The Situation
Catherine just completed her master's in counselling. Her classmates — a cohort of newly qualified counsellors — are now all certified practitioners. Word-of-mouth referrals (friends of friends) have started generating real client interest on the demand side.
This is a two-sided network that already exists organically: supply (qualified counsellors) and demand (clients arriving via social trust). The question is what, if anything, should be built on top of it — and for whom.
Framing constraint
Eric's rule: organic over mechanical, simple over complex. This is not a BetterHelp clone pitch. The question is: what is the minimum thing that makes this network more valuable for Catherine's classmates and their clients, given what already exists?
HK Counselling Market — Key Facts
HK Mental Health Patients
226K+
2015–16, growing 2–4%/yr
1
BetterHelp Price
$70–100
USD/week (global benchmark)
Private HK Session
HK$800–1,500
typical private rate
AMindset Supervision
HK$750–1,000
per session at counsellor's cost
2
HK has no dominant counsellor-client matching platform. Psychology Today covers HK but is US-centric. Open噏 exists for youth crisis chat (volunteer-staffed, 24h text). The formal hospital system is overwhelmed.3
The counselling profession in HK is unregulated — no statutory body controls who can call themselves a counsellor. This creates a trust problem: clients cannot easily verify credentials, and this is the primary reason word-of-mouth referrals dominate over cold directory searches.
Structural insight: trust is the bottleneck
Clients don't find counsellors via Google. They ask friends. The credential gap + stigma around mental health means referral chains are the dominant acquisition channel. Catherine's network is already the product — it is warm trust.
Competitive Landscape
| Player |
Model |
HK Presence |
Gap |
| BetterHelp |
Async + video, subscription |
Available, English-dominant |
Not Cantonese-native, no local trust signals |
| Open噏 |
24h volunteer text chat, youth focus |
HK-native |
Not professional therapy; volunteer-only; no booking |
| Psychology Today HK |
Directory listing |
Partial (HK section thin) |
Cold discovery, no warm referral path |
| Private practices |
In-person, referral-based |
Local but scattered |
No digital layer, no client-counsellor fit info |
| AMindset (Catherine's contract offer) |
Platform places counsellors, takes commission |
HK-local |
One-sided terms (counsellor bears cost + non-compete) |
The gap that exists
No one in HK has built a warm, trust-preserving way for people to find a counsellor through their social network + get matched based on fit. The closest analog globally is the Zocdoc model (verified practitioners, real reviews) — but that requires insurance integration and doesn't exist in HK. The gap is a Cantonese-native, trust-first, referral-augmented discovery layer.
Part I — GTM for Counsellors (Catherine's Classmates)
Catherine's cohort is freshly qualified, likely building a private practice or looking for placement. Their core needs as new practitioners:
- Getting their first clients (acquisition)
- Managing sessions, notes, and progress (practice ops)
- Getting supervision hours (clinical requirement post-graduation)
- Credibility signal to clients ("why should I trust you?")
- Billing and scheduling admin (low-value overhead)
GTM Option A — WhatsApp-based Referral Network (Lowest effort, highest leverage)
The word-of-mouth chain already exists. The friction is that referrals are informal — someone texts "my friend is a counsellor, want her number?" and the client has no way to evaluate fit before committing to a session.
What this looks like: Each counsellor in the cohort gets a clean 1-page profile (name, approach, what they specialise in, Cantonese/English, session format, approximate rate). A shared WhatsApp group or link tree acts as the referral landing. When someone asks Catherine "do you know a good counsellor for X?", she sends the link instead of a phone number.
Why this works now
Zero infrastructure needed. Catherine is already the trust anchor. The cohort already has warm social proof inside their friend groups. This is the Bob/Donna model applied to mental health — leverage owned access rather than building a platform.
GTM Option B — Donna-as-Intake (Medium effort, higher value)
Each counsellor gets a WhatsApp number backed by Donna. When a prospective client contacts the number, Donna handles the intake: what they're looking for, availability, what the counsellor specialises in, and schedules a free 15-min intro call.
