Matching methodology
Business Fit, explained honestly
PBN suggests partners with a transparent, rules-based method on structured profile data. Here is what the score means, what goes into it — and what it is not.
The short version
A relevance score, not a prediction
Business Fit is
- A 0–100 measure of how well another company’s declared profile complements your declared needs.
- Directional: your fit with them and their fit with you are calculated separately.
- Explained: every point traces back to profile data you can read.
- Deterministic: the same profiles give the same score.
Business Fit is not
- A probability or likelihood that a deal, contract or partnership will happen.
- A rating of quality, compliance or reputation.
- An endorsement by PBN of either company.
- Shown as a percentage — ever.
Reading a match
The anatomy of a Business Fit explanation
- 1
Score (0–100)
Relative relevance for this direction. Compare matches with each other — not with a percentage.
- 2
Evidence coverage
Good, Partial or Limited: how many of your stated requirements the other company has declared data for. Verification counts are shown separately.
- 3
Mutual fit
Shown when declared data is relevant in both directions — it does not mean the other side is interested.
- 4
Why they fit
Each reason points to declared profile data, with its evidence level.
- 5
Gaps to check
What is missing or undeclared — the questions to ask in your first call.
Business Fit
Example BiologicsExample Sterile CDMO
82/100
Why they fit
Needs ↔ offers: Sterile fill-finish and lyophilization — 2 of 2 declared needs covered Self-declared
Certification: EU-GMP certificate for the fill-finish site Verified
Territory: Manufacturing site in the EU, where the opportunity requires it Document provided
Gaps to check
- No FDA inspection history declared
- Batch-size range not stated
A relevance score from declared profile data — not a probability that a deal will happen.
What goes in
Six signals, weighted and published
For a pair A → B, each signal scores between 0 and 1. A signal only counts when you have a requirement in it. If you need something and the other company hasn’t declared it, that requirement scores zero — leaving data out can never raise a score.
Needs ↔ offers
weight 35How many of A’s declared needs (capabilities, services, technologies, dosage forms, collaboration types) B declares as offers.
Opportunity fit
weight 20For A’s open opportunities: the share of required terms B offers or holds, adjusted for territory.
Territory
weight 15Whether B is present where A needs presence — commercial reach, distribution, registrations, sites.
Regulatory readiness
weight 15Certifications A requires that B holds. Verified evidence counts more than self-declared.
Focus alignment
weight 10Overlap of therapeutic areas and product categories.
Commercial model
weight 5Whether the two company types typically work together (e.g. licensor ↔ distributor, biotech ↔ CDMO).
Weights shown are for methodology version fit-v1. When the method changes, the version changes and all matches are recomputed. Geography or company type alone can’t make a match: without overlap between your needs and their offers the score is capped at 25. A missing must-have requirement caps it at 49 and is listed as a blocking gap. Inactive, hidden and already-connected companies are filtered out of recommendations.
Principles
Built to be questioned
Direction matters
Coverage is shown, not hidden
Self-declared vs verified
Every claim carries its evidence level. For certifications, verified evidence weighs fully, a provided document slightly less, and a self-declaration less again.
Only public, structured data
Where AI fits in — and where it doesn’t
The score and its reasons are computed by deterministic rules, so they can be checked. Where an AI-written summary is offered, it is clearly labelled “AI-generated summary” and is written only from the explanation you can already see — it adds readability, not new claims.
Why there is no “deal probability” yet
Predicting outcomes honestly needs outcome data — which connections were accepted, which conversations led somewhere — collected with consent and validated over time. Until such a model is calibrated and tested, PBN will not show a likelihood figure. If one is added later, it will be shown separately from Business Fit, with its calibration date.
Get matches you can actually explain.
The more precisely you declare what you offer and what you seek, the more useful — and the more confident — your matches become.