§ Measured results · August 2026 cycle

We publish our numbers. Ask for any receipt.

Every engagement runs the same instrument: 200 identical buyer questions across four AI platforms, Google and Bing search data, and dated deploys with clean before/after windows. The full measurement reports below are public — confounds named, gaps admitted. Across the current portfolio, AI share of voice rose 73% (7.7% → 13.3% mean across three clients, Jul 13 – Aug 24, 2026).

The standard

Three evidence grades — gated, aligned, directional — and no claim before its pre-registered checkpoint. Every figure on this page links to the live report it comes from.

Three clients, two measurement cycles.

A local services company, a retail CPG brand, and a clinic — measured on the same panel, reported monthly, July and August cycles both public.
Boss Hog Moving · local services
Named in 1 of 4 AI answers — and still climbing with zero new work

Six weeks of foundation work took AI answer share from ~0 to 21%. Then the real test: no new deploys June 29 – August 25, and share rose anyway, 21.0% → 26.5% on identical questions. Five customers have named ChatGPT on quote forms; August logged the first chatgpt.com referral sessions in the site's history.

Cartwheel · retail CPG
AI citations 21 → 146 per month

Schema releases and expert articles were followed by citations rising exactly where predicted: 21 in May to 146 in July, consolidating at roughly twice the spring floor in August (+107% per day on the conservative basis). Customer-question clicks rose 49%. The clicks land on retail shelves — where this brand actually sells — by design.

Joybody Physio · clinic
The control: without the foundation, nothing moves

A schema deploy silently failed — and metrics ran eight weeks flat. Fixed June 30: brand appearances +43% within 12 days. The v4 upgrade followed: brand-search clicks +58%, sustained 27 days, with AI-assistant visits at twice the pre-deploy baseline. AI recommendation share is honestly reported flat at 8.5% — no claim before the pre-registered week-12 checkpoint.

Reports · Boss Hog

August cycle →  ·  July cycle →

Reports · Cartwheel

August cycle →  ·  July cycle →

Reports · Joybody

August cycle →  ·  July cycle →

How the numbers earn their place.

Marketing case studies are easy. Ours are built like measurements, because our founders shipped this discipline at Fortune-1 scale — and because PE analysts read our decks.
Standard 01
Graded
How strong is the claim?

Three evidence grades, labeled on every figure.

Gated: a dated deploy with a clean before/after window. Aligned: event-aligned movement with confounds named. Directional: small-n movement in the predicted direction. A directional read is never dressed up as a gated one.

In practice
  • 200 identical questions, 4 AI platforms, per cycle
  • Google Search Console + Bing Webmaster data
  • Dated deploys; windows drawn before results are read
Standard 02
Controlled
Would it have happened anyway?

Two natural experiments anchor the causal story.

The control: Joybody's failed deploy gave us eight weeks of flatline, then +43% within 12 days of the fix — the foundation is the gate. The persistence test: Boss Hog gained +5.5 points of answer share across eight weeks with zero new deploys — the foundation compounds unattended.

In practice
  • Confounds named in every report ("Honestly…")
  • Seasonality and press windows excluded from trends
  • Corrections published when a read improves
Standard 03
Attributed
Did it become customers?

Conversion is claimed only where it's instrumented.

Boss Hog's intake form asks every lead how they found the company — it has caught five customers naming ChatGPT. Where instrumentation isn't proven yet, the portfolio report says "not aggregable" instead of inventing a number. That line is the product.

In practice
  • Self-reported attribution on intake forms
  • AI-referral session tracking in analytics
  • Retail/booking data requested before it's claimed
§

Every claim traces to a dated, reproducible measurement. Ask for any receipt.

§ Your baseline

The first report we'd write is yours.

The free audit reads your business the way these reports read our clients' — where you're findable, where you're invisible, and the score we'd re-measure against.