Research
How often is the internet right about companies?
The same evidence that produces each Score can answer wider questions: which facts go wrong, where errors start, how long changes take to spread. Findings are published here with their method, sample and date. Nothing has been published yet.
- Now Does the Score discriminate?The first question, before any other: whether real companies spread across the scale or all land in the nineties. Scored by hand, claim by claim, on a first cohort of Danish B2B software companies. If the Score does not separate companies, the methodology changes before anything else is built.
- Now CalibrationWhether a representation rated 80% likely to be accurate turns out accurate about 80% of the time. Built from independently verified claims, including a random panel of unclaimed companies, and published with its sample size.
- Next Which company facts go wrongAccuracy by claim type: leadership, founders, category, products, pricing, ownership. Where errors start, and which sources carry them.
- Next After a rebrandHow long a new name, category or leadership takes to propagate across registries, databases, search and AI, and which sources never update.
- Next After an acquisitionHow ownership, product futures and positioning are represented in the months after a deal.
- Later Digital Reality IndexCohort benchmarks by industry, geography and company size, published annually with methodology version and sample sizes.
- Later Source reliabilityWhich sources are most often right, most often copied, and most often the origin of an error.
How research is published
- Every figure carries its sample size, date, market and methodology version.
- Findings describe what sources say. There are no “worst companies” lists.
- Companies are named only where their Score is public, and never on the basis of private answers.
- Results are published when they are unfavourable to the method, too.
Take part
Companies can claim their profile and verify facts, which is what makes calibration possible. Agencies, investors and researchers who want to score a portfolio or a cohort can email [email protected].