Skip to content

Calculator

Risk calculator: conformance gap and complaint exposure

Set the sliders to your situation. The calculator returns two numbers: the conformance gap, meaning the share of assessed criteria that are violated, and the complaint exposure as a band with a traffic light. Both formulas are in the open, so you can recompute the number and argue with it.

Your situation

Out of the success criteria that can be assessed automatically or partly so. Criteria no machine can decide never enter the calculation.

Barriers that stop checkout, login or getting in touch outright. Beyond five the message does not change, so the term saturates there.

How much of it a purely automated check would find on its own. Missing alt text and weak contrast belong here; an illogical focus order usually does not.

Accessibility statement

Is there a reachable statement with a conformance claim, non-accessible content, a date and a feedback route?

Result

Conformance gap

30 %

Measured, not modelled: the share of assessed criteria that are violated. Even one percent means „not conformant“ for the criteria concerned.

Complaint exposure

52 to 68 %

Rating: high

Modelled from the four terms above. Published as a band, because a point value would suggest a precision the model does not have.

A model calculation on a disclosed formula, not a legal forecast and not legal advice. The conformance gap is measured; the complaint exposure is a setting whose weights are stated above.

Why there is no probability of a fine here

We could easily print a percentage for „this is how likely you are to pay a fine“. It would be invented. Three facts from the legal situation decide what the model may output instead:

  1. There is no error threshold. Each success criterion is met or not met, and one violated criterion means „not conformant“. The law knows no score at which it flips.

  2. A fine does not follow from a count of defects but from a procedure: market surveillance checks actively and on complaint, asks for remediation with a deadline, and only escalates if that is ignored.

  3. There are no publicly documented individual fines yet (as of July 2026). Without a base rate any probability of a fine would be guesswork. The real near-term risk is a complaint, a warning letter and pressure to remediate.

The weights, disclosed

The complaint exposure is a model calculation from four terms. The weights are a reasoned setting derived from how enforcement works, not an empirical estimate, because there is no data set to estimate from. That is exactly why they are here rather than in a black box.

DriverWeightWhy it is in the model
Base term, as soon as anything is violated0.10An authority screening automatically does not need a complaint to find a site.
Blockers in core journeys0.35A barrier that stops checkout, login or contact outright produces the complaint that triggers a reactive check.
Share of machine-detectable violations0.30Exactly the class an authority can establish at scale without human effort. Our rules layer finds the same cases.
Missing or incomplete accessibility statement0.25Independently actionable under § 37 numbers 2 to 6 and provable in seconds, so it is the first thing a check looks for.

The full formula including the bands is in docs/specs/risk.md in the source. If the legal situation changes or documented enforcement practice emerges, the recalibration is recorded there.

Request a free scan

Tell us your domain - we run a free initial scan and show you the most important barriers.