Google Ads Context
Seed InputChannel context for Google Ads — how search traffic differs from Facebook/Instagram, Twofold's current Google Ads setup with /lp/_ and /specialties/_ landing pages, performance data, and Quality Score considerations
raw_input/google_ads_context.md
Google Ads Channel Context
This document provides context for all agents about Twofold Health's Google Ads channel. Read this alongside the other raw_input/ files when generating research or specs that should account for Google Ads traffic.
Why Google Ads Matters
Twofold currently acquires users through Facebook and Instagram ads that drive traffic to a quiz funnel at start.trytwofold.com. We are expanding to also optimize for Google Ads search traffic, which is a fundamentally different channel.
How Google Ads Differs from Facebook/Instagram
| Dimension | Facebook/Instagram | Google Search |
|---|---|---|
| User intent | Low — interrupted while scrolling | High — actively searching for a solution |
| Browser context | In-app browser (limited OAuth, unreliable scroll, no password manager) | Full system browser (Google SSO available, standard capabilities) |
| Landing page needs | Hook + quiz (engagement-first, earn attention) | Answer the query + convert (fulfillment-first) |
| Google SSO | Unavailable in in-app browser | Available — lowest-friction signup option |
| Quality Score | N/A | Must optimize: ad relevance, landing page experience, expected CTR |
| Ejection needed | Yes — must escape in-app browser for recording to work | No — already in full browser |
| User knowledge | Often unaware they need the product | Actively researching solutions |
Twofold's Current Google Ads Setup
Landing Pages (Working Well)
Twofold runs Google Ads pointing to content-rich landing pages on trytwofold.com, NOT the quiz page. These pages have high Quality Scores. Two templates exist:
-
/lp/*template (e.g.,/lp/ai-therapy-notes,/lp/ai-therapy-dictation) — Keyword-swapped pages: Hero > testimonials > features > how-it-works > pricing > privacy/security > 4 FAQs > footer. ~2,500+ words, proper H1/H2, Product schema (4.9/5 rating). Server-rendered. -
/specialties/*template (e.g.,/specialties/behavioral-health,/specialties/psychiatry) — More differentiated. Adds 6-card feature grid with specialty-specific items, richer structured data (FAQPage, Service, BreadcrumbList schemas), more FAQs (6-9). Server-rendered.
The quiz page (start.trytwofold.com) was tried as a Google Ads destination and got a low Quality Score because it has robots: noindex, no <h1>, is pure client-side React, and has no keyword-relevant text content.
Ad Groups and Keyword Themes
From raw_input/google_ads_report_2026-04-02.csv, Twofold's active ad groups cover:
Specialty-specific:
- Mental Health, Psychology, Psychiatry, Social Work (paused), Counseling, Physical Therapy
Product feature:
- Therapy Dictation, Therapy Charting, Note Types (SOAP/DAP/BIRP), Progress Notes, Treatment Plans, Therapy — Note App, Therapy — Generator, Therapy — Scribe
Compliance-driven:
- HIPAA
General:
- Therapy (Core)
Current Performance (from ads report)
- Total: 1,748 clicks, 26,856 impressions, 6.51% CTR, 7.18% conversion rate, $303.92 avg cost/conversion
- Best conversion rates: Note Types (11.22%), Therapy Core (10.85%), Psychology (10.00%), Psychiatry (8.64%)
- Best cost efficiency: Therapy Charting ($47.91/conv), Note Types ($198.70/conv)
- Highest volume: Therapy Dictation (377 clicks, 29.5 conversions), Psychiatry (301 clicks, 26 conversions)
- All ads rated "Excellent" ad strength except Psychiatry ("Good")
Target Keyword Clusters
When researching competitors and structuring recommendations, consider these keyword types:
-
Specialty-specific: "AI therapy notes", "psychiatry AI scribe", "psychology documentation software", "social work notes AI", "counseling notes generator", "physical therapy AI scribe"
-
Product feature: "AI progress note generator", "SOAP note generator", "therapy dictation software", "treatment plan generator", "AI medical scribe", "clinical documentation AI", "therapy charting software"
-
Compliance-driven: "HIPAA compliant AI notes", "HIPAA AI scribe", "HIPAA compliant therapy software"
-
Competitor targeting: "[Competitor] alternative", "best AI scribe for therapists", "[Competitor] vs Twofold"
-
Pain-driven: "reduce clinical documentation time", "stop spending hours on therapy notes", "therapy note taking too long"
Core Research Questions
The research pipeline should answer:
-
How do competitors structure their Google Ads landing pages? Do they use quizzes, content pages, comparison pages, or hybrids? What's different from their Facebook pages?
-
Should the quiz be part of the Google Ads funnel? For which keyword clusters does a quiz add value vs. create friction? Research shows quizzes work for "commercial investigation" queries but hurt conversion for transactional queries.
-
What converts best for high-intent clinical SaaS keywords? Not generic SaaS — specifically healthcare/clinical software targeting professionals.
-
What can we learn from competitors' Google Ads approach? Especially competitors like Freed, Nuance/DAX, Ambience Healthcare who target similar keywords.
-
How should the funnel differ by keyword intent? A clinician searching "AI therapy notes" has different intent than one searching "Freed alternative" or "SOAP note generator free trial."
Key Research Finding: Search Traffic Conversion
Detailed research is at outputs/search-traffic-landing-page-research.md. The most important finding for agents:
- Quiz funnels show 40-60% conversion lifts overall, but paid search traffic only completes quizzes at 35-50% (vs 55-70% for organic search)
- Message match (ad headline echoed in page headline) drives 20-35% higher conversion — this is the #1 optimization lever
- Above-the-fold content must answer the search query within 5 seconds
- For search traffic, keep pre-quiz content minimal: headline matching query, 1-2 sentence value prop, trust logos
- Google SSO being available changes the signup friction equation significantly
- Separate landing pages per ad group can drive +83% conversion improvement (BioRender case study)