◆ Sample: our own audit standard, run on a real brandGrowth audit · 17 July 2026

Crank 11

Prepared for Jack by Jack & Max at Crank 11. We went through the whole business: the site, the ads, the email, the numbers you shared. Everything worth acting on is written down here.

Book a walkthrough, we’ll take you through itOnly if you want it. The report stands on its own, and it’s yours either way.
kreators.aiREF · sample-kreators

THE VERDICT · READ THIS PAGE FIRST

Two things are capping you and the evidence does not yet separate them: the audit form was a dead end - every lead lost, and the site compiled itself in the browser, per visit.

Both scored the same severity on the evidence we have. Which one leads is the first thing to settle on the call.

What this is built on

  • 0ObservedRead directly from your public funnel, your ad library, or a connected account.
  • 0ReportedSupplied by you. We have not independently verified these.
  • 0InferredReasoned from category benchmarks. These are the lines to challenge first.
  • 1Unknownthe revenue model could not be built.

These describe the inputs this audit was built from, not a tag on each sentence.

What it is worth

Not modelled. We did not have enough of your numbers to show the working.

Contribution margin was not supplied, so this audit does not recommend a spend level.

The first move

The audit form was a dead end - every lead lost

What we still need from you

  1. After product cost, fulfilment, fees, refunds and shipping, what is left of a $100 order? Without it we will not recommend a spend level.
  2. Which numbers can you put in front of us live, so we can size this properly?

Conditional fit

We cannot yet name one constraint with confidence, so we cannot say a seat is worth it. That is the first thing the call settles.

If any of this is wrong, that is the most useful thing you can bring to the call. Mark it up.

BEFORE YOU READ IT

Two minutes on how to read this report: the thread through it, and the two sections to read if you read nothing else.

What’s in here
01

Executive verdict

A premium brand losing every lead it earns.

The design language is genuinely distinctive - and almost everything behind it was working against conversion. The site compiled itself in the visitor's browser on every visit, the audit form didn't submit anywhere, and the case-study section carried claims that couldn't be stood behind. The fixes are structural, not cosmetic - and each one is in the 30-day plan below.

The audit form was a dead end - every lead lost

critical

The final CTA was a plain anchor link. No capture, no storage, no email. 100% of quiz effort discarded.

The site compiled itself in the browser, per visit

critical

React development builds + Babel Standalone: multi-second first paint on mobile - fatal for paid traffic.

The chrome/Y2K design language is a real asset

working

Distinctive against a sea of template agencies. Keep it, make it fast.

Where the money goes today
Paid clicks lost to load time (projected)≈ 12–30% of ad spendIndustry LCP data: mobile abandonment rises steeply past 2.5s. A browser-compiled bundle put first render well past that on 4G - at £2,000/mo test spend, that's £240–£600/mo paying for visits that never see the page.
Leads lost to the unwired form100% of submissionsThe form's submit was an anchor to #thank-you. Every completed form = a lead paid for and thrown away. At any CPL, the waste is total.

This is the projected-waste view every URL-only audit receives - pre-launch, so projections rather than named account lines. Share ad-account detail and these become your numbers, not projections.

02

The CRO teardown

Also worth fixing

Client-side compilation on the critical path

critical

Every visit shipped React development builds plus a live JSX compiler (Babel Standalone) to the browser and compiled ~6 component files on the main thread before anything interactive appeared.

Index.html: react.development.js, babel.min.js script tags; ~10 JSX files compiled per visit.

Fix: Re-platform to a compiled framework with static generation. (Done: this report is served by the fix.)

Lead capture wired to nothing

critical

The 4-step audit form collected brand, spend, ROAS and email - then discarded all of it. The 'Book my audit' button was a hash link to a static thank-you screen.

Diamond-plus-pages.jsx:693 - <a href="#thank-you"> as the submit action.

Fix: Real API capture with storage, operator notification and confirmation email. (Done - this funnel.)

Case studies that couldn't be stood behind

critical

Named clients, revenue figures and ROAS multiples with no verifiable basis. One discovery call away from a credibility collapse - worse than no proof.

CaseAuroraDP: '$40K→$120K in 90 days'; pricing calculator: '+89% ROAS lift across 47 brands'.

Fix: Replaced with dogfood proof (this report) and designed slots that fill with real client results as they land.

