Skip to content

Answer · Conversion Engineering

How Do I Write a Good CTA?

14 days kickoff → live $3K–$15K+ scope-tiered WCAG 2.1 AA baseline

The short answer

Name the outcome, not the action. “Get my 5-day audit” beats “Request audit.” “Book my 20-min review” beats “Schedule a call.” Use first-person possessive (“my”) and add specificity (timeframe, deliverable). CTA copy is 5-10x more impactful on conversion than CTA color.

№ 01The longer answer

The CTA tells the visitor what they get, not what they do. Generic CTAs (“Get Started,” “Submit,” “Learn More”) underperform specific CTAs by 10-25% on B2B pages. The fix is one sentence of copy work, not a design overhaul.

The pattern that consistently works: first-person possessive (“my,” “our”) + outcome + specificity. “Get my pricing” outperforms “See pricing” by 4-9%. “Send my project brief” outperforms “Contact us” by 6-12%. The pronoun makes the action feel owned; the specificity reduces ambiguity.

Tier-match the CTA to the page. On a $50K Authority-tier case study, “Talk about a similar build” beats “Start free trial.” Buyers reading high-ticket content aren’t trial buyers. The CTA should match the deal size implied by the page.

Two CTAs (primary + secondary) consistently beat one. Primary: the high-commitment ask (“Get my audit”). Secondary: the low-commitment escape valve (“See pricing first” or “Read a case study”). The secondary catches not-ready buyers who’d otherwise bounce.

Primary CTA is always a button (high visual weight). Secondary CTA can be a button outline or a text link. Both buttons being equal weight forces the visitor to decide between two equally-prominent options, which adds friction.

№ 03How long should the CTA copy be?

2-5 words on the button itself. Longer copy belongs in reassurance text below the button. “Get my 5-day audit” on the button + “No sales calls. 5-day delivery. Refundable.” below.

№ 04Should I A/B test CTA copy?

Yes, if you have the traffic. Copy tests need less sample size than design tests because the expected lifts are bigger (10-25% vs 2-5% for color tests). Higher MDE = smaller sample required.