Your landing page’s weakest score is Trust.
How Everest ranks
68Growth scoreupdated
category scores
category scoresShowHide
Score over time
Trust
58FairProof that the product is real, competent, and safe to try.
| Trust read | |
|---|---|
| Everest | "benchmarked on DeepSWE" panel showing cost $0.412 → $0.243 (-41%) and latency 6.9s → 6.2s (-10%) |
| Author list with "Microsoft Corporation" affiliation and {hjiang, qianhuiwu, xufang.luo}@microsoft.com emails plus linke | |
| Detailed analytics dashboard screenshot showing "Sum of Cost by Model Over Time — $445.95" with per-model usage table (o | |
| Stat bar: "400T+ Monthly Tokens · 10M+ Global Users · 80+ Providers · 500+ Models" | |
| "A PALO ALTO NETWORKS COMPANY" tag under the Portkey logo |
Mobile web
58FairThe phone experience as a first-class citizen.
123- Compression feature box — Bullet text 'optimized for achieving the' is cut off at the right edge — the box overflows the viewport horizontally.
- Truncated curl command — The install command 'curl -fsSL https://install.ev…' is clipped mid-URL, undermining trust in the one-click install pitch.
- Start with $30 credit — The primary CTA sits far below the first viewport as small right-aligned bracketed text instead of a full-width tappable button.
CTA
68FairCopy, size, verb, cost-feel — does the primary CTA earn the click?
Problems
- '[ COPY ]' promises no payoff
- $30 credit offer buried bottom-right
- command truncated on mobile
- no benefit copy near submit
Information hierarchy
76FairDoes the design help or inhibit what the page is trying to say?
123- Start with $30 credit — The likely conversion CTA is stranded bottom-right in low-emphasis bracket styling, outside the scan path and weaker than the [ COPY ] button.
- Two equal terminal panels — The compression panel and DeepSWE benchmark box sit at identical visual weight, so the strongest proof (-41% cost) doesn't outrank the feature list.
- Benchmark delta labels — The -41% and -10% savings — the core promise — are rendered at the same small mono size as row labels instead of being the loudest numbers on the page.
Value props
78FairTitle, subtitle, hero — do they sell the product's true value?
Everest’s headline sells
everest saves your bills on the best closed-source models for coding.
Working
- Names real outcome: lower model bills
- Hard numbers: -41% cost, -10% latency
- Hero demos the product, not decoration
Could improve
- No subhead translating how it works
- Jargon bullets: BYOK, intent-based compression
- 'saves Everest's bills' phrasing is awkward