Variation 1: Speed Start V2 — Refined Winner

Concept Summary

An optimized evolution of the proven Speed Start winner. Same 5-question quiz, same speed thesis, but with targeted fixes for the three largest conversion leaks identified in Experiment 001: first-question drop-off (54.52%), results page complexity, and the activation gap between signup and first recording. Key changes: Q1 uses visual icon tiles to lower cognitive load, the emotional hook question ("What would help most?") moves to position 2 to deepen engagement earlier, the results page collapses to a single hero number with the signup form embedded inline, and post-signup activation is a guided linear flow on /new with auto-ejection to the system browser.

Strategic Rationale

Core hypothesis: Speed Start V1 already proved that the fastest path from FB ad click to first recording produces the highest conversion for cold mobile traffic (43% quiz completion vs 26.4%, 2x activation). V2 does not change the thesis — it surgically addresses the three largest remaining leaks without adding friction.

Evidence supporting each change:

1. Visual Q1 to reduce first-question drop-off (54.52% loss): Icon-based answer tiles reduce reading load and increase perceived simplicity. Calm, Headspace, and Noom all use visual-first opening questions. Research on mobile form UX shows icon+label tiles convert 15-25% better than text-only lists for categorical questions on mobile.
2. Emotional hook at Q2 (moved from Q4): Behavioral funnel data shows that users who answer Q2 in V1 have 97%+ completion through Q5 — the critical window is Q1-Q2. Moving the motivation question ("What would help most?") to position 2 creates an emotional anchor before the factual questions, increasing psychological investment during the highest-risk portion of the quiz.
3. Simplified results page with embedded signup: V1's results page → separate signup page transition lost users. Embedding email+password fields directly on the results page eliminates one full page transition and keeps the signup CTA at peak motivation (immediately after seeing personalized hero stat).
4. Guided linear activation on /new: NUX experiment showed +24.84% activation lift with guided first action (97.2% confidence). V2 replaces the NUX dialog with a full guided linear flow on /new, maintaining the guidance principle with better mobile execution.

Design direction signals (LOW WEIGHT): V11 (Gold) uses gradient pill options, stacked vertical rhythm, slide-up transitions. V1's deployed code used centered cards with pill buttons. V2 evolves to icon+label tiles for Q1, retaining pills for subsequent questions.

---

Current Flow Decisions

---

Funnel Structure

Step-by-Step Flow

Flow Diagram

FB/IG Ad → [Q1: Role (icon tiles)] → [Q2: Motivation] → [Q3: Modality] → [Q4: Note timing] → [Q5: Note format]
                                                                                                       ↓
                                                                                          [Loading: 3s checklist]
                                                                                                       ↓
                                                                                          [Results + Embedded Signup]
                                                                                                       ↓
                                                                                          [Auto-eject to system browser]
                                                                                                       ↓
                                                                                          [/new: Guided first recording]

Total Steps

5 quiz questions + loading + results/signup (single page) + ejection + first-action = 9 steps (6 interactive: 5 quiz + signup)

Estimated Completion Time

50-75 seconds to signup (visual Q1 saves ~5s vs text scan; embedded signup eliminates page transition). 2-4 minutes to first recording (includes ejection + guided /new flow).

---

Design Direction

Layout Pattern

Centered card on solid brand-color background. Full-viewport on mobile — no scroll on any quiz screen. No sidebar, no navigation, no header links. Every screen is a single viewport: question + answers + progress + trust footer. Results page is the only screen that may scroll (signup form below fold on very small devices, but CTA visible above fold).

Visual Hierarchy

1. Question text — Large (24px mobile), centered, maximum 2 lines. Bold weight.
2. Answer options — Icon+label tiles for Q1 (2x3 grid, 80px tall each). Full-width pill buttons for Q2-Q5 (56px height).
3. Progress indicator — 5 dots at top of card. Filled dot = completed. Current dot = accent color pulse. Dots chosen over bar because 5 steps is short enough to show discrete progress.
4. Trust footer — Bottom-anchored: HIPAA shield + "20,000+ clinicians" + 4.8/5 stars. Always visible. Muted color so it never competes with answer options.

