1. ./brands/{brand-slug}/brand-context.md

   -- brand identity, audience, USP

2. ./brands/{brand-slug}/product-marketing-context.md

   -- deep positioning, customer language, objections (read if it exists)

3. ./brands/{brand-slug}/sostac/03-strategy.md

   -- target segments, positioning, phasing

4. ./brands/{brand-slug}/sostac/04-tactics.md

   -- channel plan, priorities

agent-browser

Setup: Before running research, check if

agent-browser

is available (

agent-browser --version

). If the command is not found, install it:

npm install -g agent-browser && npx playwright install chromium

. If installation fails, use

WebFetch

and

WebSearch

tools as alternatives for all research tasks in this section.

• Write the headline for the most skeptical version of your ideal customer.
• The value prop must be specific, not generic. "Save time" is weak. "Cut reporting time from 4 hours to 20 minutes" is strong.
• Eliminate insider language, acronyms, and feature-first copy that assumes context visitors do not have.
• The visitor's first question is always "Is this for me?" — answer it in the headline or subheadline.

• Does the headline lead with the outcome the customer cares about, not the mechanism?
• Is there a specific number, timeframe, or comparison that makes the claim concrete?
• Does it speak directly to the primary traffic source persona (cold traffic needs more context; retargeting can be more direct)?
• Failing headline test: reads like a company tagline. Passing: reads like what a customer would say after using the product.

• One primary CTA above the fold. Always. No exceptions.
• CTA copy should describe what happens next ("Start My Free Trial") not what the user does ("Click Here").
• Button contrast: the primary CTA button should be the most visually distinct element on the page.
• Repeat the CTA at every natural decision point — after the value prop, after social proof, after objection handling, at the bottom.
• Remove competing CTAs that split attention (secondary links, navigation items, social icons near the primary CTA).

• Scanners (80% of visitors) must be able to extract the core message from headlines, subheadlines, and bullet points alone — without reading a single paragraph.
• F-pattern reading: front-load the most important words in every headline and bullet.
• Whitespace is a conversion tool. Crowded pages create cognitive overload.
• Hero image or video must reinforce the value prop — if it does not add context, it reduces conversion.
• Font size hierarchy: H1 → H2 → body should be visually distinct (minimum 4-6px step-down).

• Customer logos (recognizable brands from the target audience's world).
• Testimonials must have: real name, photo, company, and a specific outcome — not generic praise.
• Review counts with star ratings (G2, Capterra, Trustpilot, App Store rating — choose the platform most credible to your audience).
• Security badges near forms and payment fields (SSL, SOC 2, GDPR, payment processor logos).
• Usage numbers ("10,000+ teams" or "2.3M reports generated") only when the numbers are meaningful and verifiable.

• List the top 5 objections from customer interviews, support tickets, and sales call notes. Then check: does the page address all 5?
• Price objection: money-back guarantee, free trial, "cancel anytime," or pricing transparency.
• Risk objection: case studies with similar company profiles, named customer logos in the same industry.
• Effort objection: "Up and running in 5 minutes," setup guide, onboarding support promise.
• Relevance objection: use-case segmentation ("Built for [role]" sections), industry-specific testimonials.
• FAQ section: address objections as questions, not as features. "Will this work for a team of 3?" not "Team Features."

• Form fields: audit every field. Remove anything that does not directly serve conversion.
• Navigation: hide or remove site navigation from standalone landing pages. Navigation is an escape hatch.
• Mobile: test on a real device. Zoom-required text, mis-sized tap targets, and hidden CTAs are conversion killers.
• Load time: every 1-second delay costs approximately 7% in conversion rate. Target under 3 seconds on mobile.
• Decision fatigue: too many plan options, too many testimonials stacked without structure, walls of text.

1. Current state: What is the current conversion rate? What is the goal conversion rate? How is it being measured?
2. Traffic: What are the primary traffic sources (paid search, organic, email, direct)? What is the monthly visitor volume on this page or flow?
3. Drop-off data: Where do users drop off? (Funnel analytics, heatmaps, session recordings, form analytics.) What tool is in place?
4. Test history: What has already been tested? What worked? What did not? (Avoid re-testing known losers.)
5. Platform and constraints: What platform or CMS is the page built on? What changes can be made without a developer? What requires engineering time?
6. Audience context: Are visitors arriving cold (first visit) or warm (retargeting, email)? What do they already know about the product?
7. Business constraint: Is the conversion goal a lead, a trial signup, a purchase, or something else? What happens immediately after conversion?

