Deterministic by design
The score, the priority ranking and every value estimate are computed by a deterministic engine: the same inputs always produce the same outputs. The LLM in Kompena is only ever used to write explanatory text grounded in numbers the engine already computed; it never invents a score or a value estimate.
Four categories, weighted and rolled up
Marketing, website, tracking and brand are each scored independently against segment-specific benchmarks, then combined into the overall score using documented weights, not hidden ones. Marketing is itself a roll-up: paid search, paid social, SEO and email are sub-disciplines inside it, never categories of their own. The weights also differ by business type, so a lead-gen business is not scored on an e-commerce curve.
Confidence is part of the answer
A score built from two connected sources carries a different confidence level than one built from ten. That confidence is shown alongside the score itself, not buried in a tooltip you'd never find.
Benchmarked, not absolute
Every metric is compared against your segment's priors before it's judged. This is why setting your segment correctly at onboarding matters: it's the difference between a fair comparison and a misleading one.