Therapist network map — warm paths into the shortlist (2026-07-30)

Companion to 2026-07-30-therapist-shortlist.md (who to approach) and the operating half, process/06-outbound-call-sprints.md (how). This file answers one question: which of these prospects can be reached warm instead of cold, and who among them knows whom.

Why it matters: process/01-client-acquisition.md drops cold email (0.215% end-to-end, ~30–70 h/client — fatal against the ~15 h/wk ceiling) and makes referral the ignition channel, with seeds limited to berlein + aerias. A cold therapist list is therefore close to worthless as-is. Its value depends entirely on how much of it can be converted to warm — which is what follows.

Evidence discipline. Rows are tagged 🟢 VERIFIED (quoted from the practitioner's own site or a primary source), 🟡 PLAUSIBLE (structural — shared institution, town, or host; a hypothesis with a named cheap test), or 🔴 UNVERIFIED. Same university or same town is not evidence that two people know each other. Every 🟡 has a test that costs one question.

1. The two assets Daniel already owns

He has two HPCSA-registered practitioner clients. In a profession that runs on inter-practitioner referral, supervision groups, and shared rooms, that is the whole game.

Alexa Berlein — berleinpsych.com (client)

🟢 Clinical Psychologist, practising Rosebank, Cape Town (WC 7700). 🟢 Trained Stellenbosch University (undergrad) → UCT (Honours) → University of Pretoria (Masters). 🟢 Listed on Psychology Today ZA. 🟢 Scope: adolescents + adults; anxiety, depression, ADHD, trauma/PTSD, grief; medical psychology (chronic illness, amputee, pain).

Her three training institutions map onto three separate clusters in the shortlist:

Test, one question each: "Do you know any of these people — De Bruin, Miller, Green (Pretoria); Sadie (Stellenbosch); Grant (Claremont)?"

⚠️ Ethics line — do not cross. Ask Alexa about colleagues only. Never ask a psychologist about patients, and never ask her to introduce you to anyone she has a clinical relationship with. Colleague intros are ordinary professional courtesy; anything patient-adjacent is a POPIA and HPCSA problem and would cost the relationship.

Nadia Omidosu — nadiaoedpsych.co.za (client, invoice drafted)

🟢 Educational psychologist, Centurion — the exact sub-niche of the 10-strong ed-psych cluster. 🔴 Her affiliations, professional bodies, and supervision group are unresearched — and shouldn't be researched. Just ask her; she's a paying client mid-relationship and the answer is one WhatsApp message.

🟢 Shared-host coincidence: her site sits on 169.239.218.61, and three prospects sit on 169.239.217.x — the same provider range (see §3).

Secondary network nodes Daniel already has

2. ⭐ The single best network find: Therapy in Action

🟢 therapyinaction.co.za — multidisciplinary allied-health practice, Stellenbosch (opened 2010) + Somerset West (2013), serving Helderberg/Overberg.

🟢 Team (as published): 6 physiotherapists, 7 occupational therapists, 1 practice manager, 3 administrators, 2 clinical psychologists in association, plus dieticians, a speech therapist, an orthotist and wound-care nurses based at the practice. Associates include Elbert de Wit (psychologist) and Imke Oosthuizen (speech therapist).

🟢 Natalie Sadie (ranked lead #7) states she works "in association with Therapy in Action, Stellenbosch and Somerset West".

Why this is the highest-leverage node in the entire map:

• ~20 practitioners, one practice manager, two towns — a single relationship reaches the lot, and a practice manager is a named, reachable, non-clinical decision-maker who is not in session all day (§4 explains why that matters enormously).
• It is allied health — the same discipline as physioscpt, so a physioscpt win is directly transferable proof rather than a vertical jump.
• It plausibly contains or overlaps the Somerset West play-therapy greenfield cluster (Gerda Kriel, Lizelle Fryer, Elretha Bartlett, Heloise Breytenbach-Uys, Elana Rocher) — 🟡 test: check the practice's our team page against those five names.
• Group practices were deprioritised in July for having a diffuse decision-maker. That rule is right for a 2-partner practice; it is wrong here — 20 practitioners with a dedicated practice manager is a concentrated buyer, not a diffuse one.

