Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Form optimization for lead capture, contact, demo request, application, and checkout forms. Covers field-cost analysis, multi-step form design, validation UX, mobile optimization, and A/B testing frameworks.
.claude/skills/borghei-form-cro/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-12 | ✗→✓ | ▲ Improved | 400% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 1832% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 323% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 280% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 393% | 0% |
Production-grade form optimization framework covering field-cost analysis, layout engineering, multi-step form architecture, validation UX patterns, mobile-specific optimization, and structured A/B test design. Applicable to lead capture, contact, demo request, application, survey, and checkout forms. For signup/registration flows, use signup-flow-cro.
Before optimizing the form, confirm these inputs. If any is unknown or vague, ASK — do not assume:
Stop rule: ask only the 2-3 that most change the output. If the user says "just draft it," proceed and list your assumptions at the top of the audit.
| Question | Why It Matters | |----------|---------------| | What type of form? (lead capture, contact, demo, application, checkout) | Different types have different optimal field counts | | How many fields currently? | Establishes baseline friction level | | What is the current completion rate? | Benchmark for improvement | | Where do users abandon? (if field-level analytics exist) | Identifies the specific friction point | | Mobile vs desktop split? | Mobile forms need separate optimization | | What happens with submitted data? | Determines which fields are truly necessary | | Which fields are actually used in follow-up? | Often reveals 30-50% of fields are never used | | Any compliance requirements? (GDPR, HIPAA) | Constrains what can be removed |
Every field has a cost measured in abandonment. The question is not "what data would be nice to have" but "what data is worth the conversion loss."
| Field Type | Estimated Abandonment Cost | Justification Threshold | |-----------|---------------------------|------------------------| | Email | Baseline (near zero for gated content) | Always justified for lead forms | | First name | +2-3% drop | Justified if personalization drives follow-up | | Last name | +2-3% drop | Rarely justified for first touch | | Phone number | +5-10% drop | Only if sales will call within 24 hours | | Company name | +3-5% drop | Justified for B2B qualification | | Company size | +3-5% drop | Justified only if routing decisions depend on it | | Job title | +3-5% drop | Can often be enriched post-submission | | Industry | +2-3% drop | Can often be enriched post-submission | | Message/textarea | +5-8% drop | Justified for contact forms, not for lead capture | | Budget | +8-12% drop | Only justified for high-intent demo/sales forms | | Custom question | +3-5% per question | Must directly affect lead routing or qualification |
Before including any field, ask: Can this be enriched after submission?
| Field | Enrichable? | Method | Keep in Form? | |-------|------------|--------|---------------| | Company name | Yes (from email domain) | Clearbit, Apollo, manual lookup | Remove | | Company size | Yes (from company name) | Enrichment API | Remove | | Industry | Yes (from company name) | Enrichment API | Remove | | Job title | Partially (from LinkedIn) | Manual enrichment | Remove unless critical for routing | | Phone number | No | Must be provided | Keep only if sales calls immediately | | Budget | No | Must be stated | Keep only for high-intent forms |
| Form Type | Minimum Fields | Optimal Fields | Maximum Fields | |-----------|---------------|----------------|----------------| | Newsletter signup | Email | Email | Email + First name | | Content download | Email | Email + First name | Email + Name + Company | | Contact form | Email + Message | Email + Name + Message | Email + Name + Subject + Message | | Demo request | Email + Company | Email + Name + Company + Role | + Phone + Use case + Team size | | Application form | Varies by requirement | -- | All required fields (justified individually) |
| Factor | Single-Step | Multi-Step | |--------|------------|------------| | Total fields | < 5 fields | > 5 fields | | Field complexity | Simple text inputs | Mix of dropdowns, checkboxes, conditional fields | | User motivation | Low-commitment (newsletter, content) | High-commitment (demo, application) | | Qualification need | No routing needed | Different paths based on answers | | Mobile proportion | < 30% mobile | > 50% mobile |
Step structure:
Progress indication:
Psychological commitment:
Back navigation:
| Pattern | When to Use | Example | |---------|-------------|---------| | Above-field labels | Default for most forms | Label sits above the input | | Inline labels (floating) | Space-constrained layouts | Label moves from inside to above on focus | | Left-aligned labels | Wide desktop forms | Label to the left of field | | Placeholder-only labels | Never | Disappears on input, accessibility failure |
| Data Needed | Best Input Type | Avoid | |------------|-----------------|-------| | Email | type="email" with validation | Plain text input | | Phone | type="tel" with format mask | Plain text input | | Country | Searchable dropdown | Long static dropdown | | Company size | Button group (1-10, 11-50, 51-200, 200+) | Free text input | | Interest/topic | Checkbox group (max 6 options) | Multi-select dropdown | | Message | Textarea (3-4 rows visible) | Single-line text input | | Date | Native date picker | Three separate dropdowns |
Show additional fields based on earlier answers. This reduces visible complexity while capturing necessary data.
