Physios CPT — evidence pack (cold Repaint lead)
Outbound · Physioscpt · biz/outbound/physioscpt/evidence-pack.md

Physios CPT — evidence pack (cold Repaint lead)

Prospect: Physios CPT, 33 Belvedere Road (cnr Wyndover), Claremont, Cape Town — physiotherapy practice, "Helping Cape Town Move Better". Team of four: Nicole (Tracy-Faulmann, co-founder), Geron (co-founder), Chamoze, Natalie. Site: physioscpt.co.za (Wix, built by incumbent agency "Your Best Friend" / bestfriendagency.com). Pipeline row: PROSPECTS.md · original evidence: 2026-07-20 reference scan. All findings below re-verified live 2026-07-24, ~19:30 SAST. Internal doc — receipts name the tech; client-facing language stays outcome-only.

1. Headline: the SSL defect is STILL LIVE — and now precisely characterised

The 2026-07-20 scan said the bare apex serves a *.jnb1.host-h.net parking cert. Verified today: true, and worse-shaped than it sounded. It is not "the apex is broken"; it is a DNS lottery that breaks the practice's front door for a fraction of every day's visitors.

Receipts (all observed 2026-07-24)

The apex has three A records — two Wix, one stale ex-host leftover:

physioscpt.co.za A → 41.203.16.63     ← xneelo shared hosting; rDNS = www449.jnb1.host-h.net
physioscpt.co.za A → 185.230.63.107   ← Wix
physioscpt.co.za A → 185.230.63.186   ← Wix
NS: ns10.wixdns.net / ns11.wixdns.net ← the zone is Wix-managed: the stale record sits in THEIR OWN Wix DNS panel

What each path serves:

  • Stale IP (41.203.16.63), HTTPS: certificate CN=*.jnb1.host-h.net — SAN does not match physioscpt.co.za → full-screen browser warning ("Your connection is not private"). A cert mismatch is not the kind of failure browsers silently retry on another address — the visitor just gets the warning.
  • Stale IP, plain HTTP: an old WordPress install still running on the xneelo box (Server: Apache, X-Redirect-By: WordPress) 301-redirects http://physioscpt.co.za/ → https://physioscpt.co.za/ on the same broken host — i.e. it actively funnels visitors into the certificate warning. Anyone who clicks through the warning lands on a bare page titled "Login" (leftover WordPress login), not the practice site.
  • Wix IPs: healthy. Valid Let's Encrypt cert CN=physioscpt.co.za (SAN covers apex and www; valid 2026-05-24 → 2026-08-22), 301 → https://www.physioscpt.co.za/.
  • www: correctly CNAMEd to Wix (cdn1.wixdns.net) — always works. This is why the team has probably never seen the bug: their own bookmarks and every link they share go to www.

Real-world incidence test (5 fresh visits to https://physioscpt.co.za/, 2026-07-24):

301 185.230.63.107  ✓ works
000 41.203.16.63    ✗ SSL: no alternative certificate subject name matches
000 41.203.16.63    ✗ (same)
000 41.203.16.63    ✗ (same)
000 41.203.16.63    ✗ (same)

4 of 5 failed. Because resolvers cache one record for the TTL, this isn't a per-visit ⅓ coin toss — entire networks get pinned to the broken address for a stretch, so a given patient's phone can be consistently broken while the practice's own office wifi is consistently fine.

Honest delta vs the 2026-07-20 scan

Someone has done half the migration correctly since (or the scan sampled only the bad IP): a valid apex-covering cert now exists on the Wix side and two of three apex records point at Wix. The entire remaining defect is one stale A record in their own Wix-managed DNS zone. That changes the opener's flavour: this is a one-stale-pointer, big-visible-symptom story — trivial to fix once someone notices it (delete one record in the Wix DNS panel), and a perfect demonstrate-on-their-own-phone moment. It also means the opener has a shelf life: any competent person who looks will fix it in minutes. Use it soon; re-verify the morning of any send (one command: curl -sI https://physioscpt.co.za/ a few times, or Safari on mobile data).