Four font families, one doing the work

opportunity

Four Google Font families loaded from a third-party CDN; the display-sans was overridden to Inter Tight anyway, so Space Grotesk shipped dead weight on every visit.

Index.html fonts request: Inter Tight, Instrument Serif, Space Grotesk, JetBrains Mono.

Fix: Self-host two families + mono, subset, preload. (Done.)

Hash routing made every page unshareable

opportunity

#pricing and #methodology don't exist to crawlers or link previews - no page could rank, and every shared link landed on the homepage.

Window.location.hash router in index.html:67–110.

Fix: Real routes per page with per-route metadata. (Done.)

The design language itself

working

Conic-gradient chrome, the iridescent orb, the magnetic cursor - a coherent, memorable aesthetic that no template agency site has. This is brand equity; the job was to keep it while removing its performance cost.

Tokens.css chrome system; WebGL orb with custom shaders.

Fix: Preserved wholesale; the orb now loads after the page paints, from a static poster.

03

Who you’re up against

What’s working for each of them, and the one move that beats it.

Sample competitor - performance agency

Quantified waste call-outFree-audit applicationThe inputs you shared
  • The longest-running ad is a plain founder-to-camera piece - social proof of the format's durability.
  • Every ad links to an application-framed funnel: they filter hard for spend level before a call.
at least 34 active adsNew creative roughly every 9 days refresh

Sample competitor - creative subscription

Portfolio carouselPrice-anchoring vs agencies
  • Price transparency is their whole wedge - the ad names the monthly fee.
  • Monthly refresh cadence shows in frequency complaints in their comments.
at least 12 active adsMonthly batches refresh
04

Ad concepts, ready to shoot

5 concepts, written to shoot.

Each one is a complete brief: the angle, the headline, the primary text, the call to action and the art direction a designer or a photographer can work straight from. We generated a wider pool than this and scored every one against the same rubric before deciding which five you see. They are yours whatever you decide, and we will produce any of them with you.

5 concepts75.8 average scoreyours whatever you decide
The Artifact, concept preview88score
Deliverable-first proof

The Artifact

Seven Days, concept preview86score
Creative-fatigue clock

Seven Days

Six Weeks vs 48 Hours, concept preview76score
Agency-timeline betrayal

Six Weeks vs 48 Hours

The Diagnosis, concept preview72score
Quantified waste call-out

The Diagnosis

The Seal, concept preview57score
Guarantee-led risk reversal

The Seal

·

Creative-fatigue read

Honest-empty: no ad data sharedNo ad account was connected for this audit - so here's what this section looks like honestly empty. With account access (or even screenshots), it becomes a per-asset read: which ads are dying, how fast, and what the refresh cadence should be. For a brand at £15–30k/mo, DTC creative typically fatigues in 5–14 days; agencies shipping 10–20 assets a quarter cannot keep up with that arithmetic.
05

The plan

The first month in order of money moved, then the road past it. Nothing here waits on us. Every week-one item is something you could start tomorrow.

  1. Week 1

    Stop the waste

    • Re-platform to compiled, static-generated pages (kills the browser-compile tax)
    • Wire the audit form to real capture: storage + operator alert + confirmation email
    • Remove every unverifiable claim; swap in dogfood proof
  2. Week 2

    Make speed a weapon

    • Self-host 2 font families + mono; cut the dead family
    • Poster-first hero: WebGL orb loads after first paint
    • Real routes per page with metadata for link previews and indexing
  3. Week 3

    Launch the test

    • Ship the two concepts that scored 85+ ('The Artifact', 'Seven Days') as the prospecting pair
    • Hold 'The Diagnosis' for retargeting; rework 'Six Weeks' copy onto a stronger visual
    • Instrument every quiz step with a variant dimension for drop-off reads
  4. Week 4

    Compound

    • Abandoner recovery emails at T+1h and T+24h off the draft-save data
    • Weekly creative refresh cadence - winners scaled, losers killed on the score
    • First real client results begin replacing dogfood slots
06

Work with us

This took the engine an afternoon. Yours takes 24 hours.

Senior-directed, engine-built, honest about what it finds - including the losers.

Book a walkthrough call

Prepared by Crank 11. This report and the concepts in it remain yours whatever you decide.