Example: "What is your primary goal?" dropdown shows "Budget range" only if they select "Ready to buy" or "Evaluating solutions."
Rules:
| Validation Type | When to Use | |----------------|-------------| | Real-time (on blur) | Email format, phone format, required fields | | On submit | Complex validation, server-side checks | | Inline suggestions | Company name auto-complete, address lookup |
| Pattern | Good | Bad | |---------|------|-----| | Position | Below the field, in context | Top of form, disconnected | | Tone | "Please enter a valid email address" | "Error: Invalid input" | | Specificity | "Phone must include area code" | "Invalid phone number" | | Color | Red text + red border on field | Red banner at top of page | | Icon | Error icon next to message | No visual indicator |
| Form Type | Weak Copy | Strong Copy | Strongest Copy | |-----------|-----------|-------------|----------------| | Content download | Submit | Download Guide | Get My Free Guide | | Demo request | Submit | Request Demo | Schedule My Demo | | Contact form | Send | Send Message | Get in Touch | | Newsletter | Subscribe | Join Newsletter | Get Weekly Tips | | Free trial | Sign Up | Start Free Trial | Start Building Free |
Rules:
| Element | Best Practice | |---------|---------------| | Color | High contrast against form background, consistent with brand CTA color | | Size | Full-width on mobile, min 44px height for touch targets | | Position | Immediately below last field, no gap | | Loading state | Show spinner + "Sending..." to prevent double-submit | | Disabled state | Disabled until required fields are valid (with clear visual distinction) |
| Rule | Implementation | |------|---------------| | Touch targets | Minimum 44x44px for all interactive elements | | Keyboard types | type="email" for email, type="tel" for phone, type="number" for numeric | | Auto-focus | Focus first field on page load (with keyboard open) | | Sticky submit | Pin submit button to bottom of viewport on long forms | | Input spacing | Minimum 8px between fields to prevent mis-taps | | Dropdown alternatives | Use button groups or radio buttons instead of dropdowns on mobile | | Auto-fill | Support browser auto-fill for standard fields (name, email, phone, address) |
| Aspect | Desktop | Mobile | |--------|---------|--------| | Layout | 1 or 2 columns | Always 1 column | | Field count | Up to 8 | Max 5 per step | | Dropdown | Standard dropdown | Bottom sheet or full-screen picker | | Help text | Hover tooltips | Always-visible inline text | | Validation | On blur | On blur + on submit summary |
| Signal | Placement | Impact | |--------|-----------|--------| | Privacy assurance | Below submit button | "We'll never share your email" | | Security badges | Next to form container | SSL, SOC2, GDPR compliance | | Testimonial | Adjacent to form | Social proof reduces hesitation | | Response time | Below submit button | "We respond within 2 hours" | | Subscriber/user count | Above or within form | "Join 10,000+ subscribers" |
| Element | Purpose | Example | |---------|---------|---------| | Form header | Remind what they get | "Get your free SEO audit report" | | Bullet list above form | Reinforce value | "You'll get: Full site analysis, Priority fix list, 30-min review call" | | Expected next step | Set expectations | "After submitting, we'll email your report within 24 hours" |
Goal: Maximize completions while capturing qualified leads.
Goal: Capture qualified prospects ready for sales conversation.
Goal: Enable communication while routing to correct team.