Natalie Sadie is a referral hub in her own right

🟢 From her own site: "in association with Stellenbosch University since January 2019"; "in association with several schools, learning/tutoring/wellness centers, universities, clinics"; practised at the Centre for Student Counselling and Development (SU); affiliated with ER24, SADAG, Lifeline, Childline, Rape Crisis, GBV Command Centre.

She ranked weakest on Repaint pain (fairly modern site, gmail) — but she is the best-connected single individual on the list. Reframe her: not a build target, a doorway. Even at a small job or none, she is worth the conversation.

3. Prospect-to-prospect links — who knows whom

🟢 Verified structural links

⚠️ On "SA Medical Specialists" / samedicalwebsitedesign.co.za: a Cape Town agency specialising in medical-practice websites was already captured in the 2026-07-24 Firecrawl scan — 8-step process, SEO + copywriting, medical-niche positioning. Whether it is the same operator as Judy Grant's "SA Medical Specialists" is 🟡 unconfirmed — verify before acting. Either way it is a direct competitor in exactly this niche (like Kea Logic) and cuts two ways: a displacement incumbent, or a white-label partner (ICP-C) whose whole client base is Daniel's ICP-A. Do not pitch against it before confirming what it actually is.

🟡 Shared hosting — same provider, not proof of a shared webmaster

Same /24 means the practitioners bought from the same SA hosting reseller. It is weak-to-moderate evidence of a shared webmaster and must not be pitched as such.

🟢 Incumbent web suppliers (each = a displacement target or a partner)

Extracted from footer credits. One supplier holding several practices is one conversation instead of five.

Standing lesson: a stale copyright + a named agency = a relationship that has lapsed. Those are the softest displacement targets on the list (Olivier © 2015, Asnath © 2017), because nobody's feelings are attached any more.

4. The operational fact that shapes every call

Therapists are in 50-minute sessions all day. They cannot answer a phone, and a missed call from an unknown number reads as a patient trying to reach them — which is both intrusive and, for them, an ethical loose end.

Consequences, and they are not optional:

• WhatsApp-first, call-second. Already the standing rule for ICP-A in process/01 ("WhatsApp click-to-chat sits above the form"); for this vertical it is near-absolute. A WhatsApp message is read between sessions; a cold call is an intrusion.
• The reachable windows are the 10-minute inter-session gaps (on the hour / ten past), before ~09:00, 13:00–14:00, and after ~17:00.
• The practice manager / admin is the reachable human at any group practice — and is paid to take exactly this call. Therapy in Action has one; solo practitioners don't.
• Never leave clinical-sounding voicemail. Identify as a supplier immediately, so nobody wonders if it's an emergency.

⚠️ HPCSA advertising rules — verify before building or pitching. Health-professional advertising in SA is constrained (canvassing/touting, and patient testimonials in particular). This bears directly on delivery — a therapist site may not be able to use the review/testimonial social proof that a normal SMB site leans on — and on marketing, since a case study quoting a practitioner may need her sign-off on wording. 🔴 Unverified — confirm against the current HPCSA ethical rules before promising testimonial features or publishing client metrics. Cheapest route: ask Alexa and Nadia what they believe they're allowed to publish; they live under these rules.

5. Body-level channels (many practitioners at once)

6. What this changes about the shortlist

The shortlist ranked prospects on Repaint pain. Network position re-ranks them, and in two cases inverts the order:

The synthesis: run two tracks at once, not one. Broken sites justify a cold approach on their own merit (the defect is the intro). Everything else should wait for a warm path, because cold is a dropped channel here and the network work costs a couple of conversations. The sequencing is in process/06-outbound-call-sprints.md.

Honesty notes

• The Alexa cohort links are hypotheses, not findings. Shared universities and adjacent suburbs are structural — they raise the prior, nothing more. Each resolves with one question, which is exactly why Sprint 0 is one conversation rather than a research project.
• Nadia's network is entirely unresearched, deliberately. She's a live client; asking beats scraping.
• No LinkedIn / social-graph work was done. Alexa has a LinkedIn profile; mapping second-degree connections there would be the obvious next pass and would upgrade several 🟡 rows to 🟢.
• Firecrawl was down for this pass (see the shortlist's honesty notes), so Psychology Today ZA and TherapyRoute — the two richest sources for who practises alongside whom — were not harvested. Fixing it would materially improve this map.