Executive Summary — Funnel Variation Specs (v2, April 2026)

Research Foundation

This spec generation is the second full run of Agent 5, now incorporating Google Ads research alongside the existing Facebook/Instagram research. It synthesizes findings from:

• 36 competitor funnels analyzed across health, wellness, fintech, B2B SaaS, and DTC verticals (Agents 1-3)
• Twofold product/UX review including source code analysis and expert walkthrough (Agent 4/4.1)
• Real experiment data: Experiment 001 (Speed Start vs Deep Commitment — statistically significant), NUX Dialog experiment (+24.84% activation lift)
• PostHog funnel data: signup conversion rates (87.3% email, 38.8% Google SSO), activation rates (50% desktop, 35% mobile on /new page)
• Google Ads research: 40+ sources on high-intent search traffic conversion, 6 competitor Google landing page assessments, Twofold's own Google Ads performance data (7.18% CVR, $303.92 CPA across 18 ad groups)
• Team feedback (Agent 10): 17 items evaluated, 6 accepted, 5 accepted with caveats, 1 rejected, with specific directives on results page simplification, no-scroll constraint, signup framing, and ejection approach
• Version prototype preferences: V11 (rank #1), V16 (#2), V4 (#3), V1 (#4) rated by team; V2, V3, V5-V10, V12 hidden

Current Flow Assessment

Twofold's current flow (landing page → signup → 4-step onboarding → NUX dialog → /new page) has strong foundations:

What's working:

• Landing page: A- grade, strong trust signals, clear value prop
• Signup page: 87.3% email+password conversion rate
• Onboarding: under 2 minutes, smart defaults, specialty-specific personalization
• Google Ads landing pages: 7.18% CVR (above B2B SaaS benchmark of 4.7%), strong message match on /lp/_ and /specialties/_ pages

What's broken:

• Activation bottleneck: 50% desktop / 65% mobile users never start recording from /new page — this is the single biggest revenue leak
• Quiz first-question drop-off: 54.52% of Speed Start users who see Q1 never answer it — the entry moment needs optimization
• In-app browser friction: No Google SSO, unreliable scroll, recording doesn't work — ejection is necessary but adds a risky step
• Google SSO signup failure: Only 38.8% conversion despite being the more popular starting method — likely due to in-app browser limitations
• Post-onboarding gap: /new page is sparse with insufficient guidance, especially on mobile

The 5 Variations

Variation 1: Speed Start v2 — Refined Winner

The proven approach, optimized. 5-question quiz with visual role selection (reducing Q1 friction), simplified single-hero-stat results page with embedded signup, and guided linear activation. This is the evolutionary refinement of what already works.

Strategic bet: The fastest, lowest-friction path from FB ad click to first recording converts the most cold mobile traffic. Iterate on the winner rather than reinvent.

Variation 2: Instant Value — Zero-Quiz Direct Path

Skip the quiz entirely. A 3-card micro-configuration UI (role → format → session type) generates a visual note preview, with signup embedded alongside the preview. Tests whether eliminating the 54.52% Q1 drop-off via a completely different interaction model outperforms a traditional quiz.

Strategic bet: Cold traffic drops off because quizzes are friction, not because users need warming. Showing what Twofold does immediately converts more than asking questions about it.

Variation 3: Pain Path — Segmented Entry Funnels

Four distinct quiz variants matching ad creative: Time Crunch, Compliance Shield, Burnout Relief, and Practice Growth. Each ad links to a pain-specific 5-question quiz with tailored questions, results, and activation messaging. Tests whether message match from ad-to-funnel-to-activation lifts conversion versus a generic quiz.

Strategic bet: Matching the funnel to the specific pain that drove the ad click converts better than a one-size-fits-all quiz. Inspired by Calm's topic-specific quizzes and Notion's 60% CPL reduction from dynamic landing pages.

Variation 4: ROI Calculator — Data-Driven Value Revelation

7-question quiz that builds a personalized savings dashboard. Questions gather data for visible calculations (hours saved, money saved, patients gained, ROI multiplier). Interstitials between quiz phases show real-time computation. Results page is a 4-metric animated dashboard.

Strategic bet: Clinicians are analytical professionals. Concrete, personalized ROI numbers create a stronger conversion driver than emotional messaging. The quiz itself delivers value by quantifying the problem.