This removes the awkward first contact — many people drop off between "considering therapy" and "actually booking" because they don't know what to say. Donna lowers that bar significantly.
Constraint
WhatsApp was banned for chatbots in Jan 2026. This requires the unofficial WhatsApp API path (same risk as Sourcy), or pivots to Telegram. If WA instability is a problem, Donna-on-Telegram with a WA fallback is safer.
GTM Option C — Cohort Collective (Highest upside, most effort)
Position the cohort as a named practice collective — e.g. "Clearwater Counselling" or similar — with a shared identity, shared booking page, and rotating availability. Clients book "a session" and get matched to the available counsellor best suited to their presenting issue.
This is the early Doctors Without Borders or co-working therapy model. It builds brand equity faster than individual cold starts and makes supervision easier (cohort can supervise each other informally).
Risk
Collective identity requires consensus from the whole cohort. If Catherine is the only motivated one, this dies immediately. Only viable if 3+ classmates are genuinely interested in co-building. Ask Catherine first.
Recommended GTM path
Start with Option A (no infrastructure, immediate value). Add Option B (Donna intake) as the second step if referral volume justifies it. Option C only if the cohort self-selects into it.
Part II — Product Features: What to Build
Two target audiences. Features ranked by effort-to-value ratio.
Audience 1: Counsellors (Catherine's classmates)
P1 · Counsellors
Smart Intake Form
A short structured form (5 questions) that captures presenting issue, preferred approach (CBT, person-centred, etc.), availability, and language. Client fills it before booking. Counsellor reviews before agreeing to session. Reduces wasted first sessions.
Effort: 1 day (Typeform or Notion form) · Value: high
P1 · Counsellors
Session Notes Template
Structured SOAP note template (Subjective, Objective, Assessment, Plan) plus session summary field. Helps new counsellors build clinical documentation habits from day one. Can be a simple Google Doc template or Notion database per client.
Effort: 2 hours · Value: high for clinical practice
P1 · Counsellors
Profile Page (Trust Signal)
One page per counsellor: photo, qualifications, modalities (CBT, DBT, etc.), what they do NOT treat, session format (online/in-person/both), language, rate range. Shareable link that replaces "here's my number." Verifiable credential badge if HKPCA-registered.
Effort: 3 hours (Notion + Super or simple HTML) · Value: high
P2 · Counsellors
Supervision Log
Track supervision hours automatically. Log date, supervisor name, case themes discussed, hours counted. Exportable for registration purposes. New counsellors often need documented hours for accreditation — this removes the manual spreadsheet.
Effort: 1 day (Notion DB or Airtable) · Value: medium-high
P2 · Counsellors
Waitlist Manager
When a counsellor is full, client goes on waitlist. When a slot opens, Donna or a simple form triggers notification. Reduces the "I'll get back to you" drop-off that kills nascent practices.
Effort: 2 days with Donna integration · Value: medium
P3 · Counsellors
Peer Case Consultation Group
Private WhatsApp group for the cohort to discuss cases anonymously (no client identifiers). Standard clinical supervision practice. Reduces isolation of solo practice, keeps clinical thinking sharp. Costs nothing to set up.
Effort: 0 (WhatsApp group exists) · Value: high for wellbeing
Audience 2: Public Clients
P1 · Clients
Fit Questionnaire → Match
5-question form: presenting concern, approach preference (structured vs open-ended), language, online/in-person, budget. Output is a ranked shortlist of the cohort's counsellors who match. The key insight: clients don't know what they need — the form teaches them and reduces mismatch anxiety.
Effort: 1 day (form + mapping logic) · Value: very high
P1 · Clients
Free 15-Min Intro Call
Structured as a product feature, not just politeness. Booking page with pre-call brief: "Here's what to expect. Here's what your counsellor will ask. Here's how to prepare." Reduces the anxiety of first contact — which is the single biggest conversion blocker in mental health.
Effort: Calendly link + 1-page prep guide · Value: high conversion
P2 · Clients
Anonymous Pre-Session Check-In
Before each session, client fills in a 3-question mood check (how are you feeling, what's your main concern this week, anything your counsellor should know). Counsellor sees it before the session starts. Warms up the session and removes the awkward "so how was your week?" opener.
