opensite-ai/opensite-skills / cro-optimization
cro-optimization
VerifiedRun conversion rate optimization through hypothesis-driven testing including audit, hypothesis generation, test design, statistical analysis, and rollout decisions. Use this skill whenever the user wants to optimize conversion, run A/B tests, audit a funnel, generate test hypotheses, design experiments, or analyze test results. Triggers on conversion optimization, CRO, A/B test, split test, multivariate test, hypothesis, conversion funnel, funnel audit, experiment design, statistical significance, lift, optimization. Also triggers when the user has a conversion problem and isn't sure where to start, or when test results are ambiguous and need interpretation.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 10.7 KB |
| references/hypothesis-library.md | markdown | 10.1 KB |
references/hypothesis-library.md
View on GitHub# Hypothesis Library
Common high-impact CRO hypothesis patterns by funnel stage. Use as starting points; adapt to the specific product and audience. Each pattern includes the typical evidence that supports it and the typical mechanism behind why it works.
---
## Top of funnel: hero and value proposition
### H1: Outcome-led headline beats feature-led
**Hypothesis:** Replacing a feature-led headline ("Advanced workflow automation") with an outcome-led headline ("Ship 30% faster") will increase scroll-past-hero rate and CTA clicks.
**Evidence to look for:**
- Session replays show users bouncing within 3 seconds
- Heatmaps show low engagement below the hero
- The current headline names a feature, not a benefit
**Mechanism:** Users grant 3 to 5 seconds to decide whether to continue. Outcome language is processed faster than feature language.
**Test design:** Hero-only swap. Control vs variant. Primary metric: scroll past hero + CTA click.
---
### H2: Specific number beats vague claim
**Hypothesis:** Replacing "Save time on month-end close" with "Close in 3 days, not 2 weeks" will increase signup rate.
**Mechanism:** Specific numbers are more credible and more memorable than vague claims.
---
### H3: Customer-facing ID over jargon
**Hypothesis:** Replacing internal/industry jargon ("ETL pipeline orchestration") with customer language ("Move data between your tools without engineering") will improve hero engagement.
**Mechanism:** Industry insiders use jargon. Most prospects are not industry insiders.
---
### H4: Real customer photo beats stock
**Hypothesis:** Replacing stock-photo hero imagery with photos of actual customers using the product will increase signup rate.
**Mechanism:** Stock photos are pattern-recognized as marketing fluff. Real customer photos signal authenticity.
---
## Mid-funnel: feature blocks and social proof
### H5: Specific named customer logos beat unnamed
**Hypothesis:** Replacing "Trusted by 1,000+ companies" with named logos will improve trust and conversion.
**Evidence:** Heatmaps show users hovering near social proof but not converting.
**Mechanism:** Named logos provide verifiable trust. Numbers without names are easily fabricated and users know it.
---
### H6: Specific outcome quote beats generic praise
**Hypothesis:** Replacing "We love this product!" with "Reduced our onboarding time by 40 percent in three months" will increase trust score and conversion.
**Mechanism:** Specific outcomes are credible because they're quantifiable. Generic praise is dismissed as cherry-picked.
---
### H7: Founder quote at scale stage
**Hypothesis:** Adding a founder quote and photo on the homepage will increase early-stage trust signals.
**Evidence:** Brand is unknown. No major customer logos to display yet.
**Mechanism:** Personal accountability from a named person beats anonymous brand voice when trust signals are scarce.
---
### H8: Comparison table for competitive categories
**Hypothesis:** Adding a comparison table against named competitors will increase conversion for prospects in active evaluation.
**Mechanism:** Users in evaluation mode are comparing anyway. Honest comparison reduces objection-handling work.
**Risks:** Can backfire if comparison feels biased or if competitors are smaller than you (gives them oxygen).
---
## Form and signup
### H9: Fewer form fields wins
**Hypothesis:** Reducing signup fields from 7 to 3 will increase signup conversion.
**Evidence:** Form analytics show high abandonment, especially on optional fields.
**Mechanism:** Each form field is friction. Optional fields without value to the user feel like data collection for company benefit.
---
### H10: Progressive profiling beats upfront fields
**Hypothesis:** Moving non-essential signup fields to a post-signup onboarding flow will increase initial signup conversion without hurting profile completion.
**Mechanism:** Users commit incrementally. The hardest "yes" is the first one.
---
### H11: Social signup options reduce friction
**Hypothesis:** Adding "Sign up with Google" alongside email signup will increase signup rate.
**Risks:** OAuth providers can also reduce signups for users uncomfortable with social login. Test, don't assume.
---
### H12: Visible password requirements reduce errors
**Hypothesis:** Showing password requirements (8 chars, etc.) above the field instead of as error messages will reduce form errors and abandonment.
**Mechanism:** Errors after submission feel like punishment. Pre-emptive guidance feels like assistance.
---
### H13: Trust signals near form
**Hypothesis:** Adding "Free for 14 days. No credit card." adjacent to the CTA will increase form completion.
**Mechanism:** Reduces perceived risk at the point of commitment.
---
## Pricing page
### H14: Annotated tiers beat unannotated
**Hypothesis:** Adding "Best for" labels on each pricing tier ("Best for solo founders," "Best for teams of 10+") will increase plan selection clarity.
