FulcrumYC S25 landing page report.
← Back to YC S25

Your landing page’s weakest score is Trust.

Fix Trust

Score over time

90807060504030METR 72 on Jun 28 (Wayback)METR 75 on Jul 13 (Wayback)METR 72 on Jul 25 (Wayback)METR 73 on Aug 6 (Wayback)METR 72 on Aug 21 (Wayback)METR 72 on Sep 7 (Wayback)METR 73 on Sep 20 (Live)Apollo Research 44 on May 1 (Wayback)Apollo Research 41 on May 10 (Wayback)Apollo Research 44 on May 25 (Wayback)Apollo Research 46 on Jun 17 (Wayback)Apollo Research 68 on Aug 21 (Wayback)Apollo Research 68 on Sep 4 (Wayback)Apollo Research 72 on Sep 22 (Live)Epoch AI 72 on Jun 16 (Wayback)Epoch AI 72 on Jul 6 (Wayback)Epoch AI 74 on Jul 26 (Wayback)Epoch AI 71 on Aug 6 (Wayback)Epoch AI 74 on Aug 14 (Wayback)Epoch AI 74 on Sep 5 (Wayback)Epoch AI 76 on Sep 20 (Live)Transluce 59 on May 14 (Wayback)Transluce 56 on May 23 (Wayback)Transluce 57 on Jun 23 (Wayback)Transluce 57 on Aug 2 (Wayback)Transluce 62 on Aug 25 (Wayback)Transluce 59 on Sep 9 (Wayback)Transluce 61 on Sep 22 (Live)Fulcrum 59 on Sep 22 (Live)Sep 22

Trust

48Poor

Proof that the product is real, competent, and safe to try.

Fulcrum
Epoch AI
METR
Apollo Research
Transluce
Trust read
Fulcrum"Highlighted Work" benchmark chart supporting "Fable is SOTA at CIFAR Speedrun"
METR"Non-public information and model access provided by:" with Anthropic, Google DeepMind, Meta, and OpenAI logos on the Fr
Apollo Research"OUR PARTNERS" logo strip naming AI Security Institute, Microsoft, Meta, Anthropic, OpenAI, Amazon, NIST, DeepMind, and
Epoch AI"Trusted by leading publications and institutions" logo strip (Our World in Data, DSIT, Government of the Netherlands, N
Transluce"Our Work" cards showing real research output, including the essay "Some Focus Areas for Embedded Evaluations and How to

CTA

58Fair

Copy, size, verb, cost-feel — does the primary CTA earn the click?

Fulcrum
Apollo Research
Epoch AI
METR

Problems

  • small, muted button lacks visual weight
  • 'updates' is vague — no named payoff
  • sits below nav links, easy to miss
  • no supporting microcopy on the form

Value props

58Fair

Title, subtitle, hero — do they sell the product's true value?

Fulcrum’s headline vagueFulcrumWe work on measuring and optimizing model performance in fuzzy domains.

Working

  • Honest, no hype or buzzword stack
  • Names a real gap: measuring fuzzy traits
  • SOTA chart gives concrete proof nearby

Could improve

  • 'Fuzzy domains' unclear to newcomers
  • No outcome stated for the reader
  • Hero numbers live in the chart, not copy

Information hierarchy

62Fair

Does the design help or inhibit what the page is trying to say?

Information hierarchy annotations on the desktop capture1234
  1. Get updates button — The only conversion action is a small, low-contrast gray button buried below three paragraphs and a nav row — invisible in a one-second scan.
  2. Nav link row — Five equal-weight underlined links (research, writing, careers, contact, x) sit directly above the email capture, giving attention five exits before the CTA.
  3. Body paragraph stack — Three same-weight paragraphs follow the wordmark with no headline hierarchy — nothing tells the eye which sentence is the value proposition.
  4. Highlighted work chart — The dense annotated chart with tiny axis labels carries more visual weight than the CTA and is unreadable at this size.

Mobile web

68Fair

The phone experience as a first-class citizen.

Mobile web annotations on the mobile capture123
  1. Get updates button — Below the first viewport and desktop-proportioned; should be full-width and reachable without scrolling.
  2. Highlighted Work chart — Axis labels and annotations render at unreadable sizes when the desktop chart shrinks to phone width.
  3. Footer link row — Five inline text links with tight gaps make imprecise thumb targets instead of a stacked or collapsed menu.

Competitors

inferred from product, audience, and value prop