Documentation
Every lever
A scenario is nothing but an ordered list of levers, each with a value and the month it lands. There are twenty four of them in seven groups, and this page is all of them.
Fifteen ready made scenarios are just preset combinations of these. Nothing is hidden behind a template.
How to read the tables below
Every lever has a value and a start month. The start month is when it lands, not when its effect finishes arriving, and for most of these the gap between the two is the whole point. A price change lands in month one and reaches a contracted customer only when their term ends.
Units are percent, people, money, score, or a thing you choose such as a department or a named customer. Percent levers are relative to today, so 20 on price means twenty percent above the current price, not a price of twenty.
Ranges are hard limits in the interface. They are wide on purpose, because a model that refuses to consider a 60 percent cut cannot tell you why it would be a disaster. The wider you go the more the answer is leaning on the shape of the equations rather than on your data, and the brief says so when the band opens up.
Levers combine. Several can share a start month, and several can be stacked at different months to describe a plan rather than a single move. The raise capital template does exactly that: money in at month one, marketing and salespeople at month two.
Price
| Lever | Unit | Range | What it does |
|---|---|---|---|
| Change our price | percent | -60 to 200 | Applies to new and renewing customers. Contracted revenue moves only when its term ends. |
| Change price for new customers only | percent | -60 to 200 | Existing customers keep their price until they churn. |
| Allow sales to discount up to | percent | 0 to 60 | Reps use it when they are behind quota, which is exactly when it costs most. |
People
| Lever | Unit | Range | What it does |
|---|---|---|---|
| Add or remove salespeople | people | -100 to 200 | New reps carry hiring cost immediately and quota only after they ramp. |
| Add or remove staff | people | -2000 to 2000 | Delivery capacity and service quality follow headcount, with a lag. |
| Eliminate a department | department | You pick one from the org | Removes its cost and its contribution. Morale takes the hit across the whole company. |
| Cut headcount by | percent | 0 to 60 | Severance lands in the month of the cut, morale over the six after it. |
| Change pay | percent | -30 to 60 | Moves cost now and attrition with a lag. |
Supply
| Lever | Unit | Range | What it does |
|---|---|---|---|
| Supplier cost change | percent | -50 to 200 | How much of it you can pass on is its own assumption. |
| Switch to another supplier | percent | -40 to 40 | Cost change plus a disruption window on lead time and quality. |
| Move manufacturing or delivery | percent | -50 to 50 | Unit cost change, a one off capital cost, and months of disruption. |
Customers
| Lever | Unit | Range | What it does |
|---|---|---|---|
| Lose a named customer | customer | You pick one, or the largest | Their revenue goes, and so does the reference they were giving you. |
| Win a named account | money | Monthly revenue you set | Monthly revenue added, with the delivery load that comes with it. |
| Spend on keeping customers | money | Monthly spend you set | Monthly spend that buys satisfaction, which buys retention, slowly. |
Growth
| Lever | Unit | Range | What it does |
|---|---|---|---|
| Change marketing spend | percent | -100 to 400 | Buys pipeline with diminishing returns and a two month lag. |
| Launch a product | money | Build cost and window you set | Build cost per month for the build window, then a quality and demand lift if it lands. |
| Enter a new market | money | Entry cost and pool you set | Entry cost, a new pool of customers, and a standing start on brand. |
| Acquire a competitor | money | Price and target you set | Purchase price, revenue and customers gained, cost inherited, integration drag. |
| Invest in the product or service | money | Monthly spend you set | Monthly spend that raises perceived quality with a three month lag. |
Money
| Lever | Unit | Range | What it does |
|---|---|---|---|
| Raise capital | money | Amount you set | Cash in, and an investor whose patience is now on a clock. |
One lever, and the second half of its description is the part people forget. Money arriving creates an agent with expectations and a time horizon, and that agent starts asking questions when the numbers slip.
Market
| Lever | Unit | Range | What it does |
|---|---|---|---|
| A competitor moves first | percent | -60 to 60 | Their price changes and you decide whether to follow. |
| A competitor launches against you | score | 0 to 1, in steps of 0.05 | How much better their offer looks, on a nought to one scale. |
| Demand shock | percent | -80 to 80 | The whole market moves, not just you. |
| A rule changes | money | Monthly cost you set | A monthly compliance cost and a constraint on how fast you can move. |
These four are the ones where nothing you did changes and everything your customers are choosing between does. They are also the cheapest way to find out whether your plan only works if nobody else moves.
One lever, many values
When you do not know the number, sweep it
A scenario answers what happens at 20 percent. A sweep answers what the whole curve looks like, with the full Monte Carlo behind every point on it.
Twenty five settings at eighty replications is two thousand complete simulations, and what comes back is a response curve with a best point on it against whichever objective you chose. The interesting part is usually not the peak but how flat the curve is near the top, because a flat top means the exact number matters less than which side of it you are on.
- Any lever with a numeric range can be swept
- A second lever can be swept at the same time, which gives a grid rather than a curve
- Site settings cap the total number of runs in one sweep so nobody accidentally asks for a million simulations
Pull one and read what comes back
The brief names the agent that caused the damage, the month it happened, and how often it happened across the replications.