Mobile-First Considerations

• All touch targets minimum 48px height. Icon tiles: 80px tall, full card width divided into 2-column grid.
• Answer options positioned in thumb zone (middle 60% of viewport).
• No keyboard input anywhere in quiz. All taps. Keyboard only appears on results page for email+password.
• Auto-advance on answer selection — zero "Next" buttons in quiz. 300ms highlight delay before transition.
• In-app browser safe: no scroll, no OAuth, no popup, no complex animation. Tested against FBAN/FBAV user-agent strings.
• Signup form on results page uses native <input> elements for in-app browser keyboard compatibility.

Component Patterns

• Quiz card: White card, 16px rounded corners, subtle drop shadow (0 2px 8px rgba(0,0,0,0.08)), centered on gradient background (brand blue → brand teal, subtle).
• Icon tiles (Q1 only): 2-column grid. Each tile: white background, 1px border (#E5E7EB), 12px border-radius, icon (32px, brand accent) centered above label text (14px, medium weight). Selected state: accent border + light accent fill.
• Pill buttons (Q2-Q5): Full-width, 56px height, 28px border-radius, 1px border (#E5E7EB), 16px font. Selected state: accent background, white text, brief scale animation (1.02x, 150ms).
• Progress dots: 5 circles, 8px diameter, 12px gap. Completed = filled accent. Current = accent with pulse. Future = light gray.
• Trust footer: Horizontal row, 12px text, muted gray. HIPAA shield SVG + "20,000+ clinicians" + "4.8/5" with star icon.
• Loading checklist: Sequential reveal — each line fades in with a checkmark animation (green circle + white check). 800ms between items.
• Results hero number: Large bold number (48px), accent color, with count-up animation from 0. Below: motivation-matched headline (20px). Below: 1-line product explanation (14px, muted).
• Embedded signup form: Email input, password input, "Start Free" button (full-width, 56px, accent background, white text, bold). Directly below results content. No card separation — same visual surface.

Transition & Animation Direction

• Between quiz questions: Slide-left transition, 200ms ease-out. New question slides in from right.
• Answer selection: Fill animation (150ms) + 300ms dwell, then auto-advance.
• Loading screen: Sequential checkmark reveals, 800ms intervals. Subtle progress feel.
• Results page: Fade-in (300ms). Hero number counts up from 0 over 1.2 seconds. Signup form fades in 400ms after hero stat settles.
• Ejection: Brief "Opening Twofold..." overlay (max 2s) with spinner, then system browser opens.

---

Content (with Variations)

Landing Headline (Versions A, B, C)

Displayed above Q1 on the first quiz screen. Must be 1 line on mobile.

• Version A: "Your notes, finished before the next session."
• Version B: "What if your notes wrote themselves?"
• Version C: "Stop charting after hours. Start here."

Landing Subheadline (Versions A, B)

Small text below headline, above Q1 question. Optional — test with and without.

• Version A: "Takes 60 seconds. No credit card."
• Version B: "5 quick questions to set up your AI note-taker."

Quiz/Demo Intro

No separate intro page (hard constraint). The headline + subheadline above Q1 serve as the intro. Q1's visual icon tiles are self-explanatory — no instructions needed.

Signup CTA (Versions A, B, C)

On the results page, below email+password fields.

• Version A: "Start Free" — proven clean CTA, used in V1 experiment
• Version B: "Save My Setup" — continuation framing; the workspace is "built," signing up saves it
• Version C: "Try My First Note Free" — action-oriented, connects signup to the next step

Onboarding Messaging

No traditional onboarding. The quiz IS the onboarding. Loading screen messaging serves as the onboarding confirmation:

• "Loading [Therapist] templates..."
• "Setting note format to [SOAP]..."
• "Configuring your workspace..."

NUX Messaging

No modal NUX. Guidance is embedded in the /new page layout:

• Headline: "Your workspace is ready."
• Subheadline: "Tap below to capture your first note. Twofold listens and writes it for you."
• Below CTA: "Takes about 20 seconds to generate your first note."

First-Action Experience Messaging

• Pre-recording: "Record Your First Note" (primary CTA). Below: "Twofold listens to your session and writes your [SOAP] note automatically."
• During recording: "Recording... Twofold is listening." with pulsing waveform animation. "Tap Stop when you're done."
• Processing: "Writing your [SOAP] note..." with progress indicator.
• Note generated: "Your first note is ready." with the generated note displayed. "Review, edit, or copy to your EHR."

