Website Redesign Proposal Clarity Benchmark 2026

Methodology

• Freeze the sample before coding and retain one public source URL per observation.
• Code only buyer-visible disclosure on the sampled page; do not infer undisclosed contract terms from an agency brand, other page or private proposal.
• Treat a field as disclosed when the page gives a meaningful buyer-facing cue, not merely a keyword mention.
• Use binary coding for six fields: process, pricing/budget mechanics, timeline, ownership/handoff, revision/change control and post-launch support.
• Preserve row-level notes so ambiguous classifications can be challenged or recoded.
• Report simple counts and percentages only; do not rank providers or infer market-wide prevalence from this convenience sample.

Limitations

• The 16-page sample is small, English-language and search-discovered; it is not statistically representative of all agencies, countries or price bands.
• Public pages can change after the collection date. The downloadable dataset records the observation date, not a permanent truth about each provider.
• Binary coding compresses nuance: a brief mention and a detailed disclosure both count as present if they meet the coding rule.
• The study measures public proposal-clarity cues, not service quality, client satisfaction, delivery competence, ranking strength or commercial performance.

Exclusions

• Directories/marketplaces without a single provider disclosure model were excluded from the coded sample.
• Official standards/search documentation were used for context in other phases but are not provider observations here.
• Private proposals, sales calls, gated documents and information unavailable on the sampled public page were not inferred.
• Pages that were not clearly about website redesign delivery, procurement, cost, migration or handoff were outside this pilot.

Analysis logic

• For each field, sum TRUE values across the 16 observations.
• Divide each count by 16 to show a descriptive sample percentage.
• Use the lower-disclosure fields only as buyer-education opportunities; never label a sampled provider better or worse from this scorecard.
• Any future comparison across years must preserve or explicitly version the sample definition and coding rules.

Update frequency: Re-run annually or when a materially different research question/sample design is approved.