Screenshots from the tools that recorded the result, each with what it is evidence of written beside it. No redrawn charts, no numbers without a source.
Why this page exists
So each one here is paired with the claim it supports and the window it covers. The case studies explain the reasoning and the decisions; this page is where you check the working. Read them together: a result without a method is luck, and a method without a result is a pitch.
Read the reasoningSearch Console and Analytics for organic, the ad platforms for paid, Lighthouse and the live build for development. Select one to narrow the set.
Showing 8 of 8 exhibits
Non-brand impressions grew through a full seasonal cycle, not a spike.
Clicks and impressions over sixteen months. The shape is the point: the line holds through the off-season instead of falling back to where it started.
Sample data, not a client account
The traffic arrives on category terms, not on the brand name.
Average position per query. Query text is redacted at capture. The position column is the evidence, and the client's keyword set is theirs.
Sample data, not a client account
Users grew year on year with no paid push behind it.
Solid line is the current period, dotted is the same months a year earlier. Organic sessions only.
Sample data, not a client account
Cost per result fell while spend went up.
Four campaigns, one account. The two right-hand columns moving in opposite directions is what matters: scale usually costs efficiency.
Sample data, not a client account
Conversion volume climbed as the cost line went the other way.
Bars are conversions per week, the orange line is cost per conversion. Campaign names redacted at capture.
Sample data, not a client account
The rebuild moved the score people actually feel.
Lighthouse mobile performance, measured on the same templates before and after. LCP and CLS are the two a visitor notices.
Sample data, not a client account
The build, running, at the address it lives at.
A URL in the chrome bar does what a caption cannot: it says the work shipped and is still standing.
Sample data, not a client account
Four usable concepts from one product brief.
Generated stills, not captures of a product UI, so these need no bezel and go straight into the standard gallery frame.
Sample data, not a client account
Captured, never recreated
Every exhibit is a screenshot of the tool that recorded the result. We do not redraw a chart into our own colours. Restyling reporting is the point at which it stops being evidence.
Published only with consent
Analytics and ad accounts hold a client's commercial data. Nothing appears here without that client agreeing to the specific screenshot, and account identifiers are redacted in the file before export.
The claim never outruns the capture
If a number is not visible in the exhibit above it, we do not assert it in the copy beside it. A claim the screenshot cannot support invites the reader to check, and then to find the gap.
Tell us which number matters and how you currently measure it. We will tell you what we would expect to move, and what evidence you should hold us to.