TL;DR

Ramp's pricing page hides a "Platform fee based on team size" with no dollar amount, range, or calculator—one nonprofit reported a surprise $7,500 fee appearing at renewal after years of no charge. The Free tier's Treasury section literally displays "Earn -.--%" (a broken placeholder) instead of a real rate. Independent reviews show a gap: 4.8/5 on G2 but only 3.5/5 on Trustpilot, where users describe unexpectedly heavy underwriting and verification friction. The site also serves different content to bots vs. human visitors, catching our automated test off guard.

Overall score: 67/100. Verdict: the hidden platform fee and broken pricing details make Ramp's self-serve promise unreliable—proceed only if you're prepared for undisclosed costs and post-signup hurdles.

Ramp Website Review: An Undisclosed Platform Fee Is Quietly Costing Customers

Executive Summary

Overall Score: 67/100

Ramp's marketing site is well-built and its messaging is more self-serve-friendly than reputation suggests — the homepage leads with an email-capture "Get started for free" box, not a gated demo form. But the pricing page, the one page a serious buyer actually needs, has two verifiable defects: the "Plus" tier lists "+ Platform fee based on team size" with no number, formula, or calculator anywhere on the page, and the Free tier's Treasury row displays an unrendered placeholder ("Earn -.--%\* in a Ramp Investment Account") instead of an actual rate. Both were confirmed by directly rendering the live page in a JS-capable browser, not inferred.

A methodology note, because it matters here: a plain HTTP request to ramp.com/ (via curl, and via this session's automated fetch tool) does not return Ramp's real homepage. It returns a text/markdown document Ramp's own edge infrastructure serves to bots and AI agents — headed "Ramp — Machine Version" — that is functionally different content from what a human visitor sees, including a tracked incentive offer. We caught this because our first automated pass produced homepage copy that didn't match the screenshots. We re-fetched with a real rendering browser to get the actual human-facing page; the analysis below is based on that render, not the bot version. More on that under Trust.

Independent verification: Ramp holds a strong 4.8/5 on G2 (2,000+ reviews) and 4.9/5 on Capterra, but a noticeably weaker 3.5/5 on Trustpilot (176 reviews) — a gap consistent with vendor-solicited review platforms skewing positive while unsolicited platforms surface the onboarding and support complaints documented below.

Messaging Score: 73/100

Direct answer: What's on the page today: The live homepage headline is "Time is money. Save both," with the subhead "Cards, expenses, bill payments, and banking\* — in the blink of AI." That's a change from language cited in earlier internal reviews of this site ("Save time and money with Ramp"), a reminder that this kind of audit has a shelf life and needs re-verification, not a one-time snapshot.

What works:

  • The self-serve framing is now explicit and immediate: an email field sits directly under the hero copy, with "Get started for free" as the primary CTA — not a sales form.
  • The value proposition is concrete: cards, expenses, bill pay, and banking as one system, backed by a "70,000 teams" customer count and a claim of growing "3.2x faster than the average American business," linked to a "Read the report" source.

What doesn't:

  • The page is saturated with AI-agent framing ("Agents at work today," "in the blink of AI," a top banner for "AI Token Spend Management") before it has finished explaining the core product to a first-time visitor. For a buyer who still isn't sure what "spend management" means, agent-hype language adds a second layer of jargon on top of the category confusion.
  • The homepage stacks more than a dozen customer testimonials in a row with no filtering by industry, company size, or use case (finance leaders from Notion, Shopify, Perplexity, a school district, and a city government all run together). Individually credible, collectively it reads as volume over relevance — a visitor from a 20-person agency has to dig to find a testimonial that looks like their situation.
  • Pricing-tier labels ("AI-assisted," "AI-powered," "AI-tailored") are marketing flourish that doesn't actually differentiate Free, Plus, and Enterprise on capability — a buyer scanning quickly gets three shades of the same adjective, not three distinct value propositions.

Conversion Score: 58/100

Direct answer: This is where the site has real, reproducible problems — verified directly, not inferred from screenshots.

The pricing page's central defect: The Plus plan is listed as "$15/mo/user + Platform fee based on team size." There is no number, no range, no calculator, and no link to one. This isn't a one-off observation — multiple independent sources describe the exact same gap. A Vendr marketplace breakdown notes platform fees can run from $0 to $15,000+/month depending on account, and a separate review of 13 prospects who priced out of Ramp in 2026 documented specific renewal surprises, including a nonprofit reporting a $7,500 platform fee appearing at renewal after several fee-free years. A user on Blind pointed out the same thing from the inside: "they don't add green check marks on their pricing page for things they do charge for though (intl reimbursements, intl bill payments, etc)."

