For the people who approve the spend
Every business case arrives at ninety percent confident and none of them says why. ConditionFirst names the beliefs a plan rests on, prices them honestly, and tells you which one is worth paying to check — before the capital moves.
No account. No email. Nothing leaves your browser.
Here is a plan resting on six things. Every one of them is reasonable. Every one would pass a review without comment.
No modelling background required. If you can say what has to be true and roughly how sure you are, the instrument does the rest.
Write each belief as a claim someone could show is false. "Site managers adopt the new process after one training cycle" — not "adoption will be strong."
One slider, one tap: hard data, a comparable case, or gut feel. Your evidence sets how wide the range is held. Ninety-five percent on a feeling is carried as fifty-nine to ninety-nine, because that is what a feeling is worth.
Most assumptions degrade rather than destroy. Declaring which is which changes the answer more than any probability does — and it is the question nobody asks.
Two numbers you already know. They produce the sentence that gets a discovery phase approved.
Strip your numbers, send the sentences, get them rated independently. Where two people land twenty points apart, the belief usually isn't written clearly enough for them to be rating the same thing.
What was believed, by whom, on what date. When it goes wrong — and sometimes it will — you can show the reasoning was sound even though the draw wasn't.
No account, no trial, no seat count, no data leaving your machine. It runs entirely in your browser — you can save the page and use it on a plane. We would rather you use it and never pay us than not use it at all.
Open it nowWorks offline · Nothing stored on our servers · Export to JSON or print
ConditionFirst is aimed at the person who has to approve a plan they didn't write, and live with the call.
The method has one honest weakness: you are pricing your own beliefs, and nobody is well calibrated about a plan they are invested in. A Condition Brief is an independent pricing of the same conditions, delivered in five working days.
Priced on capital at risk because that is what the work protects. If a Brief tells you the decision is sound, you have bought a defensible record for less than a day of most people's consulting rates. If it tells you it isn't, it has paid for itself many times over.
It will not tell you the probability your plan succeeds. Nobody can compute that. What you get is the probability implied by your own stated beliefs, propagated correctly. If the beliefs are wrong, the answer is wrong — which is why the tool keeps a calibration log and scores you over time.
It will not replace judgement. It exposes judgement to scrutiny. That is the entire contribution, and it is a smaller claim than most tools in this space make.
It will not give you a number to hide behind. If your beliefs rest on gut feel, the ranges are held wide and the tool says so on the results page rather than quietly producing a confident-looking figure.
And it is new. ConditionFirst launched in 2026 and is being validated in the open. If you use it and it is wrong about something, we would genuinely rather hear that than not.
Yes, permanently, with no account and no email required. It is a single web page that runs in your browser — nothing you type is transmitted anywhere, and you can save the file and use it offline.
We make money from Condition Briefs and facilitated sessions, where the value is independent judgement rather than software. Charging for the instrument would put a toll booth in front of the thing we want people to do.
Primarily people who approve capital: appropriation reviewers, investment committee members, procurement leads assessing vendor business cases, grant and stage-gate panels. Anyone who must judge a plan they did not write and be accountable for the call.
It works just as well on your own plan — it is simply harder, because you are marking your own homework. That is what the blind rating feature is for.
No. You answer four plain questions per belief and move a slider. The arithmetic underneath uses a Monte Carlo simulation with a Gaussian copula for correlated beliefs, and you never have to know that — it is documented in the Methods tab for anyone who wants to check the working.
A risk register lists things that might go wrong and assigns them a colour. It does not tell you what the combination is worth, which risk is worth paying to eliminate, or whether your confidence has any track record behind it.
ConditionFirst starts from the opposite end: what has to go right, priced as a range, multiplied honestly, and ranked by what it would cost you to be wrong.
You send the business case or plan. We write out the conditions it rests on, price each one independently without seeing your figures, cite the base rates behind our pricing, and return a one-page summary plus a ranked and costed list of what is worth checking first.
If you have already priced the conditions yourself, we run it blind and include the disagreement analysis — which is usually the most useful part.
Published failure-rate literature — project overrun studies, adoption research, sales-cycle data. They are conservative reference points, not precision figures, and they are cited so you can check them.
They exist to make you justify a departure, not to overrule you. Being twenty points above the base rate may be entirely correct; you should just be able to say why out loud.
ConditionFirst is built by CIONAOD Inc. in Ottawa, Canada. The method joins Roger Martin's strategic cascade from Playing to Win to the calibration discipline in Annie Duke's Thinking in Bets — neither of which is ours, and both of which are worth reading directly.
One email, no form, no scheduling link. Say roughly what is at risk and what you are trying to decide, and you will get a straight answer about whether this helps — including if it doesn't.
hello@conditionfirst.com