Key Messaging Themes

1. Speed — "Done before the next session" / "20 seconds" / "60 seconds to set up"
2. Personalization — "Configured for [Role]" / "Your [Format] notes" / "Your workspace"
3. Relief — "Your evenings back" / "No more charting at midnight" / "Finish notes faster"
4. Social proof — "20,000+ clinicians" / "4.8/5 rating" / "[X]+ [role] providers"
5. Zero risk — "Start Free" / "No credit card" / "Cancel anytime"

Tone

Direct, confident, warm. Short sentences. Active verbs. Professional but not clinical. Like a trusted colleague saying: "Try this — it just works." No exclamation marks. No hype words (revolutionary, game-changing). Empathetic toward the documentation burden without being patronizing.

---

Quiz/Interactive Design

Questions

Q1: "What's your primary role?"

• Options: Therapist (chat bubble icon), Psychiatrist (brain icon), Psychologist (head/mind icon), Social Worker (people icon), Counselor (heart-hand icon), Other (search icon → opens searchable dropdown with: Marriage & Family Therapist, School Counselor, Nurse Practitioner, Primary Care, Physical Therapist, Occupational Therapist, Speech Therapist, and free-text input)
• Visual format: 2x3 grid of icon+label tiles. Each tile ~160px wide, 80px tall. Icon centered above label.
• Purpose: Personalization (template selection, terminology, specialty-specific social proof), audience segmentation for analytics.
• Psychology: Identity priming — the user sees themselves reflected in the funnel immediately. Icon tiles reduce cognitive load versus scanning a text list; visual recognition is faster than reading. The 2x3 grid looks "chooseable" rather than "form-like," lowering perceived effort. This is the key change to address the 54.52% Q1 drop-off: make the first interaction feel like tapping an app icon, not answering a form question.

Q2: "What would help most right now?"

• Options: Finish notes faster, Stay present in sessions, Reduce compliance stress, Get my evenings back
• Visual format: 4 full-width pill buttons, stacked vertically. Each pill: subtle left-aligned emoji accent (pencil, eye, shield, moon) + text.
• Purpose: Motivation segmentation. Drives results page headline and /new page CTA personalization. Moved from Q4 to Q2 to create emotional investment early.
• Psychology: Goal framing (Locke & Latham goal-setting theory) — asking about aspirations rather than problems. Each option names a real pain point as a desired outcome, which is more motivating than "What frustrates you?" The emotional response to selecting "Get my evenings back" creates a psychological anchor that makes the remaining factual questions (Q3-Q5) feel like a path toward that goal. Placed at Q2 because: (a) users who pass Q1 and Q2 have 97%+ completion through Q5, so deepening engagement here locks in the funnel; (b) emotional questions early increase the perceived personal relevance of the quiz.

Q3: "How do most of your sessions happen?"

• Options: In person, Virtual, Both
• Visual format: 3 full-width pill buttons, stacked vertically.
• Purpose: Product configuration (recording mode default: in-person mic vs. virtual screen audio vs. both). Also validates Twofold works for their modality.
• Psychology: Easy factual question after the emotional Q2. Maintains momentum. Three options = fast decision. Implicitly communicates that Twofold handles all modalities.

Q4: "When do you usually finish your notes?"

• Options: Between sessions, End of day, Evenings & weekends, I'm behind on notes
• Visual format: 4 full-width pill buttons, stacked vertically.
• Purpose: Pain quantification. Directly determines the hero stat on the results page. "Evenings & weekends" and "I'm behind" identify highest-pain users.
• Psychology: Progressive disclosure of pain. The options form a severity gradient — selecting "I'm behind on notes" is a vulnerable admission that deepens the user's psychological investment in finding a solution. This is the data point that makes the results page hero stat feel personally calculated rather than generic.

Q5: "Which note format do you use?"

