Improve conversion

Conversion work fails when people do it out of order: they polish clever copy on a page whose structure buries the CTA, or they A/B test a button color before the value proposition is even legible. This playbook fixes the sequence. Get the skeleton right, then the words, then the persuasion, then prove the lift with a real test. Each layer assumes the one below it is already solid.

Governing formula, carried from conversion-cro: Conversion = Desire minus (Labor plus Confusion). Structure removes labor and confusion, copy and psychology raise desire, analytics tells you whether you actually moved it.

When to use this

• One page or flow gets real traffic but converts poorly, and the user wants it fixed.
• The user pastes a URL or describes a hero/pricing/checkout and asks "make this convert".
• A rewrite is needed and you want copy that survives on a well-built page, not in a vacuum.
• The user is about to run an A/B test and needs to know what to change and how to measure it.

When NOT to use this (reach for instead)

• You do not yet know which stage of a multi-step funnel is leaking. Diagnose first with the

sibling fix-a-leaky-funnel playbook, then come here for the page it points to.

• The price or model is the real blocker (people understand the offer and still won't pay):

use pricing-monetization, optionally via the pricing-and-monetization-review playbook.

• There is no traffic to convert yet: acquisition first (paid-acquisition, seo-content,

organic-social, launch-gtm).

• The whole strategy is wrong (wrong audience, wrong channel, no PMF): strategy-fundamentals.

The diagnostic order (run these in sequence)

Step 1: Structure: fix the page skeleton

Pull conversion-cro. Get from it: the above-the-fold/hero diagnosis (the hero does ~90% of the work), the fixed 7-section page structure, CTA placement, trust and friction levers, and checkout or signup field surgery. Decide WHERE every element goes and whether the page passes the "do they know what you sell in 5 seconds" test before you polish a single sentence. A great sentence in the wrong slot does nothing.

Step 2: Words: write the copy that fills the skeleton

Pull copywriting-messaging. Get from it: the headline and value prop (one competitors cannot also claim), features turned into benefits, objection handling, CTA and micro-copy, and a framework (PAS, BAB, AIDA) to structure the argument. Mine real customer language so the page sounds human, not like a pitch deck. This step fills the slots conversion-cro defined; do not let it move the slots.

Step 3: Persuasion: tune the psychology

Pull marketing-psychology. Get from it: the levers that raise perceived value without changing the product, anchoring and price-display framing, social proof placement, scarcity and urgency used honestly, and CTA phrasing that lowers resistance. Apply these on top of solid structure and clear copy. Persuasion applied to a confusing page just makes the confusion louder, which is why it comes third.

Step 4: Proof: verify with an experiment

Pull analytics-data. Get from it: the right conversion event to track, an A/B test sized for real significance, and a causal read (holdout or before/after with guardrails) so you know the lift is real and not seasonal. Ship only funnel-moving changes to test; the winner becomes the new baseline. Without this step you have opinions, not a lift.

Why this order

Structure caps what copy can achieve; copy caps what psychology can amplify; psychology is only worth measuring once the first three are in place. Running them backward (test → psychology → copy → structure) means every test fights a broken foundation and you learn nothing durable.

How to adapt

• Quick win vs. full rebuild: if traffic is low, the single highest-value micro-change is

rewriting the hero header (Steps 1 to 2 only). Save the full four-step pass for pages with enough traffic to test.

• Ecommerce product/checkout page: add ecommerce after Step 1 for cart, AOV, shipping, and

scarcity-framing specifics; keep conversion-cro for the field-level checkout surgery.

• Pricing page specifically: run the pricing-and-monetization-review playbook instead, since

the number and model matter more than the layout there.

• SaaS signup/onboarding flow (not a marketing page): the "conversion" is activation; route to

product-led-growth for onboarding and time-to-value, still closing with analytics-data.

• Too little traffic to reach significance: skip Step 4's A/B test and use judgement plus

session review from analytics-data; do not fabricate significance on a tiny sample.

From God of Skills: a curated, hand-tested directory of AI skills, prompts, templates and image style guides. Source: https://godofskills.com/skills/improve-conversion?ref=claude-skill