplier moved. On the invented Harborline record, supplier spend is about 68 percent of cost of goods and the single sourced billet is about 59 percent of that spend, so a 15 percent rise from that one supplier is about 9 percent blended. Enter the blended number, or the run will overstate the damage by two thirds.
Cost of goods is accounts times revenue per account times one minus gross margin times the supplier index. Entered at face value on Harborline, 15 percent across a 0.68 share lifts the index 10.2 percent and takes about 83,000 dollars a month out of a company making roughly 74,000 dollars a month of operating profit. Entered at the blended 9 percent it is about 50,000. Same lever, very different meeting.
The new price applies to everything you buy from the month it lands. There is no stock at the old price working through, no hedge, no fixed price agreement running to its end date. If you have one of those, delay the lever to the month it expires rather than modelling the protection.
Every month the supplier index also drifts by a small normal draw centred just above zero, which is where the ordinary, noisy input inflation lives. It is small next to the shock and it is the reason two replications of the same scenario do not have identical cost lines.
Once the supplier index is above 1.05, and if price is below any regulatory cap, the finance lead passes part of it on. The rise applied is the index increase times the pass through assumption, which defaults to 45 percent. On the Harborline example at an index of 1.102, that is a price rise of about 4.6 percent on both the list price and the new business price. This is on a six month cooldown, so it happens once in a normal horizon.
A pass through is a price rise, so it inherits everything from that scenario: contracted revenue reprices only at renewal, competitors match a share of the move once after their lag, named accounts jump to the new price at their own renewal month, and partners may drift if the price index goes above 1.1.
Fifty five percent of the rise, at the default, is simply absorbed. That is a permanent reduction in gross profit, and it is the part that reaches runway, the investor and, if it goes far enough, the finance lead's own cost cutting rule. A supplier increase is the scenario most likely to end in a headcount decision that nobody connected to the supplier.
The ecosystem
The supplier is the agent that starts this and the least active one in the run. The interesting behaviour is on your side of the invoice.
Objectives: hold their own margin as their costs move at weight 0.5, keep your volume at 0.3, be paid on time at 0.2. In this scenario their decision is imposed rather than simulated, because you are asking what it costs you, not whether they will do it.
Objectives: protect margin at 0.45, protect cash at 0.4, keep the forecast credible at 0.15. Owns the pass through decision and owns the cost cutting rule underneath it. Both are on cooldowns.
Objective: pay as little as the job allows. They do not know your input costs moved. They see a price change and respond with the same elasticity they would to any other one.
Objective: renew at a price they can defend internally at weight 0.35. A cost driven increase is easier to defend than an opportunistic one, and the model does not know the difference. Treat the named account losses in this scenario as a worst case.
Objective: take share when you give them an opening at 0.45. They match a share of your pass through, once, after their lag. If they buy from the same supplier they would be raising prices too, and the model does not know that either.
Objective: keep price defensible in the field at weight 0.2, which is the objective a cost driven rise puts under most pressure. With discount authority granted, three months under quota turns into field discounting that gives the pass through straight back.
The ledger
| Assumption | What it does here | Why to check it first |
|---|---|---|
| Share of cost from suppliers | Multiplies the percentage you entered before it reaches the index | The single biggest lever on the size of this answer, and easy to read off a purchase ledger |
| Share of a cost rise you can pass on | Sets how much of it reaches customers and how much stays in your margin | Defaults to 45 percent. Your last increase is better evidence than any prior |
| Gross margin | Turns the index move into money | A thin margin business is levered to this scenario and a fat one barely notices it |
| Price elasticity of demand | Decides what the pass through costs you in volume | The reason passing it all on is usually not the answer |
| Revenue under contract and months left | Decide how slowly the pass through actually reaches revenue | A cost rise that lands immediately against a price rise that lands over a year is a cash flow problem as well as a margin one |
| Supplier concentration | Nothing. It is computed into the ledger and the simulation does not read it | It is there to tell you whether the blended number above is the right one to enter |
| Supply lead time | Nothing. Also in the ledger and not read by the engine | Named here so you do not assume the model is using it |
Two of the rows in this table exist in the ledger and do not reach the simulation. They are listed because a model that quietly ignores a number you entered is worse than one that says so.
Honestly
The margin arithmetic here is the most reliable thing in the engine. Everything around the arithmetic is thinner than the decision needs.
The pass through question
The rise weighted by that supplier's share of your total supplier spend. Fifteen percent from a supplier who is 59 percent of your spend is about 9 percent.
The supplier concentration figure in the ledger is there to help you work it out, even though the simulation does not read it. If you do not know the split, the purchase ledger will tell you in an hour and it is the highest value hour in this scenario.
Run it three ways: the default pass through, none of it, and all of it, against the same baseline on the same seeds. The difference between the three is the actual decision. Passing none of it is not the safe option, because a permanent margin reduction reaches the runway rule and the runway rule cuts people.
There is a separate lever for it. It moves the cost in either direction and adds a disruption window on quality and delivery while the new supplier learns your business, with the drag lifting when the window closes. Run both and compare. A cheaper supplier who costs you three months of on time delivery is not obviously cheaper.
Because a permanent margin reduction eats runway, and below a threshold set by the finance lead's own risk appetite they take out between 4 and 20 percent of headcount and 30 percent of marketing. That is not in the scenario you wrote. It is what this company does at that runway, and it is the most useful thing the run tells you.
Next
A single sourced input is a concentration risk in the same way one large customer is, and the graph shows both. This scenario puts a number and a band on one of them.