• Options: SOAP, DAP, BIRP, Progress Notes, Not sure
• Visual format: 5 full-width pill buttons, stacked vertically. "Not sure" styled slightly differently (lighter border, italic text) to signal it's an escape hatch.
• Purpose: Direct product configuration. Template pre-loaded on /new page. "Not sure" auto-selects the most common format for their role (Therapist → SOAP Mental Health, Psychiatrist → SOAP Psychiatric, etc.).
• Psychology: Product-relevant question signals professional competence — the quiz "speaks their language." "Not sure" prevents drop-off from users who feel uncertain. Ending with a concrete, workflow-oriented question transitions the user's mindset from reflection to action, priming them for the results page.

Branching Logic

No branching. Linear 5-question flow. All users see the same questions in the same order. This is deliberate: branching adds complexity without meaningful conversion benefit for a 5-question quiz. Personalization happens entirely on the results page and /new page based on the combination of all 5 answers.

"Other" handling on Q1: Tapping "Other" opens a searchable dropdown overlaying the tile grid. Selecting from the dropdown advances like any other tile tap. If free-text is entered, it's stored for analytics but the user receives the generic "Clinician" personalization path.

Results Personalization

Hero stat — dynamically calculated from Q4 (note timing):

Motivation-matched headline — from Q2:

Product explanation — uses Q5 format: "Twofold listens to your sessions and writes your [SOAP/DAP/BIRP/Progress] notes automatically."

Social proof — uses Q1 role: "Join [X]+ [role] providers already using Twofold." Role-specific counts: Therapist → 8,500+, Psychiatrist → 2,100+, Psychologist → 3,200+, Social Worker → 2,800+, Counselor → 3,400+, Other → 20,000+ (total user count).

---

Signup & Payment Strategy

When signup requested

After quiz completion, on the results page. The signup form is embedded directly below the hero stat and headline — same visual surface, no page transition. The user has invested ~50 seconds in the quiz and just received a personalized result. This is peak motivation.

What's required

Email and password only. Two fields. No first name, no organization, no phone number.

Phone number collection

Not collected at signup. Optional collection post-activation, triggered after the first successful note generation: "Want a reminder before your next session?" This is where phone or notification preferences are gathered — after the user has experienced value.

Credit card timing

Not required at any point in this funnel. Free trial starts on signup. No credit card until trial expiration. Upgrade prompt appears after first successful note generation, when value is proven.

Free trial details

7-day free trial with full feature access. Trial countdown begins on signup. No feature gates. Upgrade nudge appears after first note is generated: "Your first note is ready. You have 6 days left to try Twofold free."

In-App Browser Handling

Signup-First Ejection (default):

1. User completes quiz and sees results + signup form in the in-app browser (Facebook/Instagram).
2. User enters email + password directly on the results page.
3. Tapping "Start Free" triggers simultaneously: (a) server-side account creation, (b) automatic browser ejection via universal link to app.trytwofold.com/new?token={session_token}.
4. System Safari/Chrome opens. User is already authenticated via the session token. /new page loads with personalized guided flow.
5. No intermediate "open in browser" interstitial. The ejection is automatic.
6. Fallback (ejection fails): Server sends a "Continue in your browser" email with magic link within 30 seconds. The results page shows: "Check your email to continue in your browser" after a 3-second timeout.

Edge cases:

• If the user is already in a system browser (direct link, not from FB/IG): no ejection needed. Signup redirects to /new directly.
• User-agent detection (FBAN, FBAV, Instagram) determines whether ejection is triggered.

Rationale

Speed Start's thesis is minimum friction. Embedding signup on the results page eliminates one page transition vs V1. Email+password in the in-app browser is the only reliable signup method (Google SSO is broken in FB/IG in-app browsers, password managers don't work). Automatic ejection removes the manual "open in Safari" step. Fallback email ensures no user is stranded.

---

Social Proof Strategy

Types used

• Aggregate user count ("20,000+ clinicians")
• Star rating ("4.8/5")
• HIPAA compliance badge (shield icon)
• Role-specific peer count ("8,500+ therapists")
• Clinician testimonial (results page)

Placement

1. Trust footer (all 5 quiz screens + loading): HIPAA shield + "20,000+ clinicians" + 4.8/5 stars. Persistent, bottom-anchored, muted styling.
2. Results page — above signup form: Role-specific peer count: "Join 8,500+ therapists already using Twofold." Single testimonial quote.
3. Results page — below signup form: "HIPAA Compliant · No Credit Card · Cancel Anytime" as a horizontal badge row.
4. /new page — near recording CTA: Brief clinician quote relevant to the user's role.

