How Mercury Built a $650M Bank on Word-of-Mouth and a Glossary

Here's the uncomfortable truth about Mercury, the $650M-revenue startup bank: its growth engine has almost no salespeople in it.

Founder and CEO Immad Akhund says it plainly: "Even today, our sales team does about 5% of our accounts. Most people just self-service." Roughly 80% of customer acquisition is word-of-mouth. Nearly half of customers first heard about Mercury from a friend, a fellow founder, or an investor.

And yet underneath that founder-network growth sits a quiet, ruthless content machine printing the equivalent of $2.54M every month in Google Ads traffic — for free. Two engines. Most teardowns only see one. Let's pull both apart.

By the numbers

Mercury is one of those rare fintechs where the financials aren't a vanity slide.

• ▮$650M annualized revenue as of September 2025, with three-plus consecutive years of GAAP profitability — almost unheard of in fintech.
• ▮300,000+ customers, growing ~50% year-over-year. One in three U.S. startups bank with Mercury.
• ▮$248B in transaction volume in 2025, up 59% from $156B the year before.
• ▮NPS of 73.8 against a banking-industry average of ~34. That gap is the word-of-mouth.
• ▮Funding: a $300M Series C led by Sequoia at a $3.5B valuation, reportedly ~$5.2B by May 2026.

On the marketing side, the organic footprint is just as loud: ~249,123 monthly organic visits, 12,534 ranked keywords, 294 sitting at #1, and a paid-search budget of about $1,500/month. Read that again. A bank doing $650M in revenue spends roughly the price of a used MacBook on Google Ads. This is an organic company.

The engine

The real growth story starts with patience most founders don't have.

Akhund spent ~18 months building before launch. His line: "I don't think you can iterate yourself to a new bank." While Silicon Valley shipped MVPs in three months, he refused — because trust in a bank can't be A/B tested into existence.

The payoff was immediate and absurd. "Four days after launch, someone signed up, never spoke to us, and transferred $1 million into a Mercury bank account." That's the entire flywheel in one sentence: a self-serve product so obviously good that a stranger trusts it with seven figures, unprompted. From there, 30-40% month-over-month growth for the first year.

Then came the accelerant nobody plans for. When SVB collapsed in March 2023, Mercury onboarded ~8,000 customers in a week or two and took in over $2 billion in deposits. They shipped "Mercury Vault" — expanded FDIC coverage — over a single weekend ("started on it Saturday and shipped Monday"). Akhund called it "by far, the biggest thing that ever happened in terms of signups," and noted the strangest part: "SVB was our main competitor." A crisis handed them the market.

But here's what makes Mercury durable rather than lucky: it didn't stay a tech-startup tool. By end of 2025, 73% of new customers came from outside the AI/tech core — ecommerce at 21% of new customers, professional services climbing, personal banking launched. The 2025 Annual Letter: "Customers don't rely on Mercury for just one task — they use cards, payments, invoicing, and more." Multi-product depth turned a checking account into a platform.

The stack

This is where most analysts get fooled. A tag scan of mercury.com finds almost nothing: Next.js and React on the front end, AWS plus an ALB, Cloudflare CDN and Bot Management, and Cloudflare Browser Insights as the only analytics. No Google Analytics. No Segment. No HubSpot. No Marketo firing in the browser.

The naive read is "unsophisticated." The correct read is the opposite. As a regulated bank, Mercury keeps client-side third-party scripts minimal for compliance, privacy, and performance — and moves the GTM machinery server-side, into the warehouse. The public scan understates their sophistication. The hiring board is the real tell.

And the hiring board screams. 155 open roles, headcount up +22.3% in 12 months, and a GTM org being rebuilt in three directions at once:

• ▮An outbound sales engine from scratch: SDRs (including "SDR - New Grad"), a Sales Development Manager, multiple Account Executives, and a Head of Revenue Enablement. You don't hire an enablement head unless you're formalizing a real sales org on top of the old self-serve funnel.
• ▮A data-first RevOps core: a Revenue Technology - Data Strategy & Operations Lead, Revenue Strategy associates, a Salesforce Developer (Salesforce is the confirmed system of record), and Strategic Finance - GTM roles — dedicated FP&A for marketing spend. The instrumentation lives in the warehouse, exactly matching the lean browser footprint.
• ▮A channel play: Strategic Partnerships - Accounting, Senior Relationship Managers for accounting partners, Implementation Managers. The accountant/advisor referral channel, productized.
• ▮Brand as a moat: a Head of Growth Marketing AND a Head of Brand Marketing, plus an Executive Creative Director and multiple Staff Brand Designers. That's unusual creative weight for a fintech — a bet that brand itself generates demand.
• ▮The forward bet: a cluster of "API & Agentic Banking" PMM/PM/engineering roles, an AI Solutions Architect, and an AI Enablement engineer. Mercury is positioning for AI-agent and embedded/programmatic banking as the next wedge. (They onboarded 2.4x more AI companies year-over-year.)

The motion is shifting from pure PLG to layered PLG + sales-assisted + partner-led — funded by the data backbone, fronted by the brand.

The clever bit

Now the part you can actually steal: the content engine.

Strip out the brand term "mercury" (368,000 search volume, ~45% of all organic traffic, homepage absorbing ~132,000 visits) and what's left is a finance/accounting glossary built to match exactly what a founder Googles.

Look at the pages: /blog/assets-vs-liabilities-vs-equity (~7,719 visits), /blog/analyze-profit-loss-statement (~7,435), /blog/retained-earnings-formula-calculate (~5,960). They rank #1 for "p&l profit and loss" (18,100 vol) and #1 for "assets liabilities equity." ACH vs wire, gross margin, COGS, unit economics, opportunity cost, cost-benefit analysis — the entire vocabulary of a startup finance operator, each post evergreen and pulling 1,000-8,000 visits.

Then the bottom-funnel layer: "best invoicing software for startups" carries a $218,000 estimated traffic value by itself. Commercial-investigation intent, parked right next to the product. Plus free tools like a profit-margin calculator — engineering as marketing.

The momentum number is the kicker: 6,966 NEW keywords vs only 1,582 declining — a ~4.4:1 expansion ratio. This isn't a library coasting on old wins. It's a factory in active expansion.

That's the clever bit. Brand demand from word-of-mouth fills the top. A glossary built around buyer vocabulary captures everyone researching the problem before they know Mercury exists. The two reinforce each other, and paid spend stays at $1,500.

What this costs you

Don't romanticize this. Mercury's playbook is expensive in ways a startup can't shortcut.

Word-of-mouth is not a channel — it's a lien on product quality. You can't "do 80% word-of-mouth." Mercury earned it with 18 months pre-launch and a product good enough that strangers wire $1M cold. If your product doesn't trigger that reaction, no growth tactic fakes it.

The SEO engine is a multi-year compounding bet. Dozens of explainers ranking #1 take quarters to mature. There's no version where you publish ten posts and capture $2.5M/month in traffic value next quarter.

The server-side, first-party data stack is real engineering. "No third-party tags" means you build the warehouse instrumentation yourself — Salesforce developers, a Revenue Technology Data Lead, a marketing data scientist. That's headcount most teams don't have.

And the SVB windfall? Pure luck. Plan as if your competitor will never collapse on your schedule.

Steal this this week

1. ▮

   Build the glossary, not the blog. List the 20 terms your buyer Googles before they know your category exists — Mercury's were "p&l," "ACH vs wire," "retained earnings." Write the definitive #1-ranking answer for each. These compound for years at near-zero marginal cost.

2. ▮

   Park one comparison page next to your product. Mercury's "best invoicing software for startups" is worth $218k in traffic value. Find the one commercial-investigation query adjacent to what you sell and own it. Bottom-funnel intent, organic price.

3. ▮

   Move analytics server-side and instrument the warehouse. Stop drowning your site in third-party tags. Mercury runs on first-party RUM and a data backbone — better privacy, better performance, and a GTM brain a tag-scanner can't even see. Start with one clean event pipeline into your warehouse.

Mercury proves the quiet engine beats the loud one: build a product worth talking about, write the glossary your buyer is already searching, and let the brand term do the rest.

Sources: Mercury 2025 Annual Letter; First Round Review (Immad Akhund interview); Fortune (May 2023, SVB coverage); DataForSEO organic + technology data; Apollo company record + job postings.