Variation 5: Peer Proof — Community-Driven Conversion

6-question quiz with peer data reveal after each answer ("4,200+ therapists use Twofold daily"). Results page is a community snapshot showing specialty-specific peer stats. The quiz transforms from data collection into continuous social proof delivery.

Strategic bet: Clinicians trust other clinicians more than any marketing claim. Peer data validates each quiz answer, creating momentum through belonging ("People like me use this").

Recommended Testing Order

1. First test: Variation 1 (Speed Start v2) vs Variation 2 (Instant Value) — Tests the fundamental question: does a quiz help or hurt? Speed Start v2 is the refined winner; Instant Value challenges the quiz paradigm entirely. This test has the highest learning value because it validates the foundational funnel architecture.

2. Second test: Winner of Test 1 vs Variation 3 (Pain Path) — Tests whether ad-to-funnel message match lifts conversion beyond a generic entry point. This requires coordinating ad creatives with funnel variants.

3. Third test: Winner of Test 2 vs Variation 4 (ROI Calculator) or Variation 5 (Peer Proof) — Tests the optimal value delivery mechanism within the winning structure. Choose V4 if the audience responds to quantitative value props; choose V5 if social proof is the stronger signal.

Rationale: This order maximizes learning per test. Test 1 resolves the biggest open question (quiz vs no-quiz). Test 2 validates whether personalization depth matters. Test 3 optimizes the value delivery mechanism.

Next Experiment Recommendation

Test Variation 1 (Speed Start v2) against Variation 2 (Instant Value).

Why this is the highest-value next test:

1. Speed Start is the current winner — any new variant should beat it, not just beat an unoptimized baseline.
2. The 54.52% Q1 drop-off is the single biggest leak — Variation 2 completely bypasses this by eliminating the quiz. If it works, it redefines the funnel architecture.
3. The test is clean — quiz vs no-quiz is a binary structural test, not a content test. Results will be unambiguous.
4. Implementation complexity is comparable — both variants are buildable within 1-2 sprint cycles.
5. The experiment addresses the untested hypothesis from the research: "Does a quiz actually help conversion for cold FB traffic, or is it just adding friction?" Every competitor quiz funnel study assumes quizzes are better — but none were tested against a well-designed direct path for B2B professional audiences.

If Speed Start v2 wins: The quiz model is validated. Proceed to test Pain Path (V3) for segmentation uplift. If Instant Value wins: The entire funnel strategy shifts. Focus subsequent tests on optimizing the direct-to-value model (adding peer proof elements, ROI calculations as optional enhancements).

Shared Elements Across All Variations

These elements are constant across all 5 variations:

1. Mobile-first design — all layouts optimized for 375px viewport first
2. No scrolling on quiz pages — question and widget centered, fixed position every screen
3. One CTA per page — single clear action, no competing elements
4. Signup fields: email + password only — no first name, no phone (at signup)
5. HIPAA badge + "20,000+ clinicians" + 4.8/5 rating — visible at signup moment
6. Signup-first ejection — complete signup in in-app browser, auto-eject to system browser for /new page activation (no intermediate "open in browser" page)
7. Guided linear activation on /new — strong default toward recording, subtle "not ready to record?" escape link
8. "Save your setup" / continuation framing — signup positioned as claiming personalized result, not cold registration
9. "Start Free" CTA — not "$0 Today" (tested slightly negative)
10. Progress indicators — format-appropriate (dot steps for ≤5Q, percentage bar for longer quizzes)
11. Twofold app aesthetic as default — warm palette (navy, salmon/coral, cream), professional typography, purposeful color and animation for conversion

Key Assumptions to Validate

1. Reducing Q1 friction (visual role selection, configuration UI) will meaningfully reduce the 54.52% drop-off
2. In-app browser ejection can be made seamless (no intermediate page)
3. Guided activation on /new will reproduce the NUX experiment's +24.84% lift in the quiz funnel context
4. Google Ads traffic converts better with content-rich pages + direct CTA than with a quiz
5. Google SSO prominence on the signup page will improve the 38.8% Google signup completion rate
6. Specialty-specific peer data and testimonials are available at sufficient quality for V5
7. The /lp/_ vs /specialties/_ template question has a performance-measurable answer for Google Ads