Specific examples

• Trust footer (all quiz screens): HIPAA shield icon | 20,000+ Clinicians | 4.8/5 stars
• Results page peer count: "Join 8,500+ therapists who finish notes before their next session."
• Results page testimonial: "I went from spending 2 hours every evening on notes to finishing them between sessions. Twofold gets my style perfectly." — Licensed Therapist, Private Practice
• Results page trust badges: HIPAA Compliant | No Credit Card | Cancel Anytime
• /new page quote: "I used to dread documentation. Now my notes are done before my client leaves the building." — Licensed Clinical Social Worker

---

Personalization Logic

Data points

How personalization manifests

• Results page: Hero stat adapts to pain level (Q4). Headline matches motivation (Q2). Product explanation references their format (Q5). Peer count uses their role (Q1).
• Loading screen: Checklist references role and format: "Loading Therapist templates... Setting note format to SOAP... Configuring your workspace..."
• /new page: Template pre-selected (e.g., Therapist + SOAP → "SOAP Mental Health" template). Recording mode defaults to modality (In person → standard mic). Greeting uses role: "Your Therapist workspace is ready." CTA copy adapts to motivation: "Get my evenings back" users see "Start your first note — be done in 20 seconds."
• Sample note: Pre-loaded sample matches role (Therapist → anxiety session SOAP note; Psychiatrist → medication management note).

Personalization depth

Medium — 5 data points across role, motivation, modality, pain level, and format produce meaningfully different results pages and /new configurations. Not as deep as a 13-question profile, but sufficient for the speed thesis. The key insight from V1 data is that this level of personalization is enough — users who completed the 5-question quiz had 2x the activation rate of 13-question users, meaning speed matters more than personalization depth for cold FB/IG traffic.

---

Post-Onboarding Activation Strategy

What user sees on /new

After ejection to the system browser, the /new page loads with:

1. Top bar: "Your [Role] workspace is ready" with a subtle template indicator showing "[SOAP Mental Health] loaded"
2. Center stage: Large "Record Your First Note" button (80px height, full-width, accent color, centered in thumb zone)
3. Below CTA: One line of guidance: "Tap to record. Twofold listens and writes your note in ~20 seconds."
4. Sidebar (desktop) / Below fold (mobile): Pre-loaded sample note matching their role, showing what the output looks like
5. Bottom: Subtle "Not ready to record?" text link

No modal. No multi-step wizard. Single-focus page designed to drive one action.

First-action guidance

Guided linear flow (building on NUX experiment's +24.84% activation lift):

1. Page loads. Template indicator confirms: "[SOAP Mental Health] ready."
2. Single prominent CTA: "Record Your First Note" — large, centered, impossible to miss.
3. Inline guidance below CTA: "Tap to record. Twofold listens and writes your note in ~20 seconds."
4. User taps record → microphone permission prompt appears with pre-prompt context: "Twofold needs microphone access to capture your session. Recordings are never stored."
5. Recording state: pulsing waveform animation + "Recording... Tap Stop when you're done."
6. User taps stop → "Writing your [SOAP] note..." with progress bar (3-8 seconds).
7. Note appears inline: "Your first note is ready. Review, edit, or copy to your EHR."
8. Below completed note: "Start another recording" + "Explore your workspace."

Alternatives to live recording

"Not ready to record?" appears as a subtle text link below the primary CTA (never competing visually). Tapping reveals two options:

• "Try a demo recording" — A 30-second script the user reads aloud. Script example for Therapist: "This is a demonstration session. The client reports feeling anxious about an upcoming job interview. We discussed cognitive behavioral techniques including thought challenging and progressive muscle relaxation. The client will practice these techniques daily and return next week." Generates a real note from the demo, proving the product works.
• "See a sample note" — Opens the pre-loaded sample note matching their role. Viewable but clearly labeled as a sample. Ends with: "Ready to try your own? Tap Record."

