What was verified, and what was dropped
This site carries the strictest honesty carve-out in the wave: every statutory fact is verified at ato.gov.au or asic.gov.au, or it is not published. This section documents both sides of that rule.
The rule worked to
No statutory figure, threshold, deadline or ASIC/ATO statistic appears on the page unless it was checked directly at ato.gov.au or asic.gov.au during this build — not at a law firm’s summary of it, not at a secondary commentary site quoting it. Where a secondary source and the primary source disagreed, the primary source won and the figure was rewritten to match it.
Verified directly at the source
- The three SBR eligibility conditions — total unsecured debt under $1,000,000 excluding employee entitlements; employee entitlements including superannuation paid; tax lodgements up to date — confirmed at ato.gov.au (Small business restructuring, tax professionals guidance) and asic.gov.au.
- The 21-day Director Penalty Notice window, and that appointing a restructuring practitioner within it is one of the recognised ways to have a non-lockdown penalty remitted — confirmed at ato.gov.au, Director penalties, which states the four remitting actions (pay in full, appoint an administrator, appoint a restructuring practitioner, or begin winding up) verbatim.
- 3,388 SBR appointments between 1 July 2022 and 31 December 2024, of which 2,820 reached a plan, and construction as 24% of external administration appointments in the first eight months of 2025–26 — confirmed at asic.gov.au, Report 810 (Review of small business restructuring process 2022–24).
- 9,307 companies entered external administration in the first eight months of FY2025–26, to 28 February 2026, down 1.3% year on year, a ratio of 0.40% of registered companies — confirmed at asic.gov.au, Corporate Insolvency Update Issue 39 (March 2026).
Corrected against the concept document
CONCEPTS-W56.md instructed: “9,300 companies entered external administration in the year to 28 February 2026, 76% above the previous five-year average.” That figure does not appear at asic.gov.au and the closest real statistic contradicts its framing.
ASIC’s Corporate Insolvency Update, Issue 39 (March 2026), states: “9,307 companies entered external administration during the first eight months of the 2025–26 financial year, down 1.3% from the 9,432 companies recorded for the same period in 2024–25.” That is eight months, not a year to 28 February 2026, and it is a 1.3% fall, not a 76% rise against a five-year average. No asic.gov.au page this build could find states a “76% above the five-year average” figure for any period ending February 2026.
Per the carve-out, the wrong figure was not published. The page instead states the real, sourced number — 9,307 companies, first eight months of 2025–26, to 28 February 2026 — without the unverifiable trend claim attached. Flag for the wave lead: this looks like a hand-typed or misremembered statistic rather than one read off the log, the same failure mode W52 and W55 both recorded for contrast ratios. Recommend the same fix: paste stats into the concept document from a fetched source, never from memory.
The concept
Clearpoint Restructuring advises directors of small companies that cannot pay what they owe, from a West Perth office, with appointments taken across the Perth metro. The page’s single job is to turn a frightened director’s ATO letter into one phone call today. It gives no advice and promises no outcome — it explains a process, states what the ATO and ASIC publish, and stops there.
Palette
Argued from the brief: red would read as cruel, navy would read as the bank doing the chasing. The accent is the outcome — a business that keeps trading — so it is eucalypt green. Clay carries the 21-day clock and validation states without shouting. All ratios pasted from contrast-w56.log, insolvency block.
#FAF8F4Page ground. Ink reads at 16.29:1.
#EFEBE3Card ground — the proof element, form, FAQ.
#1C1815 hover — Dark structural sections: stats, footer, Sapience band, sticky bar.#2F6B45The outcome. CTAs, the “met” checkmark, the under-threshold gauge fill.
#8C3A22The 21-day strip and validation errors. Paper/panel only — forbidden on the map, so it never reads as a place.
#57504A — Body text and secondary text.Type
Fraunces at 500/600, a restrained optical size, for headings — a warm serif carries gravity without the cold corporate-navy register this audience already associates with the bank chasing them. Public Sans for body, no monospace anywhere: this is a letter to one director, not a dashboard.