2. The rest of what's broken (secondary receipts)

  • Booking is punted to RecoMed. 3 homepage links go to recomed.co.za/physiotherapist/cape-town/nicole-tracy-faulmann/…. RecoMed is a per-booking-fee marketplace that lists their competitors on the same platform; the patient relationship, the rebooking loop and the data live with the middleman, not the practice. (RecoMed as a channel is fine — as the only funnel it's a tax.)
  • Search surface is near-zero. Homepage <title> is the Wix default "HOME | Physios Ct" (brand truncated/misspelled); no meta description (og:title is the same weak string); two H1s on the homepage ("About Physioscpt" + "Book Your Physiotherapy Session Online"); nav is 3 pages total (home / about-us / contact) — no per-service or condition pages for a practice with four practitioners and multiple specialities. For "physio Claremont"-class local queries there is almost nothing for Google to hold on to. (Not claimed: schema markup — can't be reliably detected via curl/static fetch; check in a browser before asserting anything.)
  • Page weight. Homepage HTML alone is 723 KB; the contact page is 994 KB — before a single image or script loads. 46 image references on the homepage; team photos ship as raw editing exports (Nicole Edit 2.png, Geron Edit.png, Chamoze Edit.png) — uncompressed PNG portraits with working filenames visible in production. Wix template ceiling, not a tuning knob.
  • No email on the homepage — admin@physioscpt.co.za appears only on the contact page. Phone 082 557 2311 is prominent; a WhatsApp start-chat widget exists (Wix third-party app).

3. What's worth keeping (say this in every conversation)

  • The brand is likeable. "Helping Cape Town Move Better" is a real tagline; the team bios are warm and personal (adventure-seeker, fitness-enthusiast, mental-health interest). Keep the voice and the people-forward content wholesale.
  • The WhatsApp instinct is right. They already put a WhatsApp chat button on the site — the correct SA patient channel. It just isn't wired into an owned enquiry/booking funnel.
  • The www path and Wix-side SSL are healthy. Whoever half-fixed the migration did the hard part; give them that.
  • RecoMed presence = distribution worth keeping as a secondary channel, plus social proof (Nicole's public profile/reviews).
  • Consistent NAP basics (address + phone visible), robots.txt sane.

4. What it's costing them — in patient-booking terms

Honest framing: we can't quantify their traffic without their numbers — that quantification is exactly what the Signal Audit sells. Per-defect, qualitatively:

  1. The apex lottery loses the highest-intent patients they have. Word-of-mouth and GP-referral patients — the ones told "look up physioscpt.co.za" — type the bare domain into a phone. On the losing networks they get a security warning that reads as scam/closed/unsafe, on a health practice where trust is the entire product. These are patients already sold on the practice, lost at the door, invisibly (no analytics event fires for a cert interstitial — the practice will never see these losses in any dashboard).
  2. Every RecoMed booking pays a marketplace fee and trains patients to book where competitors are listed. No owned rebooking loop, no follow-up capture — physio revenue is course-based (multi-session), so the rebooking loop is the money loop.
  3. Local search invisibility compounds monthly. With a default title, no description and a 3-page site, "physiotherapist Claremont / Cape Town" queries go to competitors and directories (including RecoMed itself — who then charges for the booking that lands back at the practice).
  4. ~1 MB HTML pages on mobile data = slow first paint for the phone-first patients everything above is funnelling.

5. The before/after angle (proof pattern, not a promise)

Daniel's proven play for exactly this shape of client — small SA practice, template site, unowned funnel — is the before/after rebuild preview on a private *.danielslater.dev subdomain: live examples include nadiapreview (ed-psych practice), mozguide (safari operator) and africansurfaris (surf-travel operator). The pattern: same content and people, rebuilt clean, fast, mobile-first, with the enquiry funnel owned by the practice (WhatsApp-first) — shown side-by-side with the live original so the owner judges the difference in ten seconds.

Decision already made (01-client-acquisition §6): physioscpt is a cold lead → NO on-spec preview. They go through the ladder — free 2-minute look → paid Signal Audit → Repaint, [scope on call]. In outreach, existing previews may be shown as the pattern ("here's a before/after of a practice rebuild I did recently") — never promise a physioscpt preview that doesn't exist, and pick the showable pair carefully (see strategy.md — prefer mozguide/africansurfaris; hold nadiapreview back, it's another prospect's brand).

Displacement note for every artefact: they pay an incumbent agency ("Your Best Friend", credited in the footer). The stale record straddles a host migration and may predate the incumbent — assign zero blame anywhere, in writing or on calls. Full sensitivities in strategy.md.

Appendix — reproduction commands (for re-verification before sending)

dig +short physioscpt.co.za A                     # expect 3 records incl. 41.203.16.63 while defect lives
dig +short -x 41.203.16.63                        # www449.jnb1.host-h.net
echo | openssl s_client -connect 41.203.16.63:443 -servername physioscpt.co.za 2>/dev/null \
  | openssl x509 -noout -subject                  # CN=*.jnb1.host-h.net
for i in 1 2 3 4 5; do curl -s -o /dev/null -w '%{http_code} %{remote_ip}\n' \
  --max-time 10 https://physioscpt.co.za/; done   # the lottery

Verified 2026-07-24. If dig no longer returns 41.203.16.63, the defect has been fixed — the opener changes (see strategy.md §"If the defect gets fixed").