We have not moved price since 2023
Margin arrives in month one. The customer answer takes a year. The competitor answer lands in the middle, and for most companies it lands harder than the board expects.
Raise priceFor chief executives
The decision you have moved to next month three times
It is usually a price, a person, a plant or a product. Everyone in the room has a view, the views are incompatible, and none of them come with a number you can hold anybody to. Counterfirm turns that meeting into a run: the same decision against the same company doing nothing, four hundred times, with a list of which actor caused the damage.
Build a twinWhat it cannot tell you
No implementation. Upload the files you already send your board and ask the first question the same afternoon.
You are not short of opinions. You are short of a range.
The finance view and the sales view of a price rise are both defensible and they cannot both be right. What is missing is not another opinion, it is the shape of the outcome: how bad the bad case is, when it arrives, and which of the two views the answer actually depends on. That last part is measurable, and the software measures it.
The deferred five
The questions that sit on a chief executive for a year
Each one ships as a scenario with the levers already set, and each has its own page saying what the model does with it and where it is weakest.
One customer is too big and everybody knows it
Concentration is a slide until the month it is an event. The run makes it the event, including the reference that account was quietly giving your salespeople.
Lose the largest customerA competitor has started undercutting us
Hold, match or split the difference. Three runs, one baseline, one seed set, and a table that shows which of the three is still standing in month nine.
A competitor movesThere is a department nobody can defend
The cost line is the easy half. What that function was holding up shows itself in month three, in churn or in delivery, and the run says which.
Eliminate a departmentWe should be in the next market over
Entry cost is cash now against share later, with an incumbent who already lives there and reacts to you rather than standing still.
Enter a new marketSomeone wants us to buy them
You buy revenue and you inherit a company. The second half is where the money goes, and it is the half a model of the first half will not show you.
Acquire a companyWhat comes back
An answer with its own argument attached
A run returns a median with a tenth and a ninetieth percentile around it, the month by month path, and a mechanism section that names the agent responsible. Not "demand softened", but the segment that left, the month it left, the contract renewals that happened to fall in that window, and the competitor who matched forty five percent of your move in month four.
Under that sits the assumption ledger: every number the run stood on, where it came from, and how wide the band is. Industry defaults are listed first, because a default sitting under a seven figure decision is the single most useful thing the software can point at.
- The bad case stated as a number, not as a caveat
- The tripwire: the earliest month reality can tell you the model was wrong
- Sensitivity, measured rather than asserted, so you know which argument in the room actually matters
- The same seed tomorrow gives the same answer, so nobody can quietly re-run it into a better shape
Why the room disagrees
Everyone is modelling a different company
Your finance lead is modelling a business with no competitors in it. Your sales lead is modelling one where every customer behaves like the last one they spoke to. Both models are internally consistent. Neither has the other one in it.
Counterfirm builds all nine classes of actor into the same run: customers split into segments that behave differently, named accounts one at a time, competitors, suppliers, salespeople, executives, employees, investors, regulators and partners. When the answer turns, you can see who turned it.
What the first fortnight looks like
This is not a transformation programme. It is a folder, a question and an argument you can have properly.
- 1
Send it the pack you already produce
Management accounts, the customer list, the last strategy deck, your price list, the org chart. It reads CSV, Excel, Word, text and markdown directly, and a PDF if it has a text layer. It tells you plainly which files it could not use.
- 2
Read the readiness score before you read anything else
It is the share of the model standing on defaults rather than on your own record. A low score is not a failure, it is a to do list in priority order.
- 3
Run the question you have been deferring
Four hundred simulations, about a tenth of a second on the machine this was built on for a twelve month horizon with around forty agents. Fast enough that the follow up question costs nothing.
- 4
Argue with the ledger, not with each other
When your sales lead says the churn number is wrong, change it in front of them and re-run. That is a two minute loop, and it moves the meeting from position taking to evidence.
- 5
Take the brief to the board
The answer, the path, the mechanism, the bad case, the tripwires and the assumptions, with the shakiest ones at the top. A board that can see the assumptions argues about the right thing.
One question, six hundred and forty outcomes
What "it depends" looks like when you draw it
Honestly
What this will not do for you
It will not tell you what your market does next year. It has no news, no feed and no knowledge of your sector beyond what you uploaded and a table of industry starting points. If a regulation changes, a category collapses or a new entrant arrives from outside your industry, the model finds out when you tell it.
It will not make the decision. It narrows the range and names the mechanism, and then somebody still has to weigh the parts that were never in the files: what your people can absorb, what you promised a customer in a room, what kind of company you are trying to build.
It will not survive a bad brief. If you upload three files and ask about a product line that appears in none of them, you will get a confident answer about a company that is mostly industry defaults. The readiness score is there so that this is visible rather than flattering.
- No market intelligence. It knows your record, not your sector
- No answer to a question your files cannot see
- Confident output is not the same as correct output, which is why every run ships with its own list of weaknesses
Where it sits
Against the three things you do today
| Ask the team | Board pack model | Bring in advisers | Counterfirm | |
|---|---|---|---|---|
| Answer this week | yes | yes | no | yes |
| Shows a range, not one line | no | no | sometimes | yes |
| Competitors who react to you | no | no | yes | yes |
| Names the actor that caused the loss | no | no | sometimes | yes |
| Every number traced to a line in a file | no | sometimes | yes | yes |
| You can ask the next question without a meeting | no | sometimes | no | yes |
| Understands your market better than you do | no | no | sometimes | no |
| Carries accountability for the outcome | yes | yes | partly | no |
The last two rows are the honest ones. A simulation does not know your sector and it cannot be held responsible for a decision. What it can do is make the argument explicit before you commit.
What a chief executive usually asks second
How much of my time does this take?
The upload is one folder. Everything after that is reading. If somebody in finance owns the twin and keeps the files current, your own time is the half hour it takes to read a brief and the ten minutes it takes to challenge an assumption and watch it re-run.
My team will say the model is wrong.
Good. That is the productive version of the argument, because the ledger makes "the model is wrong" into "this specific number is wrong", which somebody can then go and check.
Every assumption is editable. If your sales lead sets churn where they think it belongs and the answer does not change, you have learned something about the decision. If it flips, you have found the thing worth spending a week on.
Can I put this in front of the board?
The brief is written to be put in front of a board: the answer, the path, the mechanism, the bad case, the tripwires and the assumptions. It names its own weak points, which tends to go down better than a deck that does not.
Is there a worked example I can look at first?
Yes. Harborline Components, an invented contract manufacturer with twenty four customers, 14.24 million dollars of revenue, thirty one percent gross margin, forty six people, a largest customer at eighteen percent of revenue and a price unchanged since 2023. It is invented, and it is labelled as invented everywhere it appears. Every question on this page is already set up on it.
Bring the decision you have deferred twice
One question and the files you already have. If you would rather see it work on somebody else's company first, the worked example is invented, public and set up with every scenario on this page.