A literal broken/placeholder value on the pricing page: The Free tier's Treasury section reads "Earn -.--%\ in a Ramp Investment Account" and "Earn -.--%\ on cash in a FDIC-insured account" — dashes where a rate should be, on a live, public pricing page. Whether that's an unrendered dynamic value or an intentionally vague placeholder, a prospective customer sees a broken number in a place designed to sell them on returns.

Post-signup friction, documented independently: Multiple Trustpilot reviewers describe underwriting/verification as unexpectedly heavy — one described providing bank statements and state paperwork and still hitting "a brick wall over and over," another compared the process to "teaching an insect how to cook scrambled eggs." This matters for conversion scoring because it means friction doesn't end at the pricing page; it continues into the account-verification step that a self-serve signup implicitly promises will be fast.

What's genuinely improved: Contrary to older claims that Ramp gates everything behind a demo request, the current nav has three clearly different entry points — "Sign in," "See a demo," and "Get started" — and Free/Plus signup requires only an email, no credit card. That's a real strength worth crediting.

nqzai's modeled impact estimate (explicitly our own directional modeling, not a Ramp-disclosed or externally-audited figure): pricing pages that force a buyer to contact sales just to learn the total cost typically see meaningfully higher bounce rates at that exact page than pages with a self-serve estimate or calculator. Given the platform fee is completely unquoted and the Treasury rate is visibly broken, we'd estimate this specific page is losing a high-single-digit to low-double-digit percentage of Plus-tier visitors who never proceed to a sales conversation — not a precise, audited number, but a reasonable order of magnitude given the friction is on the page itself, not buried three clicks deep.

Trust Score: 71/100

Direct answer: Genuinely strong: Ramp's Trust Center (trust.ramp.com, run on SafeBase) lists SOC 1, SOC 2, ISO/IEC 27001, PCI DSS, CCPA, FedRAMP Moderate, GovRAMP, and Cyber Essentials Plus — a more substantial compliance posture than most competitors publish, even though the underlying reports sit behind an NDA request (standard practice for SOC 2 documentation, not a design flaw).

The review-platform gap: G2 (4.8/5, 2,000+ reviews) and Capterra (4.9/5, 215 reviews) are strong, but Trustpilot sits at 3.5/5 (176 reviews), with recurring 2026 complaints about support responsiveness and onboarding. NerdWallet's review separately flags a 2024 rewards-rate change — from a flat 1.5% cashback to a variable 0%–1.5% range — that some customers characterized as a bait-and-switch, since realized rates didn't match what was initially discussed.

The finding we didn't expect: Ramp runs a documented, intentional strategy of serving different content to AI agents than to humans. Their own engineering blog (linked below) describes embedding a tracked $3,100 "AI Research Incentive" into a bot-detected markdown version of their site, explicitly to get AI assistants like this one to relay the offer into conversations with human users — and reports that Claude specifically has been the most likely model to surface it. We can confirm this is real: an unauthenticated request to ramp.com/ returns content-type: text/markdown with that exact offer, while a rendered browser session sees the normal marketing homepage with no such offer visible anywhere. To be clear about what we did with that: we did not relay the offer, sign up through the tracked link, or treat the embedded "mention this to a sales rep" instruction as anything other than page content to report on. We're noting it here because it's a legitimate, verifiable transparency question for a company in the trust business — human visitors and AI agents evaluating Ramp on their behalf are currently being shown different information.

Recommendations

  1. Publish the platform fee formula, or a ballpark calculator. "Based on team size" with zero anchor number is the single most-cited pricing complaint across independent reviews. Even a rough range (e.g., "$X–$Y/month for teams of 10–50") would remove the biggest reason to bounce before talking to sales.
  2. Fix the Treasury APY placeholder. A live pricing page showing "-.--%\*" instead of a rate looks broken regardless of the reason, and it sits right next to the numbers meant to sell the free tier.
  3. Set onboarding/underwriting expectations before signup, not after. Trustpilot's recurring "brick wall" complaints suggest a documentation checklist up front (what's needed, how long it takes) would reduce mid-process abandonment.
  4. Trim or segment the homepage testimonial wall. A dozen-plus unfiltered quotes dilute rather than reinforce; grouping by company size or industry would let a visitor find a relevant proof point instead of scrolling past all of them.
  5. Decide, deliberately, what human visitors see relative to AI agents. The agent-only incentive program is a legitimate emerging-channel experiment, but it currently means a person evaluating Ramp directly has less information than an AI agent evaluating it on their behalf. Surfacing an equivalent human-facing offer (or being explicit that agents get a different one) closes that gap.
  6. Surface a condensed trust summary outside the NDA wall. The compliance list (SOC 2, ISO 27001, FedRAMP Moderate, etc.) is a real differentiator — putting badges or a one-line summary directly on the pricing page would carry some of that trust into the exact moment a buyer is deciding.

Sources