How scoring works
The investor path is the clearest walk-through. Hiring organisations, people consultancies and charities run the same pipeline with different configuration. Nothing here requires you to operate Yareta’s internals. This is how the product behaves.
1. Model what success looks like
Yareta, with you or from a pre-built benchmark, defines success for a role and a specific use case, for example a pre-seed CEO raising a first round. Reference models are built from validated constructs and, where relevant, from people whose outcomes are known (for example founders who went on to raise). The result is a Success Framework for that role and use case, covering skills, requirements, competencies, culture, motivation and traits, which can be adapted across situations.
You can run several frameworks at once (CEO, CFO, CTO, a sales-team framework) under one organisation. The engine is the same; the configuration changes. See Success Frameworks.
2. Connect and absorb evidence
Recommended entry: your CRM. For each person you’re evaluating, Yareta can ingest materials you already hold:
- Pitch decks and PDFs
- Financial models
- CVs
- Call transcripts (where you record and take notes through your CRM or a connected meeting tool)
From this evidence, Yareta builds a behavioural profile, inferring traits, competencies, culture and motivation from how people write, reason and communicate, rather than relying on a questionnaire alone. You can also add people and files manually in app.yareta.ai.
3. Optional historical calibration
Where you provide historical pipeline (for example founders you backed vs founders you passed on who later raised), Yareta identifies the behavioural profile behind each outcome: the traits and patterns that separated who succeeded from who didn’t. Our psychologists and your team refine the framework from there. This is a managed programme with Yareta, not something you configure in a hidden admin screen.
4. Outcomes make frameworks predictive
The engine is only as good as its outcome data. When people, and the role they were assessed against, are tagged with what actually happened (raised / did not raise; hired / succeeded / did not; team held / team broke down), those labels link constructs to results.
You share outcomes with Yareta as part of an ongoing programme, via your CRM/ATS integration or directly. Yareta recalibrates framework weightings toward what actually mattered, and each recalibration ships as a new framework version. You don’t need to run a data-science project yourself.
Once a framework’s predictive validity is proven against real outcomes, its score is promoted to a Predictive Number. Until then, it remains a Success Framework Score: a measure of fit, not a proven prediction.
5. Score new people in real time
Each new person entering the funnel is scored against the attached framework (via CRM, the app, Zapier or the API). A real-time management due diligence report typically includes:
- the Success Framework Score, or Predictive Number where the framework has earned one
- diligence flags on specific framework components
- value-creation drivers
- probing questions to ask to raise confidence
You can optionally release the Yareta Assessment (and Insights, where enabled). Each new interaction and each answered probe re-scores the person and can move the confidence band. This is an ongoing loop, not a one-shot PDF.
6. Third-party psychometrics
If you already hold reports from third-party tools such as Hogan, the platform can ingest them, interpret the results against your framework, and produce probing questions and focus areas: the first-pass interpretation a psychologist would typically do.
7. Confidence and non-destructive updates
Each piece of evidence is typed (CV, transcript, psychometric, notes and so on), given a contribution to confidence, and summed to Low / Medium / High. Thin files produce a lower-confidence score plus a checklist of what would raise it. A completed Yareta Assessment typically lifts confidence toward High.
When a report is regenerated, Yareta uses all sources plus the prior report. Findings are carried, enriched, added or retired, never silently deleted. Every point keeps its source provenance.
8. Engine boundary (what the numbers mean)
A deterministic scoring engine owns every number. It runs first, against fixed rubrics and your framework configuration. A language model never generates or adjusts a score. It only helps with text-to-construct inference and writing the narrative. Scorecards show the engine version and framework version, so any score can be traced to the exact framework it was produced under, including after recalibration.
Treat AI-written prose as explanation, not as a second score.
9. Human psychologist on request
If you want a psychologist in the loop, request one in the product. Yareta can place a psychologist to run and interpret the assessment on your behalf, on top of the automated report. The automated report remains decision support. Your named reviewer, or the psychologist you engaged, makes the people decision. See How decisions are made.
10. Two populations you can score
- Your own funnel: founders or candidates entering through your channels, CRM or the portal.
- Yareta’s assessed network (where your contract includes it): people already assessed by Yareta who have consented to be discoverable. Because any profile can be scored against any framework, Yareta can surface people who fit your framework, and you can request an introduction.
11. Team-level scoring
When several people from one team are scored, you see coverage, concentration and gaps against the framework, plus a view of team breakdown risk informed by outcome-labelled management teams. As more team outcomes accumulate, these build into Team Success Frameworks by function and company stage.
Post-investment, the same engine supports portfolio teams that are underperforming despite having raised, using transcripts and interviews alongside team psychometrics. The same approach applies to divisions (sales, marketing) and scales with company size.