The principles
PSI rests on five observations about how major operational decisions actually play out. They are not abstract: each is something practitioners watch fail in real deals, real implementations, and real renewals, and each has support in the implementation research (see the evidence review). The formula that combines them is still a hypothesis being tested against outcomes.
Two questions, not one
Every major purchase asks two distinct questions that almost no one separates. Does this solution create value? And: will this organization actually capture the value it creates? Most evaluations conflate them. PSI measures the second.
Pain is real, and rarely honest
Buyers under-report change cost because under-reporting is what gets the deal across the finish line. Sellers under-price it because over-pricing it loses the deal. The truth (technical complexity, organizational disruption, risk exposure, and the burden of getting the team to actually use the thing) sits somewhere both parties have a small reason to hide. PSI surfaces it as four sub-scores, separately, on the record.
Adoption decides the outcome
PSI's working hypothesis is that whether value gets realized depends less on the feature set, the price, or even the strategic case than on whether people actually adopt the change: a champion with authority, a team with capacity. Time to value, training load, projected adoption, support burden, and champion strength form the Experience term. Together they can move the score up or down by as much as 50%.
Time has weight
A deal that hasn't measurably advanced in three months is a different deal than the one you scored on day one. Staleness is signal, not noise. PSI applies a time-decay factor so a dormant deal carries less commit weight, and a stalled deal never scores better for having stalled. If you wouldn't commit it cold today, you shouldn't commit it cold in your forecast.
Confidence is data
A score from a deal where the inputs are documented and a score from a deal where the inputs are guesses are not the same score, even if the numbers match. PSI carries confidence through to a Monte Carlo band (p10 to p90), so the reader sees both the central estimate and how soft it is. A wide band on a Strong score is a flag, not a victory.
How PSI is used
One instrument, three lanes. Sellers, buyers, and organizations all read the same number, but each has a different question they need it to answer.
For sellers
Pipeline discipline
Score every deal at every stage transition. Tighten the weakest input before commit. Let the top lever (reduce Pain, raise Gain, or improve Experience) be your coaching cue. Keep low-PSI deals out of commit, or commit them with eyes open.
Score a deal →For buyers
A reality check before signing
Before authorizing a major purchase, walk through the questions a vendor wouldn't ask. Are we really ready to absorb this? Is the champion strong enough? Has someone made the change-cost honest? PSI gives you a structured way to find out, and a number to remember when implementation gets harder than expected.
Self-discover →For organizations
A calibration corpus that gets smarter
Every score and every outcome enters a corpus that is used to retune the rubric over time. Scores are stored with their rubric version so they can be replayed under each new version, showing how calibration shifts. The instrument that scored the deal becomes the instrument that learns from it.
Read the methodology →Try it
Three ways to start.
01 · Lab
Try the instrument
Sliders and live math. Build intuition for what moves PSI.
02 · Score
Score a real deal
23 questions (28 for a past deal). In motion or historical. Behaviorally anchored.
03 · Discover
Self-discover as a buyer
25 questions, buyer perspective. About 10 to 15 minutes.
What this is not
PSI is a structured decision input. It is not investment advice, not a substitute for diligence, and not a replacement for human judgment. The score is a tool: useful when treated as one, dangerous when treated as truth.
Have your CFO, COO, or outside counsel validate financial assumptions before acting. The model is in active calibration; scores may be retroactively updated as the rubric tunes against real outcome data. See the methodology for the full math, and the evidence review for outside research that supports and challenges the idea.