Mobile-specific considerations

• Large touch target for record button: 80px height, full-width, centered in thumb zone.
• No scroll needed to see CTA and guidance text on any device 320px+ wide.
• Microphone permission pre-prompt before native iOS/Android permission dialog to increase grant rate.
• If recording on mobile while not in a session (common for FB/IG ad traffic — user is on their couch): demo recording is promoted after 10 seconds of inaction. Tooltip appears above "Not ready to record?": "Not with a client right now? Try a 30-second demo instead."
• Waveform animation is lightweight CSS, not canvas-based, to prevent in-app browser rendering issues.

Personalization of /new page

Users not ready to record

Layered escape paths, from least disruptive to most:

1. 10-second inaction trigger: Tooltip appears near "Not ready to record?" — "Not with a client? Try a 30-second demo."
2. "Not ready to record?" link: Two options — demo recording or sample note viewing.
3. "Remind me later" option (appears after 30 seconds): "Send me a reminder before my next session." Collects phone number (optional) or sets an email reminder. This is where phone collection happens — after the user has context for why.
4. /new page state persists: If the user closes the browser and returns later (via email reminder, bookmark, or direct), they see the same guided flow. No progress is lost.
5. Re-engagement email (24h after signup, no recording): "Your [SOAP] workspace is ready — try a 30-second demo to see how Twofold works." Links directly to /new.

---

Psychological Principles Applied

---

Trust & Compliance

Trust signals

• HIPAA compliance badge (shield icon) — persistent on all screens
• "Recordings are never stored" — on /new page near recording button
• BAA (Business Associate Agreement) available — linked from signup
• 4.8/5 star rating with review count
• "20,000+ clinicians" aggregate proof
• Role-specific peer counts
• "No credit card required" on results page
• "Cancel anytime" on results page

Compliance requirements

• HIPAA badge must be visible on the signup screen (results page)
• Terms of Use link on results page below signup form
• Privacy Policy link on results page below signup form
• BAA link accessible from results page (in trust footer or as text link below form)
• Clear data handling statement on /new page near recording CTA: "Recordings are processed securely and never stored. HIPAA compliant."
• Cookie consent / privacy banner if required by jurisdiction (auto-detected)

Placement

---

Expected Metrics

Funnel completion rate target

18-22% ad click to signup (end-to-end). Breakdown:

• Ad click → Q1 view: ~95% (page load)
• Q1 view → Q1 answer: 50-55% (improved from 45.48% via icon tiles — targeting 5-10pp lift)
• Q1 answer → Q5 answer: 93-96% (consistent with V1's ~2-3% per-step drop-off)
• Q5 answer → results page view: 98% (loading screen is passive)
• Results page → signup: 38-45% (improved from V1 by embedding signup on same page — eliminating one transition)

Overall: 95% x 52% x 94% x 98% x 42% = ~19.2% (vs V1's measured ~13.45% from quiz start)

Activation rate target

60-70% signup to "pressed start recording." Benchmark: NUX dialog achieved 62.42%. Guided linear flow on /new with personalized template and demo recording option should match or slightly exceed. Target: 65% blended (70% desktop, 55% mobile).

Subscription rate target

9-12% trial to paid. Benchmark: NUX dialog achieved 8.92%. Speed Start V2 users benefit from better activation + template personalization, which should improve perceived value during trial. Conservative target accounts for cold traffic quality.

Benchmarks

Key risk factors

1. Icon tiles may not solve Q1 drop-off. The 54.52% drop may be driven by traffic quality (users who never intended to engage), not question design. Mitigation: A/B test icon tiles vs text pills on Q1 specifically.
2. Q2 position change could backfire. Emotional questions require more cognitive effort than factual ones. If users aren't ready for "What would help most?" at step 2, it could increase Q2 drop-off. Mitigation: Monitor Q1→Q2 drop-off vs V1's Q1→Q2 rate (which was ~2-3%).
3. Embedded signup on results page may feel rushed. Some users might want to read the results before being asked to sign up. Mitigation: Signup form is below the fold on results — users scroll to it after absorbing the hero stat.
4. Browser ejection reliability. Universal links are not 100% reliable across all FB/IG in-app browser versions on all devices. Mitigation: Fallback email with magic link. Monitor ejection success rate in PostHog.
5. Activation gap persists. V1's 5.99% signup → first recording rate was poor. The guided /new flow targets 60-70%, which is a 10x improvement. If this target is not met, the entire V2 thesis weakens. Mitigation: The NUX experiment (62.42%) provides high-confidence evidence that guidance works. Demo recording option provides an activation path for users not in-session.