Goal: Capture enough detail for accurate quoting.
| Test | Hypothesis | Success Metric | |------|-----------|----------------| | Remove phone field | Removing phone increases completion by 5-10% | Completion rate | | Single-step to multi-step | Multi-step increases completion for 6+ field forms | Completion rate + submission quality | | CTA copy change | Value-specific copy increases clicks by 10-20% | Click-through rate | | Add social proof | Testimonial near form increases trust | Completion rate |
| Test | Hypothesis | Success Metric | |------|-----------|----------------| | Field order change | Easiest fields first increases step-1 completion | Step completion rates | | Inline validation | Real-time feedback reduces form errors | Error rate + completion rate | | Add progress bar | Visual progress on multi-step increases completion | Completion rate | | Embedded calendar vs form | Calendar reduces friction for demo requests | Booking rate |
| Metric | Formula | Benchmark | |--------|---------|-----------| | Form completion rate | Submissions / Form views | 20-40% for lead forms, 5-15% for long forms | | Field-level drop-off | Abandonment per field | Identify the highest-drop field | | Time to complete | Avg seconds from first interaction to submit | < 60s for simple forms, < 3min for complex | | Error rate | Users who see error / total users | < 10% | | Mobile completion rate | Mobile submissions / Mobile form views | Should be within 20% of desktop rate |
Track these events in your analytics:
| Artifact | Format | Description | |----------|--------|-------------| | Form Audit Report | Issue/Impact/Fix/Priority table | Per-field analysis with estimated abandonment cost | | Recommended Field Set | Justified list | Required vs optional vs enrichable fields with rationale | | Field Layout Specification | Annotated outline | Order, grouping, label style, validation rules, mobile adaptations | | CTA Copy Options | 3-option table | Button text variants with reasoning and expected impact | | A/B Test Plan | Prioritized table | Top 5 tests with hypothesis, variant, metric, and priority | | Mobile Optimization Checklist | Checkbox list | Mobile-specific fixes with implementation notes |
Purpose: Score a form against CRO best practices across field count, field types, CTA quality, mobile readiness, and trust signals.
bashpython scripts/form_scorer.py form_config.json python scripts/form_scorer.py form_config.json --json
| Flag | Required | Description | |------|----------|-------------| | form_config.json | Yes | JSON file with form fields, CTA, and context metadata | | --json | No | Output results as JSON |
Purpose: Calculate the estimated abandonment cost of each form field and recommend fields to remove, keep, or make enrichable.
bashpython scripts/field_cost_analyzer.py form_fields.json python scripts/field_cost_analyzer.py form_fields.json --json
| Flag | Required | Description | |------|----------|-------------| | form_fields.json | Yes | JSON file with form fields and their types | | --json | No | Output results as JSON | | --monthly-visitors | No | Monthly form visitors for dollar impact estimate (default: 1000) | | --current-rate | No | Current form completion rate as percentage (default: 25) | | --value-per-lead | No | Dollar value per lead for ROI calculation (default: 50) |
Purpose: Calculate A/B test sample size, duration, and statistical significance for form optimization experiments.
