Inclusive and Accessible Public-Service Benchmarking
Accessibility standards specify essential requirements for digital content and interfaces. Conformance matters, but it does not prove that a person can understand the service, verify identity, move between channels and receive the intended outcome without avoidable dependence on someone else.
Marketways benchmarks the complete service through technical tests, user experience and operating evidence. The combined assessment shows where access fails, whom the barrier affects and which redesign should receive priority.
Define the service outcome and the people who need it
An accessibility review needs a defined task, not an abstract user. Applying, paying, submitting evidence and appealing a decision create different demands.
Marketways identifies the required outcome, material user conditions and channels available for completion. We include permanent, temporary and situational barriers where they change the journey.
This scope prevents a successful test of one webpage from standing in for the full service. It also connects each barrier with a consequence that management can address.
Use standards to establish necessary requirements
Technical standards provide stable, testable criteria for digital products. Automated and expert tests can identify missing labels, keyboard failures, contrast problems and other defects that block assistive technology.
Marketways selects representative pages, functions and technologies and records conformance evidence consistently. We distinguish a requirement from optional guidance and avoid claiming broader accessibility than the tested scope supports.
Conformance establishes a necessary baseline. It does not reveal every problem that a person encounters while completing the service.
Observe people completing the actual task
Instructions may satisfy a technical standard and still remain difficult to understand. Authentication may also work with assistive technology while requiring information or timing that a user cannot provide.
Marketways observes intended users performing realistic tasks with the channels and assistive technologies they normally use. We examine comprehension, independence, repeated effort and the point at which a participant requests help or abandons the journey.
The research explains why a technically valid interaction fails in practice. It also reveals workarounds that operating data record only as ordinary assisted contact.
Follow the complete service across channels
Public services often move between a website, call centre, physical location and specialist review. A person may repeat information or lose support at the transition even when every channel performs well in isolation.
Marketways maps the journey from the initial need to confirmed completion. We link digital events, assisted-channel use, appointments, submissions, decisions and complaints where lawful and feasible.
The complete trace distinguishes channel accessibility from service accessibility. It also identifies a transfer that preserves access for one group but creates a new barrier for another.
Interpret differences without hiding small groups
Average completion can conceal a severe barrier affecting a smaller population. However, very small samples can also produce unstable rates that management may overinterpret.
Marketways combines quantitative outcomes with direct evidence and uncertainty. We report the observed difference, sample limitation and consequence instead of treating every numerical gap as a precise population estimate.
Low frequency does not make a material exclusion unimportant. Legal obligations and the severity of the barrier can require action regardless of the number observed in the study.
Prioritise redesign by consequence and obligation
Not every defect produces the same harm. A minor presentation inconsistency differs from a barrier that prevents identity verification, payment, submission or appeal.
Marketways compares severity, reach, obligation, available alternatives and feasibility. The resulting register names the owner, remedy, evidence needed to confirm correction and reassessment date.
Prioritisation makes the redesign manageable without turning uncommon needs into residual work. A barrier that blocks a consequential service can remain urgent even when it affects a small group.
Test AI-mediated access directly
Generative AI may explain complex content, translate instructions or support conversational access. It can also omit a qualification, misinterpret a dialect or produce an interaction that works poorly with assistive technology.
Marketways tests the actual AI service with intended users, relevant languages, assistive technologies and difficult cases. We compare its answer with the governing source and verify escalation when eligibility, rights or deadlines are involved.
Natural language does not make a service automatically accessible. The organisation must prove that the interaction preserves meaning, independence and a route to authorised help.
Measure independent and equitable completion
Management should measure successful completion, time, repeated attempts, assisted-channel use, technical failures and user confidence. We also examine whether a redesign improves one route while shifting effort to another.
Assistance is not inherently a failure. The benchmark distinguishes a chosen support option from dependence created by an inaccessible service.
Repeated benchmarking should confirm that corrections survive content, process and technology changes. Accessibility becomes an operating measure rather than a one-time release check.
How Marketways conducts the benchmark
Marketways defines the service tasks, maps the complete journey and combines standards review, accessibility testing, user research and operating evidence. We then prioritise redesign by consequence, reach, obligation and feasibility.
The engagement can combine Customer Satisfaction and Experience Research, Mystery Shopping and Service Quality Audit and Observation and Mystery Shopping. Where AI forms part of the service, Marketways also designs representative evaluation cases and human escalation.