Acquisition.md
← Playbooks
Landing Pages · 22 min

Landing Page System

A repeatable framework for building high-converting pages.

There are ~9 modules in every high-converting landing page. Your job is to pick which 5-7, in what order, for this audience.

tl;dr
  • Build 1 page in Framer + Relume, not 12 pages in Figma.
  • Use Clay + PostHog to learn what converts before you polish copy.
  • Default to Attio and n8n; Salesforce only if you enjoy tax-grade friction.
  • Personalise 1:1 with Clay + Mutiny when intent crosses 70/100.
  • Ship with one operator, 11x, and Cursor; dead tactic is committee review.
Action plan
Action plan0/3 · 0%
Progress saved locally. Sign in to sync
Week one
  1. 01Day 1: Define 1 ICP and score 200 accounts into 3 intent bands.
  2. 02Day 2: Write 3 headline variants and pick 1 CTA rule for the whole page.
  3. 03Day 3: Generate the page skeleton in a wireframe tool and move it into the site builder.
  4. 04Day 4: Add proof blocks: 1 quote, 1 logo bar, and 1 metric per objection.
  5. 05Day 5: Wire analytics events, form routing, and a preview deploy.
  6. 06Day 6: Launch 1 variant and set the weekly conversion query.
  7. 07Day 7: Review winners, archive losers, and queue the next enrichment batch.

1. Map the offer to intent, not to opinions

Start with 3 intent bands in Clay: cold, warm, and high-fit. Use Common Room or Keyplay to score accounts, then push the top 20% into Attio with a single offer per segment. A landing page is not a brochure; it is a decision engine with 1 CTA. Use Claude 4.5 Sonnet or GPT-5.4 to compress the promise into 12 words, then validate it against 50 enriched records. If the message does not lift click-to-submit by 20% in PostHog, kill it. The winning page usually comes from 1 tight promise, 1 proof block, and 1 frictionless form.

  1. 01

    Score intent first

    Use Clay + Common Room to rank 500 accounts by firmographic fit and recent signal strength.

    Prompt: 'Classify each account as cold, warm, or high-fit using ICP, visit recency, and funding signal.'
  2. 02

    Write 1 offer per segment

    Use Claude 4.5 Sonnet in Cursor to turn 3 pains into 1 outcome line and 1 CTA line.

    Output: 'Book a 15-minute teardown' for warm traffic, not 'Learn more'.
  3. 03

    Route into CRM

    Push high-fit leads from Clay into Attio via n8n and tag source, segment, and page variant.

    Decision rule: if intent >= 70/100, create an Attio deal and assign an 11x task.
IfThen
you need more than 1 CTAyou do not have 1 offer; collapse the page to 1 action and 1 proof point.
Clay fit score is below 60/100do not show the pricing or demo CTA; send to a lighter educational page.

2. Build the page in Framer, then harden it in Vercel

Use Relume to generate the first structure, then move it into Framer for 80% of pages. If the page needs custom logic, use Next.js on Vercel and keep every component server-light. Arcads or Runway can generate 1 UGC-style clip, but only if it supports the CTA above the fold. Use n8n to pull testimonial snippets from Attio notes and Common Room mentions into the page schema. The dead tactic is endless design review; the winning tactic is 1 draft, 1 QA pass, 1 deploy. Most pages should ship in 48 hours and hit 95 Lighthouse performance by day 3.

  1. 01

    Generate the skeleton

    Use Relume to produce the wireframe, then paste into Framer and replace placeholder copy with segment-specific lines.

    Prompt: 'Build a 5-section page for mid-market ops leaders with 1 CTA and 2 proof blocks.'
  2. 02

    Add proof automatically

    Use n8n + Attio to inject case study quotes, logos, and role titles from approved records.

    Decision rule: if proof is older than 18 months, archive it and replace with fresher Common Room evidence.
  3. 03

    Ship and monitor

    Deploy to Vercel, connect PostHog events, and watch scroll depth plus CTA clicks for the first 72 hours.

    Example event: 'cta_clicked' with properties page_variant='A' and intent_band='warm'.
IfThen
a section cannot be explained in 1 sentencedelete it; complexity kills mobile conversion.
Framer constraints slow a custom interactionmove that interaction into Next.js on Vercel instead of forcing a workaround.

3. Personalise only where intent pays

Do not personalise everything; personalisation below 1,000 monthly visits is usually theatre. Use Mutiny or Intellimize for 1:1 hero swaps only when Clay enrichment shows firmographic match and Common Room shows repeat engagement. For routing, use Attio plus n8n to split leads by territory, segment, or product line. If the source is paid search, let Google Performance Max feed the top-level page and reserve dynamic variants for high-fit accounts. Use Relevance AI or Lindy to write follow-up summaries and hand them to 11x within 5 minutes. This is where operator speed wins: 1 page, 3 variants, 1 routing graph.

