How we measure results

The three kinds of evidence we use

• Measurable and repeatable: anyone can run the same tool against the same public URL and see a similar figure. Page speed scores and technical audits sit here.
• Client-reported: the client told us what they are seeing. We publish their wording, we do not convert it into a percentage, and we do not present it as a measured result.
• Arithmetic: a sum of published prices. Software savings comparisons sit here, and the inputs stay visible so you can change them.

Speed and technical scores

Speed figures come from Google's own Lighthouse and PageSpeed Insights run against the live public URL, on the same page anyone else can test. Scores move with the network, the device profile, and the day, so a single run is a snapshot rather than a fixed property of the site.

What that does not mean: a high score is not a ranking, and it is not revenue. It means the page loads quickly and cleanly for real visitors, which removes one common reason people leave before the page finishes.

Search and AI visibility

Search visibility is read from Google Search Console for the property itself, which reports impressions, clicks, and average position from Google's own data rather than an estimate.

Referrals from AI assistants are harder to attribute. Assistants often send a visitor without a referrer, and the visitor arrives having already decided. Where clients tell us that enquiries mention ChatGPT, Claude, Gemini, Perplexity, or Grok, we publish that as client-reported and we do not attach a number to it.

Software savings

The savings comparison is arithmetic on list prices. We total the published monthly price of the tools a business currently pays for, at the seat count they actually use, and compare it against the monthly cost of the platform that replaces them.

What it excludes: setup fees, per-transaction fees, contractor time spent maintaining integrations, and staff time lost to duplicate data entry. Those all sit on the current-stack side of the ledger, so leaving them out makes the comparison more conservative rather than less.

• Inputs are visible and editable in the savings calculator, so you can replace our assumptions with your own invoices.
• We use published list prices, not negotiated or promotional rates.
• We do not count savings from staff reductions. Consolidation is about output, not headcount.

Delivery timelines

The two to eight week range describes builds we have delivered, measured from the approved architecture to launch. Simple builds land at the short end, and enterprise platforms with extensive custom logic land at the long end.

The clock depends on you as well as us. Content, brand assets, and feedback turnaround are the most common reasons a build sits still, which is why the discovery meeting sets those expectations before anything is quoted.

What we will not publish

• Invented metrics, rounded up from a client's enthusiasm into a percentage.
• Revenue or lead figures a client has not explicitly approved for publication.
• Before and after numbers where the before was never measured.
• Screenshots of dashboards belonging to one client shown as though they represent typical results.