bashpython scripts/ab_test_calculator.py --baseline 25 --lift 10 --traffic 500 python scripts/ab_test_calculator.py --baseline 25 --lift 10 --traffic 500 --json
| Flag | Required | Description | |------|----------|-------------| | --baseline | Yes | Current conversion rate as percentage (e.g., 25 for 25%) | | --lift | Yes | Minimum detectable lift as percentage (e.g., 10 for 10% relative lift) | | --traffic | Yes | Daily traffic (visitors per day to the form) | | --confidence | No | Confidence level as percentage (default: 95) | | --json | No | Output results as JSON |
| Problem | Likely Cause | Solution | |---------|-------------|----------| | Form completion rate below 15% | Too many fields or high-friction fields present | Run field_cost_analyzer.py to identify and remove high-cost fields; target email-only for first-touch lead forms | | Mobile completion rate 50%+ lower than desktop | Form not optimized for touch input | Ensure 44px touch targets, single-column layout, native keyboard types; replace dropdowns with button groups on mobile | | Users start but do not finish the form | Friction in middle fields (phone, budget, message) | Move high-friction fields to later steps in a multi-step form; capture email in step 1 | | High error rate on email or phone fields | Validation too aggressive or unclear error messages | Validate on blur (not keystroke); use specific error copy ("Please include @ in email") not generic ("Invalid input") | | A/B test results are inconclusive after 4 weeks | Insufficient sample size or too small a lift target | Use ab_test_calculator.py to confirm required sample size; consider testing bigger changes (field removal vs copy tweak) | | CTA clicks are low despite good page traffic | CTA copy is generic or button is not prominent enough | Replace "Submit" with value-specific copy ("Get My Report"); ensure CTA is full-width on mobile, high-contrast color |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 15,742 | 17,187 | +9% | 1 | 1 | 0% | 2,353 | 7,914 | +236% | 0 | 0 | — |
case-02 | fail→fail | 28,160 | 37,568 | +33% | 1 | 1 | 0% | 4,233 | 11,155 | +164% | 0 | 0 | — |
case-03 | fail→fail | 32,976 | 24,908 | -24% | 1 | 1 | 0% | 5,683 | 9,146 | +61% | 0 | 0 | — |
case-04 | pass→pass | 12,034 | 18,294 | +52% | 1 | 1 | 0% | 1,861 | 7,938 | +327% | 0 | 0 | — |
case-05 | pass→pass | 11,551 | 14,937 | +29% | 1 | 1 | 0% | 1,712 | 7,397 | +332% | 0 | 0 | — |
case-06 | pass→pass | 10,342 | 11,859 | +15% | 1 | 1 | 0% | 1,534 | 7,042 | +359% | 0 | 0 | — |
case-07 | pass→pass | 9,814 | 7,930 | -19% | 1 | 1 | 0% | 1,353 | 6,393 | +373% | 0 | 0 | — |
case-08 | pass→pass | 9,271 | 12,327 | +33% | 1 | 1 | 0% | 1,451 | 7,428 | +412% | 0 | 0 | — |
case-09 | pass→pass | 9,171 | 6,788 | -26% | 1 | 1 | 0% | 1,291 | 6,225 | +382% | 0 | 0 | — |
case-10 | pass→pass | 8,136 | 6,733 | -17% | 1 | 1 | 0% | 1,267 | 6,240 | +393% | 0 | 0 | — |
case-11 | pass→pass | 8,951 | 7,233 | -19% | 1 | 1 | 0% | 1,432 | 6,428 | +349% | 0 | 0 | — |
case-12 | fail→pass | 8,490 | 6,859 | -19% | 1 | 1 | 0% | 1,227 | 6,139 | +400% | 0 | 0 | — |
case-13 | fail→pass | 2,196 | 4,423 | +101% | 1 | 1 | 0% | 315 | 6,086 | +1832% | 0 | 0 | — |
case-14 | fail→pass | 7,854 | 3,160 | -60% | 1 | 1 | 0% | 1,378 | 5,829 | +323% | 0 | 0 | — |
case-15 | fail→pass | 9,359 | 1,852 | -80% | 1 | 1 | 0% | 1,463 | 5,554 | +280% | 0 | 0 | — |
case-16 | pass→pass | 6,151 | 7,031 | +14% | 1 | 1 | 0% | 1,012 | 6,401 | +533% | 0 | 0 | — |
case-17 | pass→pass | 12,187 | 11,754 | -4% | 1 | 1 | 0% | 1,782 | 7,062 | +296% | 0 | 0 | — |
case-18 | fail→pass | 9,908 | 13,032 | +32% | 1 | 1 | 0% | 1,489 | 7,346 | +393% | 0 | 0 | — |
case-19 | pass→pass | 7,432 | 9,776 | +32% | 1 | 1 | 0% | 1,021 | 6,721 | +558% | 0 | 0 | — |
case-20 | fail→fail | 16,814 | 18,195 | +8% | 1 | 1 | 0% | 2,583 | 8,030 | +211% | 0 | 0 | — |
case-21 | fail→fail | 15,744 | 20,394 | +30% | 1 | 1 | 0% | 2,371 | 8,285 | +249% | 0 | 0 | — |
case-22 | fail→pass | 18,685 | 26,024 | +39% | 1 | 1 | 0% | 2,741 | 8,869 | +224% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +27 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.