Effort: 2 hours (form + email automation) · Value: high retention
P2 · Clients
Between-Session Resources
Curated by counsellor per client: breathing exercises, CBT worksheets, mood diary prompts. Sent via WhatsApp or email after sessions. Not a content library — counsellor selects 1-2 things that match what was discussed. Extends the therapeutic work between sessions.
Effort: Resource curation + template (3 hours) · Value: medium
P3 · Clients
Progress Tracking (Client-Facing)
Weekly mood log (1 question), session attendance tracking, and optional milestone notes ("I spoke up at work today"). Visible to client only. Gives clients a sense of momentum — critical for retention in talk therapy, which often feels slow. Uses validated PHQ-2 or GAD-2 screeners.
Effort: 3 days to build properly · Value: medium-high retention
P3 · Clients
Psychoeducation Content (Cantonese)
Short explainers (300–400 words each) on common topics: what is anxiety, what is CBT, how do I know if I need therapy. In Cantonese-register Chinese. Catherine's cohort can co-write based on what clients actually ask. Doubles as SEO and trust-building.
Effort: 2 hours per article · Value: long-tail acquisition
Critical Assessment — What Could Go Wrong
Reasons to pursue
- Supply already exists — no recruiting needed
- Demand already arriving via word-of-mouth
- Eric has direct access to Catherine (zero cold start)
- Mental health demand in HK is structurally growing
- No dominant local platform exists
- Donna infrastructure is already built (intake bot, WhatsApp)
- Low capital required to start (forms + profiles = days)
Reasons to be careful
- Catherine may not want to be the distribution channel
- Counselling is emotionally heavy — burnout risk if scaled poorly
- Regulatory void cuts both ways: easy to start, hard to differentiate
- Cohort may be heterogeneous in motivation (some just want jobs)
- Clinical liability: bad matching or bad intake = real harm risk
- Mental health platform graveyard is real (dozens of dead apps)
- Eric is already at 3 active projects — adding this risks overload
The most important unknown: does Catherine want this?
Everything above is irrelevant if Catherine doesn't want her social network and professional network merged into a product. She just finished a master's — she may want to practice, not build. The first conversation should be with her, not about her. Ask: "Do you and your classmates want help with any of this?" before building anything.
Failure Base Rates — What History Says
Mental health platform failures are well-documented. Key patterns:
| Platform |
Model |
Outcome |
Cause of death |
| RewardMe |
Digital loyalty + wellness |
Dead ($1.1M raised) |
Engagement collapse post-onboarding |
| Lunchclub |
AI professional matching |
Wound down 2024 |
Cold matching lacks warmth; no trust layer |
| Spring Health (earlier stage) |
Employer mental health platform |
Still alive — B2B only |
Survived by going B2B, not B2C |
| Lyra Health |
Employer EAP replacement |
Alive — B2B only |
Same pattern: B2B survivable, B2C brutal |
| Headspace standalone therapy |
Consumer therapy app |
Pivoted / scaled back |
Acquisition cost too high vs LTV |
Pattern: B2C mental health has near-100% failure rate as standalone platforms.
Survivors all went B2B (employer-sponsored) or became referral/directory tools rather than direct care platforms. The lesson is not "don't build" — it's "don't build a platform." Build a layer that makes the existing word-of-mouth network work better.
Recommended Sequence
| Week |
Action |
Effort |
Expected output |
| This week |
Ask Catherine: does she want any of this? Which classmates are interested? |
1 conversation |
Go/no-go signal |
| Week 2 |
Build profile pages for willing counsellors (3 hours per person) |
3–9 hours |
Shareable trust signal |
| Week 2 |
Build smart intake form + referral link tree |
1 day |
Warm referral path live |
| Week 3 |
Add Donna intake bot (Telegram) for each counsellor who wants it |
1–2 days |
Automated first contact |
| Month 2 |
Review: how many sessions booked? Client feedback? What's breaking? |
— |
Go/pivot/kill decision |
Kill condition: If after 8 weeks there are fewer than 5 sessions booked through the new system, shut down the infrastructure (keep the WhatsApp group). The referral chain doesn't need tech — it needs people who want to refer.