• Funnel analytics: Google Analytics 4 or Mixpanel funnel reports — where does the funnel lose the most volume?
• Heatmaps: Hotjar, Microsoft Clarity (free), or FullStory — where do users click, scroll, and ignore?
• Form analytics: Hotjar forms or Funnellytics — which fields cause abandonment? Which fields take longest to fill?
• Session recordings: Watch 20-50 sessions of users who did NOT convert. Look for confusion patterns, rage clicks, and abandonment moments.
• A/B test results: Check prior test history before recommending anything. Never re-test a confirmed loser.

• Post-signup surveys: "What almost stopped you from signing up?" (Typeform or in-product survey, ask 3-7 days after signup to engaged users.)
• Exit-intent survey: "What stopped you from signing up today?" — one question, 4 multiple-choice answers the team has hypothesized, plus an "other" free text.
• Support ticket analysis: Common questions and objections are a CRO research goldmine.
• Sales call recordings: What objections does the sales team handle on every call?
• User interviews: 5-7 conversations with recently converted users reveal the real decision-making journey.

• Start with heatmaps and funnel analytics to identify WHERE the problem is.
• Use session recordings and surveys to understand WHY it is happening.
• Use that understanding to write better A/B test hypotheses.

• Hypothesis: "If we [change X], then [metric Y] will improve because [user behavior reason Z]."
• Primary metric: The single number that determines the winner.
• Secondary metrics: Guard rails (e.g., "signup rate should not drop while we test headline").
• Minimum sample size: See Section 6 for calculation guidance.
• Expected duration: Based on current traffic volume.

• Just ship it: The current version is clearly broken (missing CTA, illegible on mobile, wrong headline). No need to test obvious fixes.
• Run an A/B test: The change is a hypothesis with uncertain outcome, the page has sufficient traffic, and the current version is performing reasonably.

1. Value proposition / headline (largest potential swing)
2. CTA copy and placement
3. Form length (removing fields)
4. Social proof type and placement
5. Pricing page structure and anchoring
6. Page layout and visual hierarchy
7. Button color, size, shape (smallest impact — test last)

• One change per test variant. Multi-variate requires dramatically more traffic.
• Run tests for a minimum of 2 business cycles (typically 2 weeks minimum) to account for day-of-week variation.
• Do not stop tests early because one variant looks like a winner — regression to the mean is real.
• Document every test: hypothesis, date range, traffic split, results, confidence level, and decision made.

• Visitors have no prior awareness of the brand.
• Must establish credibility and context immediately.
• Longer pages perform better — more space to build trust and address objections.
• Social proof must include recognizable reference points (logos, industries, role titles).
• Headline: lead with the pain or problem, not the solution name.

• Visitors have seen the brand at least once.
• Skip the basic explanation — get to the compelling offer faster.
• Shorter, more direct pages often outperform.
• Headline can reference their prior action ("Still thinking about [product name]?").
• Stronger urgency or specific incentive (limited-time offer, bonus content) can close the gap.

• Visitors are actively evaluating — they are in buy mode.
• Comparison content, guarantee language, and risk reduction are highest priority.
• Minimize steps between landing and conversion. Remove all friction.
• Live chat or demo booking CTAs can dramatically improve conversion at this stage.

• Visitors arrive with context from the referring source. Leverage it.
• Custom landing pages per referral source outperform generic pages by 20-40%.
• Match the message to the referring context ("Welcome, [Partner Name] customers").

1. Read brand context and SOSTAC (Section 0). Always. Non-negotiable.
2. Identify the CRO domain (Section 1). Read the appropriate reference file before producing recommendations.
3. Ask the diagnostic questions (Section 3) if the user has not already provided this information. Do not produce recommendations without knowing current CVR and traffic source.
4. Research the page (Research Mode) using

   agent-browser

   if a URL is available. Screenshot the current state before recommending changes.
5. Apply the universal priority framework (Section 2). Start with value proposition, not button colors.
6. Deliver structured output (Section 5). Quick Wins first, then High-Impact, then Test Hypotheses, then Copy Alternatives.
7. Save deliverables to

   ./brands/{brand-slug}/campaigns/cro/

   .
8. Recommend next steps: What to ship immediately, what to test, when to review results.

• Significant UX redesign or new page build -- recommend involving a designer and developer; CRO brief can define the conversion requirements.
• Paid traffic strategy (the ads sending traffic to the page) -- route to marketing-paid-ads skill.
• Email nurture for leads captured via forms -- route to marketing-email skill.
• SEO implications of page changes (meta, structure, content) -- route to marketing-seo skill.
• Brand messaging or positioning questions that surface during CRO audit -- route to product-marketing-context or brand-context update.