  1. 01

    Set the personalisation gate

    Use Clay enrichment to require firmographic match, role match, and signal match before a page swap occurs.

    Prompt: 'Only personalise if industry=fintech, employee count 200-2000, and visit recency < 7 days.'
  2. 02

    Swap only the hero

    Use Mutiny to change headline, proof logo, and CTA; keep the rest of the layout fixed.

    Decision rule: if variant count exceeds 3, stop and run an A/B test instead.
  3. 03

    Automate follow-up

    Use Lindy or 11x to summarise form context, create the Attio task, and send the first reply.

    Workflow: 'new_form_submit' -> Attio record -> 11x draft -> human approve -> send.
IfThen
monthly traffic is under 1,000 sessionsuse static copy and one proof block; skip Mutiny until data exists.
lead score is above 75/100route to 11x immediately and suppress nurture emails for 24 hours.

4. Instrument the page like an experiment, not a vanity asset

PostHog is the source of truth for behaviour; Northbeam is the source of truth for paid spend; Attio is the source of truth for pipeline. Track 6 events: page_view, hero_scroll, proof_click, form_start, form_submit, and booked_meeting. Use Cursor or Claude Code to write RevOps SQL that joins source, variant, and lifecycle stage weekly. If a variant lifts form submits by less than 10% after 300 sessions, archive it. If one testimonial increases scroll depth by 15%, pin it. Every page should produce 1 weekly learning memo generated by an MCP server that reads PostHog, Attio, and Clay outputs.

  1. 01

    Define event names once

    Lock the schema in PostHog and mirror the same names in n8n, Attio, and your dashboard.

    Events: page_view, form_start, form_submit, booked_meeting, with variant_id on every event.
  2. 02

    Run weekly SQL

    Use Cursor with Attio exports to calculate conversion by traffic source, intent band, and page variant.

    Prompt: 'Write SQL to compare variant A vs B on booked_meeting rate by source_medium.'
  3. 03

    Feed the agent loop

    Use an MCP server to summarise losers, identify winning proof, and open the next Clay enrichment batch.

    Decision rule: if a proof block wins 2 weeks in a row, promote it into the default template.
IfThen
you cannot attribute a conversion to a variantyour event schema is broken; fix tracking before changing copy.
paid traffic CPA rises 20% for 7 dayspause the lowest-intent source and reallocate spend to the best page variant.
Pitfalls

What goes wrong

Mistake

Building 5 pages before validating 1 offer

Fix

Use Clay and PostHog to test 1 page, 1 CTA, and 3 headlines first.

Mistake

Letting design lead copy

Fix

Write the CTA in Claude 4.5 Sonnet before Framer opens.

Mistake

Personalising every visitor

Fix

Mutiny or Intellimize only after intent crosses 70/100.

Mistake

Routing leads manually in spreadsheets

Fix

Use Attio + n8n and let 11x handle first-response drafting.

Mistake

Tracking vanity clicks instead of meetings

Fix

Optimise for booked_meeting and qualified_lead, not button hover.

FAQ
When should I use Framer instead of Next.js on Vercel?▾

Use Framer for 80% of landing pages because speed beats engineering ceremony in 2026. Move to Next.js on Vercel only when you need custom routing, server-side logic, or deep component reuse. If the page cannot ship in 48 hours inside Framer, the issue is usually scope, not tooling.

How much personalisation is enough?▾

Use 1:1 personalisation only when traffic and intent justify it. If you have fewer than 1,000 monthly sessions, static copy plus 1 proof block is usually the better trade. Clay, Mutiny, and Common Room should trigger personalisation only when the fit score is above 70/100.

What is the best CRM for this system?▾

Attio is the default in 2026 because it stays light and lets operators move fast. Salesforce is a tax on your GTM unless you already have the team to absorb it. Pipe everything through n8n, then let 11x or Lindy create the first response and task trail.

How do I know a landing page is actually working?▾

Track booked meetings, qualified lead rate, and source-level conversion in PostHog and Attio. If a variant improves form submits by less than 10% after 300 sessions, it is noise. Northbeam should confirm whether paid traffic is buying real pipeline or just clicks.

What is the fastest way to improve conversion?▾

Change the offer before you change the pixels. Use Clay to segment the audience, Claude 4.5 Sonnet to rewrite the promise, and Relume plus Framer to ship the new page the same week. Most gains come from tighter intent matching, not prettier gradients.

Get one system like this in your inbox each week

No fluff, no upsell. Unsubscribe anytime.