The Donna Angle — What Eric Can Actually Build
Eric's real leverage here is Donna, not a counselling platform. Donna already handles intake conversations (Sourcy, restaurant WhatsApp pilots). The counselling intake problem is identical:
- Client reaches out unsure what to say
- Donna asks structured questions (what brings you here, what have you tried, what matters to you in a therapist)
- Donna routes to the right counsellor based on responses
- Counsellor gets a brief before the intro call
This is not a mental health app. It is a scheduling and qualification bot with a warm tone — the same thing Donna already does. The clinical work stays entirely with Catherine and her classmates.
Commercial model (if this scales): Counsellors pay HK$300–500/month for Donna-managed intake + profile page. At 10 counsellors that's HK$3K–5K/month with near-zero marginal cost. Not a business on its own — but a proof of concept for Donna B2C professional services verticals.
Verdict — CONDITIONAL
Don't build a platform. Augment the existing referral network.
The supply and demand already exist. The only missing piece is a trust layer that makes referrals more reliable and first contact less terrifying. That's solvable in days, not months, using tools Eric already has.
The first step is not a product decision — it's a conversation with Catherine. She and her classmates are the product. If they're willing to participate, the GTM is already done (friends-of-friends referrals). If they're not, nothing Eric builds matters.
If Catherine is in: start with profile pages + intake form. Add Donna-on-Telegram as the second step. Kill if 5 sessions aren't booked within 8 weeks. The whole experiment costs 2–3 days of Eric's time and $0 in infrastructure.
If Catherine is not interested: drop it. Don't build a counselling platform for a market Catherine isn't in.
Sources
1. Wikipedia —
Health in Hong Kong.
Mental health patients 226K+ in 2015–16, growing 2–4%/yr; competitive education pressure cited as driver for youth anxiety.2. Internal — Catherine assist file (crm/reports/assist/catherine.md), 2026-02-11 session. AMindset contract terms: supervision at Catherine's cost HKD 750–1,000/session, mandatory 10hr/wk, 6-month non-compete.
3. Open噏 platform (openup.hk). 24-hour youth text counselling, volunteer-staffed, Cantonese-native; reviewed Feb 2026. Confirms gap in professional (non-volunteer) digital counselling for HK adults.
4. BetterHelp —
How BetterHelp Works.
Global benchmark: 30,000+ licensed therapists, $70–100/week, matched via intake questionnaire. 1.7M reviews, 4.9/5 rating. Not Cantonese-native.5. Wikipedia —
Online Therapy.
History, modalities, advantages (stigma reduction, accessibility, convenience) and disadvantages (tech reliability, credential verification, privacy). Research confirms online CBT as effective as in-person for most conditions.6. MoodGYM —
moodgym.com.au.
Evidence-based self-help CBT program, 1M+ users. Validates demand for structured digital mental health tools. 2016 meta-analysis (12 studies): significant reduction in depression + anxiety. Used as reference for between-session resource format.7. Memory — PCRM MEMORY.md, 2026-02-18. Restaurant loyalty platform graveyard: RewardMe ($1.1M dead), Mogl ($35M/13yr dead), Bink (£20M+ dead). Pattern: B2C engagement platforms without operational utility collapse. Applied to mental health vertical.
8. Memory — PCRM MEMORY.md, 2026-02-16. Knowledge marketplace failure base rates: Clubhouse, Lunchclub, Google Answers, Keen.com all dead. Only B2B expert networks (GLG, AlphaSights) survive. Applied to counselling platform risk assessment.
9. Psychology Today —
HK Therapist Directory.
Reviewed Feb 2026. HK section routes to US listings — confirms Psychology Today has no meaningful HK presence. Local directory gap confirmed.10. Memory — catherine.md, 2026-02-11. Word-of-mouth referral structure: Catherine's classmates all now qualified counsellors. Client demand arriving via friends-of-friends chain. Organic supply + demand already operating without infrastructure.