Test messages, not landing pages
Four landing-page variants on a small budget produce four uninterpretable results instead of one useful one.
The instinct is understandable: you have four ways to describe the product, so you build four pages and let the market pick. On a limited budget this turns one readable result into four thin ones.
The arithmetic of splitting traffic
Detecting a landing-page conversion difference needs traffic per arm, and conversion events are expensive because they're bought with clicks:
| Comparison | Visits per variant | Media at $1.50 a click |
|---|---|---|
| 4% vs 8%: a doubling | 553 | $1,659 |
| 5% vs 10% | 435 | $1,305 |
| 5% vs 7.5% | 1,471 | $4,413 |
Those are per-arm figures, for effects far larger than most copy changes produce. A $150 budget split 4 ways gives each page about 25 visits.
Ad messages are measured on impressions, and impressions are cheap
Move the comparison up the funnel. Separating 0.8% from 1.6% click-through takes about 2,900 impressions per message, costing roughly $87 at a $15 CPM. The tighter read costs about $193.
The design is several messages, one page. Each message makes a different bet. Click-through shows which bet earns attention.
It also answers the more useful question
“Nobody wants this” and “nobody wants it described this way” demand different fixes. If every angle fails, change the proposition. If one wins, use it.
Landing-page variants come later, when the budget supports them. Until then: one strong page, several honest messages.