Internal application · September 2026

Applying behavioral pretesting to this website.

Using our own public website to examine how clearly we communicate the work, support our claims, and invite a relevant conversation.

This is an internal implementation case. The design process combined external research, alternative copy and layouts, synthetic critique, and browser checks. It does not establish an increase in inquiries or sales.

The decision

The previous site presented a broad practice through long service lists and uneven visual hierarchy. The redesign needed to help businesses, governments, and institutions identify relevant capabilities and find supporting work without implying a narrow sector boundary.

What we compared

• Copy. A previous opening and three alternatives emphasizing capabilities, opportunities, or decisions.
• Visual presentation. Three rendered layouts with the same copy, portrait, and evidence, examined on desktop and mobile.
• Buying context. Service-specific propositions considered against business and institutional decision scenarios.

What changed

The implementation introduces a clearer opening, grouped capabilities, more visible evidence routes, a consistent visual system, and direct contact options. Responsive checks cover the site's public HTML pages; keyboard entry and contact interactions were also checked.

What the testing can tell us

Synthetic role scenarios helped identify plausible interpretation problems and competing design preferences. They were not real participants, a representative population, or a forecast of behavior. We reviewed the findings critically and rejected interpretations that were unsupported by the page or its evidence.

The next evidence comes from real use: whether visitors understand the offer, find relevant work, and begin qualified conversations. No measured conversion improvement is claimed.

Research behind the approach

The work drew on website credibility research, accessibility guidance, maintained design systems, current practitioner analysis, and the limitations of synthetic user research.