Techniques
No libraries. The debt gauge (scale, ticks, the $1,000,000 limit flag, the live debt bar and its clay overshoot, and the three condition marks) and the service-area map are hand-drawn inline SVG, built and updated from JavaScript rather than pre-authored, so every mark always matches the live model — there is no second drawing and no second source of truth. The favicon is an inline SVG data URI. Everything else is HTML and CSS — native input[type=range] for the slider, segmented buttons for the two toggles, and native <details> for the FAQ.
The three conditions — eight reachable states
One comparison (debt against the $1,000,000 SBR limit) and two booleans. Licensed by check-w56.mjs, insolvency block: the full 8-state truth table below, the threshold comparison at exactly $1,000,000, the excess arithmetic, and the 59-position slider span (50,000 to 1,500,000 in 25,000 steps).
| Debt ≤ $1m | Entitlements paid | Lodgements up to date | Card |
|---|---|---|---|
| No | – | – | Over the threshold by $X — SBR is not the door. Ring us about the others. |
| Yes | No | No | Two things to fix first: entitlements, then lodgements. |
| Yes | No | Yes | Entitlements including super must be paid before an appointment. |
| Yes | Yes | No | Lodgements first — your BAS is the cheapest thing on this list. |
| Yes | Yes | Yes | The give-away (below) |
The “over threshold” row collapses four reachable states (all combinations of the two booleans once debt exceeds $1m) into one message, because the excess arithmetic is identical regardless of the other two answers — that collapse is itself asserted in iv-states.mjs against all four cells of the truth table, not just the one shown in the spec.
Why the fully-qualified state gives revenue away
When a director meets all three conditions, the obvious page would say “you qualify — book now.” This one does not. It says a restructuring plan costs money, an ATO payment arrangement does not, and offers to say which is cheaper on the first free call. A restructuring practice telling its most qualified lead to try the free government option first is the least deniable trust signal this page can print — a template or a hungry competitor would never write that sentence, because it can cost a $12,500 engagement. It is the same move W52 and W54 both found converts hardest: publish the number, or in this case the option, that the trade normally keeps quiet, and say plainly that you are doing it.
Why this converts
The proof element answers the one question every director actually has — does SBR apply to me — in under thirty seconds, with no email address collected and no wall in front of the answer. The give-away line, where a fully-qualified visitor is told to ask about a cheaper ATO payment arrangement first, does more for trust than any testimonial could, because it is a sentence that can cost the business a sale. The 21-day DPN strip sits directly beside the headline as information rather than a countdown, because the honesty this audience needs from a stranger on the internet holding their finances is calm, not pressure. The enquiry form asks only for a name, a phone number, the company, a rough debt band and what has already arrived — five fields, because a frightened director abandons a long form and the practitioner can ask everything else on the call. Every FAQ answer is drawn straight from ato.gov.au or asic.gov.au rather than paraphrased, so a director who checks the source finds the page told them exactly the truth, which is the only thing that earns a call from someone in this position.
Local SEO we’d implement for the real business
This page ships no LocalBusiness, Review or AggregateRating structured data — fictional structured data in a real index is dishonest markup. This is the schema a real Clearpoint-shaped client would get, as a code sample only.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "LegalService",
"name": "Clearpoint Restructuring",
"image": "https://example.com/logo.png",
"telephone": "+61-8-XXXX-XXXX",
"priceRange": "$$",
"address": {
"@type": "PostalAddress",
"streetAddress": "[real street address]",
"addressLocality": "West Perth",
"addressRegion": "WA",
"postalCode": "6005",
"addressCountry": "AU"
},
"areaServed": { "@type": "City", "name": "Perth" },
"openingHoursSpecification": {
"@type": "OpeningHoursSpecification",
"dayOfWeek": ["Monday","Tuesday","Wednesday","Thursday","Friday"],
"opens": "08:00", "closes": "18:00"
},
"aggregateRating": {
"@type": "AggregateRating",
"ratingValue": "4.9",
"reviewCount": "37"
}
}
</script>
The aggregateRating block only ships once real, consented reviews exist — via Google Business Profile sync or a review platform’s own schema, never hand-typed. We would also register the practice on Google Business Profile against the “insolvency service” and “bankruptcy service” categories, build a citation set matching the exact NAP on this page, and target “director penalty notice Perth” and “small business restructuring WA” as the two highest-intent local queries.