**Mechanism:** Users self-select faster when given decision shortcuts.
---
### H15: Annual emphasis beats monthly default
**Hypothesis:** Defaulting the pricing toggle to annual (with savings highlighted) will increase annual plan adoption.
**Mechanism:** Defaults stick. Loss-framed savings ("Save $X") are more compelling than gain-framed.
---
### H16: Clear "what's included" reduces confusion
**Hypothesis:** Adding a checkmark feature comparison below the pricing tiers will reduce sales-assist requests and improve self-service conversion.
**Mechanism:** Many prospects abandon when they can't quickly verify a needed feature is in their tier.
---
### H17: Social proof on pricing page
**Hypothesis:** Adding tier-specific customer testimonials ("Why we picked Team plan") will increase plan upgrade rate.
**Mechanism:** Tier-specific proof reduces fear of overpaying or under-buying.
---
## Checkout / cart
### H18: Free shipping threshold visible
**Hypothesis:** Adding a "$X away from free shipping" message in the cart will increase average order value.
**Evidence:** Cart abandonment correlated with sub-threshold orders.
**Mechanism:** Goal gradient. Users near a goal accelerate toward it.
---
### H19: Guest checkout option
**Hypothesis:** Adding a guest checkout option will increase conversion of first-time buyers.
**Mechanism:** Account creation is friction at the worst possible moment (when the user has just decided to buy).
---
### H20: Multiple payment methods
**Hypothesis:** Adding [PayPal / Apple Pay / Google Pay / etc.] at checkout will increase mobile conversion.
**Mechanism:** Saved payment methods reduce friction. Mobile typing is painful.
---
### H21: Visible total throughout checkout
**Hypothesis:** Showing the running total persistently in the checkout flow will reduce abandonment.
**Evidence:** Step-by-step abandonment shows users dropping after seeing fees.
**Mechanism:** Surprise cost is the #1 cause of cart abandonment.
---
### H22: Trust badges at checkout
**Hypothesis:** Adding security badges (SSL, PCI, money-back guarantee) at the payment step will increase completion.
**Mechanism:** Security concerns peak at payment entry.
---
## Onboarding (post-signup)
### H23: Single first action vs feature tour
**Hypothesis:** Replacing a feature tour with a single guided first action will increase activation.
**Mechanism:** Feature tours teach. First actions deliver value. Users who experience value retain better than users who learn features.
---
### H24: Sample data on first login
**Hypothesis:** Pre-loading the new user's account with sample data will increase first-week retention.
**Mechanism:** Empty states are demotivating. Pre-populated data shows "what good looks like."
---
### H25: Time-to-value reduction
**Hypothesis:** Reducing the steps from signup to first valuable action from 12 to 4 will increase activation rate.
**Evidence:** Funnel analysis shows huge drops in the middle of the onboarding flow.
**Mechanism:** Each step is a chance to bounce. Fewer steps = more activations.
---
## Email and notifications
### H26: Subject line specificity
**Hypothesis:** Replacing "Update from [Brand]" with a specific benefit-led subject ("Your invoice is ready" or "Save 20% this week") will increase open rate.
**Mechanism:** Inbox is competitive. Specificity earns the open.
---
### H27: Single CTA per email
**Hypothesis:** Reducing email CTAs from 3 to 1 will increase click-through to primary action.
**Mechanism:** Choice paralysis. Multiple CTAs dilute the primary action.
---
### H28: Personalization beyond name
**Hypothesis:** Personalizing email body content based on user behavior (last action, product viewed) will increase reactivation rate.
**Mechanism:** Generic emails feel like spam. Behaviorally relevant emails feel like service.
---
## Mobile-specific
### H29: Tap target sizing
**Hypothesis:** Increasing tap targets to meet 44x44 minimum will reduce mobile bounce rate.
**Evidence:** Mobile heatmaps show mis-taps and rage clicks.
**Mechanism:** Small targets fail more often than large ones.
---
### H30: Sticky CTA on mobile
**Hypothesis:** Adding a sticky bottom CTA on mobile pages will increase conversion.
**Mechanism:** Long mobile pages bury CTAs. Sticky keeps the action available.
**Risks:** Can cover content. Test for accessibility issues.
---
### H31: Mobile-specific headlines
**Hypothesis:** Shortening headlines for mobile viewports (vs desktop) will reduce bounce rate.
**Mechanism:** Long headlines on mobile become walls of text. Short headlines preserve hierarchy.
---
## Compounding tests
The biggest CRO wins come from compounding learnings. After running 5 to 10 tests, you should know:
- What kind of headlines win in your category
- How long your forms should be
- Which trust signals move your audience
- Which features matter to your buyer
- What price framing converts
These insights compound across tests. Document each test's lesson, even when the test loses.
A pattern to watch:
- **Test #1:** Outcome-led headline wins
- **Test #5:** Outcome-led email subject wins
- **Test #10:** Outcome-led ad copy wins
You haven't just run 10 tests. You've discovered that "outcome-led messaging" is your audience's preference, broadly. That insight is worth more than any single test.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |