Why pricing pages are hard
Because they don't present information. They help someone decide.
A mediocre feature page just fails to impress. A mediocre pricing page leaves the visitor stuck on "which one am I supposed to buy" — and they leave, at the exact moment they were closest to paying.
Ask an AI for a pricing page and the default output is always the same shape: three cards, the middle one slightly taller, each with a column of ticked feature names. It looks professional and it does nothing to help anyone decide.
Four typical failures
1. Feature lists written in internal jargon
The model lists what you gave it, which is usually your feature names:
- ❌ "Webhook integration support"
- ✅ "Get notified in your own system the moment an order changes status"
Nobody buys features. They buy outcomes.
2. The tiers aren't distinguishable
Default output: Basic has 5 features, Pro has 8, Enterprise has 12. The visitor has to diff three lists to find the difference.
The fix is to lead with who each tier is for:
Starter: one person's project Team: 3–20 people who need collaboration and permissions Enterprise: you need SSO, audit logs, self-hosting
One line each. The feature list becomes supporting evidence, not the argument.
3. Objections go unanswered
Four questions surface simultaneously on a pricing page, and default AI output answers none of them:
- Can I cancel?
- What happens if I exceed the limit?
- Are there hidden fees?
- Can I try it first?
Leave these unanswered and no amount of feature listing helps.
4. The "most popular" badge is decoration
The badge does real work — it reduces decision cost. But AI puts it on the middle tier purely because the template has it there.
Tell it which tier to recommend, and why.
A pricing page prompt that works
Page: SaaS pricing page
Tiers (use exactly these; do not add or remove):
- Starter $19/mo: for one person's project
- Team $49/mo: for 3–20 person teams needing collaboration
and permissions (mark "most popular")
- Enterprise, contact us: needs SSO, audit logs, self-hosting
Structure:
1. Heading + one line explaining the pricing model (what is billed)
2. Monthly/annual toggle, annual labelled "2 months free"
3. Three cards, each: tier name → one-line "who it's for" →
price → CTA → feature list
4. Detailed comparison table for people who want to diff
5. FAQ, which must cover: cancellation, overage handling,
hidden fees, trial availability, upgrading and downgrading,
accepted payment methods
6. Closing CTA
Copy rules:
- Describe features as outcomes for the visitor, no internal jargon
- Invent no numbers, client names or "used by X companies" claims
- Use exactly the prices above; do not adjust them
Style: see the design spec block. Three cards equal height and aligned;
emphasise the recommended tier with a border rather than by scaling it up
(scaling breaks the alignment).
That final note comes from a specific failure: AI loves to scale the recommended card 5% larger, which misaligns the card tops and looks especially bad on mobile.
Three checks before launch
- Show it to someone unfamiliar for 10 seconds and ask which tier they'd pick. If they can't answer, the differentiation isn't clear.
- Look at it on a phone — how do three columns behave on a narrow screen? Horizontal scroll, stacked, or collapsed? The default is usually a cramped mess.
- Click every CTA and confirm each goes to the right checkout link.
The third sounds trivial. It's also the most common real bug in AI-generated pricing pages: all three buttons pointing at the same href.
MotionSites has several polished pricing-page prompts, comparison table and FAQ structure included.