---

Implementation Complexity

Technical requirements

• Q1 icon tiles component: New component — 2x3 grid of icon+label tiles with selection state, "Other" searchable dropdown overlay. Estimated: 1 day frontend.
• Quiz question reordering: Move motivation question from position 4 to position 2 in the quiz state machine. Estimated: 1 hour.
• Embedded signup on results page: Move signup form into results page component. Remove separate signup page route. Estimated: 0.5 day frontend.
• Server-side in-app browser detection: Already exists from V1 (FBAN/FBAV user-agent parsing). No change needed.
• Browser ejection mechanism: Already implemented in V1 (universal links / deep links with auth token). Needs testing across FB/IG in-app browser versions. Estimated: 0.5 day QA.
• Guided /new page flow: Modify /new page to show single-focus guided recording flow with template pre-loaded from quiz data. Remove multi-step onboarding wizard. Add "Not ready to record?" escape path with demo recording option. Estimated: 2-3 days frontend.
• Demo recording feature: Pre-written scripts per role. Recording and note generation from demo audio. Estimated: 1 day backend + 0.5 day frontend.
• PostHog event tracking: Quiz step events, signup event, ejection success/failure, /new page actions, first recording, demo recording usage. Estimated: 0.5 day.
• Fallback email sender: Already exists from V1 (magic link email on ejection failure). No change needed.
• Results page dynamic content: Already exists from V1 (hero stat, motivation-matched headline). Minor copy updates. Estimated: 0.5 day.

Content requirements

• 6 role-specific icon assets (32px SVG icons for Q1 tiles)
• Role-specific peer count numbers (verified from analytics)
• Role-specific sample notes (Therapist, Psychiatrist, Psychologist, Social Worker, Counselor — 5 samples)
• 5 demo recording scripts (one per role)
• 1-2 clinician testimonials (already available from V1)
• 2-3 versions of results page copy and signup CTA for A/B testing

Estimated effort

Medium — Most infrastructure exists from V1. Key new work: Q1 icon tiles component, results page with embedded signup, guided /new page flow, demo recording feature.

Dependencies

• V1 quiz infrastructure (quiz state machine, results page, ejection mechanism) must be deployed and stable.
• Browser ejection validated on iOS Safari in-app (Facebook), iOS Safari in-app (Instagram), Android Chrome in-app (Facebook), Android Chrome in-app (Instagram).
• PostHog experiment setup for V1 vs V2 comparison (same experiment framework as Experiment 001).
• Role-specific icon assets from design team.
• Demo recording scripts reviewed by clinical advisory team for accuracy.
• /new page modifications coordinated with product team (affects all users, not just quiz traffic — feature flag required).

---

Google Ads Addendum

Should this variation use a quiz for Google traffic?

No. Google Ads traffic is high-intent search traffic — these users are actively looking for a clinical documentation solution. They do not need a quiz to build motivation or identify pain. A quiz would add unnecessary friction between search intent and product access. Google traffic should go to content-rich landing pages with direct signup CTAs.

Entry point for Google Ads visitors

Existing /lp/* and /specialties/* content pages, adapted per keyword cluster:

Landing page hero / above-the-fold content

Specialty-specific pages (/specialties/therapists):

• Headline: "AI-Powered Notes for Therapists — Done Before Your Next Session"
• Subheadline: "Twofold listens to your sessions and writes SOAP, DAP, and BIRP notes automatically. HIPAA compliant. Trusted by 8,500+ therapists."
• Primary CTA: "Try for Free" — large button, above fold
• Secondary CTA: "See a Sample Note" — text link below primary CTA
• Hero visual: Screenshot of a completed SOAP note for a therapy session, with the Twofold interface visible
• Below fold: How it works (3 steps: Record → Generate → Review), testimonials from therapists, pricing, FAQ

Product feature pages (/lp/ai-clinical-documentation):

