Ideas
Evidence-backed experiment ideas for pricing, onboarding, activation, retention, and expansion.
The Growthtyping idea library is an evidence-backed gallery of growth experiment patterns - not a generic prompt list. Each idea is tagged, sourced, and ready to render in your brand once you have a profile.
What the gallery covers
Patterns span the full PLG funnel:
- Pricing - parity pricing, usage sliders, lifetime deals, plan recommenders, enterprise contact cards
- Onboarding - checklists, progress gates, activation surveys, welcome modals
- Activation - ROI calculators, social proof tickers, product tours, feature discovery widgets
- Retention - upgrade prompts, usage nudges, re-engagement offers
- Expansion - referral waitlists, integration marketplaces, community join cards
Every record includes funnel surface, effort estimate (S/M/L), expected impact grounded in published evidence, and a link to the source teardown or case study.
Where patterns come from
Two sources, both rendered the same way in the gallery:
- The authored library - patterns written against published evidence, each carrying an
evidence_urlyou can read before you build. - Spotted in the wild - patterns captured from live product surfaces, shown with a screenshot and credited to the site they were found on, linking back to the original page.
Public SEO surface
Each pattern is a public page at /ideas (and /ideas/[slug]). This is deliberate distribution: the gallery markets the tool with zero ad spend, following the pricingpages.com playbook. Visitors discover proven experiments; signed-in users render them on-brand.
Free gallery vs paid render
The public gallery is free - browse patterns, read evidence summaries, see generic example renders. Paid unlocks rendering in your brand profile plus HTML/React export. Free tier is the marketing surface; on-brand output is the product.
How to pick an idea
Use three filters before you generate:
- Funnel stage - match the surface you control (pricing page, onboarding, in-app modal, email-embedded widget).
- Effort - prefer S/M patterns when you need something credible this week; reserve L for bets worth a longer eval cycle.
- Evidence - read the linked case study or teardown. If the source experiment moved a metric you care about, the pattern is a better bet than a blank prompt.
Start from proof, not vibes
Growthtyping patterns ship with evidence_url and evidence_summary fields. The gallery is curated from growth.design case studies, pricing page teardowns, Growth Unhinged AI UX patterns, and high-conversion task taxonomies - not LLM improvisation.
From gallery to prototype
- Open /ideas and filter by category or funnel surface.
- Pick a pattern that matches your current experiment backlog item.
- Select your brand profile and generate - the compiler injects your tokens, tone, and ICP into the pattern template.
- Fork variants on the canvas if you want to test framing before export.