Three iteration passes
What changed
The Sapience demo band shipped a first pass with the primary link styled .btn-dark — the same dark colour as its own section background, so the button was invisible as a shape and read as floating bold text. Caught by screenshotting the band at 375 rather than trusting the class name. Fixed to .btn-primary (green fill) so the two calls to action are visually distinct against the dark band, and the unused .btn-dark rule was deleted rather than left as dead CSS. Spacing, the services-grid price chip weight, and the FAQ disclosure marker (a plus that becomes a minus rather than a rotating chevron, cheaper and clearer) were reviewed at the same pass.
What changed
The condition checklist started as three lines of plain text. It now carries hand-drawn SVG check/cross badges (green filled circle with a check for “met”, an outlined circle with a cross for “not met”), toggled via setAttribute rather than the SVG-inert .hidden property, giving a second read: on the first pass a visitor reads the sentence, on the second they notice the badges tracking every toggle they touch. The debt gauge fill changes token — green under the $1,000,000 line, clay over it — so the “over threshold” state is legible from the colour alone before the card text is read, and the entrance animation eases the gauge in from zero on cold load only (wall-clock, cubic ease-out, never re-entered on resize).
What changed
Verified at 375/768/1440 with zero horizontal overflow; the h1, primary CTA and trust line all finish above the sticky bar’s effective fold. prefers-reduced-motion is honoured both in CSS and via a matchMedia check before the entrance animation starts — under reduced motion the gauge settles at its final value immediately rather than easing in, asserted in iv-hardening.mjs. The phone field is validated with a JavaScript regular expression only; no HTML pattern attribute is used, which sidesteps the whole v-flag compilation trap W52 and W54 both hit. Removed one ornament: an early draft of the gauge carried a second tick mark at the slider’s midpoint that added nothing the label didn’t already say, so it came out. All three passes re-ran the full 8-state truth table and the empty/junk/filled form drive against the live page, not against a description of it.
What changed
The wave-56 sweep failed drawn-subject-in-thumbnail: the debt gauge was a 400×54 sliver whose actual painted marks spanned 19px, because the page had already moved the real proof — the three-condition checklist — into DOM boxes and type. The gauge and the checklist are now one hand-drawn SVG: a debt scale with tick marks, the live debt bar against the $1,000,000 limit flag, the excess drawn in clay once the slider passes it, and the three conditions drawn as three marks that fill in as each is satisfied — all read from the same model.debt / entitlementsPaid / lodgementsUpToDate the result card already uses, so there is still exactly one source of truth. The green fill uses the markGreen token (the mark-licensed alias of green go), never green go itself, on a paper/panel ground — the same lesson the livestocktransport builder paid for this wave with its map pins. The old checklist markup is kept as a visually-hidden (.sr) list so screen readers still get the full sentence form; the drawing carries the visual read. Re-verified: qa-w56.mjs 56/56, check-w56.mjs 182/182, all eight proof states and the empty/filled form flow re-driven in headless Chrome, zero console errors, no overflow at 375/768/1440. The construction-share stat was also brought into line with the concept document’s correction — 24% of EXTERNAL ADMINISTRATION appointments in the first eight months of 2025–26, quoted verbatim from ASIC Corporate Insolvency Update Issue 39, not 27% and not an SBR-specific share. The first correction swapped the number but kept the wrong population label; the second fixed the label. Two ASIC datasets, two different denominators.
What would change for a real client
- Xero or MYOB ledger access (read-only), so the pre-insolvency options review and the safe-harbour test could reference actual aged payables and BAS lodgement history instead of the director’s recollection of them on the phone.
- A real intake CRM (the enquiry form currently only shows a local confirmation state) wired to whichever practice management tool the firm already runs, with the same five fields feeding a genuine callback queue.
- A verified Google Business Profile and review feed for the LocalBusiness schema above, once real client consent exists.
- A live ASIC/ATO figure refresh — the three context statistics on this page are current as of this build and would need a quarterly check against the same two sources, not a one-off.