• Headline: "Your Clinical Notes, Written Automatically"
• Subheadline: "Twofold captures your sessions and generates compliant clinical documentation in seconds. SOAP, DAP, BIRP, and more."
• Primary CTA: "Start Free — No Credit Card"
• Hero visual: Before/after comparison — hours spent on notes vs Twofold's 20-second generation

Competitor pages (/lp/alternative-to-freed):

• Headline: "Looking for a Freed Alternative? Try Twofold Free."
• Subheadline: "Same AI documentation. Better accuracy, lower price, real-time capture. Switch in 60 seconds."
• Primary CTA: "Try Twofold Free"
• Comparison table: Feature comparison below the fold

Signup flow changes

• Google SSO prominent: "Sign Up with Google" button is the primary signup method on Google Ads landing pages (no in-app browser restriction). Email+password as secondary option.
• No quiz: Direct landing page → signup → /new page. No quiz interstitial.
• No ejection: Google traffic arrives in the system browser. No in-app browser handling needed.
• Signup page: Standard signup page with Google SSO + email/password. Not embedded in a results page (no results to show — no quiz was taken).
• Post-signup: Same guided /new page flow as quiz traffic. Template defaults to most common for their specialty (detected from landing page URL, e.g., /specialties/therapists → SOAP Mental Health template). If specialty is unknown (generic landing page), show a quick 1-question role selector on /new before the recording CTA.

Google ad copy alignment

Specialty-specific (e.g., "AI notes for therapists"):

• Headline 1: "AI Notes for Therapists"
• Headline 2: "SOAP & DAP Notes in Seconds"
• Headline 3: "Try Free — No Credit Card"
• Description: "Twofold listens to your therapy sessions and writes compliant clinical notes automatically. Trusted by 8,500+ therapists. HIPAA compliant. Start free today."
• Landing page: /specialties/therapists — message matches "therapists" + "SOAP & DAP" + "Try Free"

Product feature (e.g., "AI clinical documentation"):

• Headline 1: "AI Clinical Documentation"
• Headline 2: "Notes Done in 20 Seconds"
• Headline 3: "HIPAA Compliant · Free Trial"
• Description: "Stop spending hours on clinical notes. Twofold captures your sessions and generates SOAP, DAP, BIRP notes automatically. Used by 20,000+ clinicians."
• Landing page: /lp/ai-clinical-documentation — message matches "AI clinical documentation" + "20 seconds" + "Free Trial"

Competitor targeting (e.g., "Freed AI alternative"):

• Headline 1: "Freed AI Alternative"
• Headline 2: "Better Accuracy, Lower Price"
• Headline 3: "Switch in 60 Seconds"
• Description: "Looking for a Freed alternative? Twofold offers real-time AI note generation, HIPAA compliance, and $49/mo pricing. Try free — no credit card."
• Landing page: /lp/alternative-to-freed — message matches "Freed alternative" + "lower price" + feature comparison

What stays the same as the Facebook variation

• Post-signup guided /new page flow (single CTA, guided recording, demo option, "not ready?" escape)
• Template personalization logic (role → format → pre-loaded template)
• Trust signals (HIPAA, clinician count, rating)
• Free trial structure (7 days, no credit card)
• Recording and note generation experience
• PostHog event tracking schema

Technical requirements specific to Google

• Google SSO integration on all /lp/* and /specialties/* pages — "Sign Up with Google" as primary CTA
• Message match validation: Ensure each ad group's keywords map to a landing page whose headline and CTA match the ad copy. Track landing page → signup conversion by ad group in PostHog.
• URL parameter passthrough: gclid, utm_source, utm_medium, utm_campaign, utm_content passed through signup to PostHog for ROAS attribution
• Landing page load speed: Google Ads Quality Score requires <3s LCP. Optimize /lp/* and /specialties/* pages for Core Web Vitals.
• 1-question role selector on /new: For users arriving from generic landing pages (no specialty in URL), show a quick "What's your role?" question (same as quiz Q1, icon tiles) before the recording CTA. This ensures template personalization without a full quiz.
• Dynamic keyword insertion (DKI) in ad headlines where applicable (e.g., "AI Notes for {Keyword:Clinicians}")
• Conversion tracking: Google Ads conversion pixel fires on signup completion and first recording (dual conversion goals for Smart Bidding optimization)