How Ramp built the fastest-growing SaaS and fintech business

Executive overview

Most B2B growth failures come from investing in expensive channels before understanding customers. Ramp reached $100M ARR in two years by sequencing channels cheaply first, then layering in data and automation to make each one more efficient than competitors.

Growth leader Sri Batchu argues the secret isn't a unique channel mix — it's bringing engineering and data to every channel before rivals do. The competitive moat is not which channels you use, but how technologized each one is.

Early growth tactics

• Cap table as growth strategy: co-founders recruited early-stage founders and operators onto the cap table, who became first customers
• VC introductions helped, but peer customer referrals drove more conversions than investor recommendations
• New customer acquisition — not expansion revenue — drove the majority of growth then and still does today
• Fundraising announcements doubled as market moments, generating measurable top-of-funnel awareness
• PR works when you have a genuinely interesting story; without one, it wastes resources

Growth team structure

• Channel-based teams: paid marketing, lifecycle/CRM, field marketing, SEO/website
• Growth engineering team dedicated to sales efficiency — uses AI and third-party data to automate prospecting, lead scoring, messaging, and response drafting
• Engineers on the sales support team own pipeline and payback metrics, not product metrics — this creates direct accountability for revenue
• Small skunk works / innovation team tests new channels (TikTok, Reddit), referral mechanics, first-party events
• Self-serve activation sits within each channel team

Velocity culture

• Internal tool days.ramp.com shows elapsed days (with decimal precision) since founding — displayed at every board meeting and all-hands to reinforce daily urgency
• Two-week sprints across the full growth team; channel teams cross-prioritize together rather than operating in silos
• Responsiveness on Slack is visible and fast; clear ownership and deadlines assigned even before tasks are complete
• Calendar audits — personal and team-level — used to align time allocation with stated priorities ("you ship your calendar")
• Engagement during all-hands tracked by the IT team; near 100% chat participation reflects genuine excitement rather than obligation

North star metrics and translation layers

• Good north star metrics sit between "too lagging" (revenue) and "too narrow" (team-specific): they must be intuitive and show clear linkage to value
• At Instacart: monthly active orders (MAOs); at Ramp: dollars of SQL pipeline
• Each sub-team tracks its own direct metric (e.g., app load time, checkout conversion), with a translation factor converting it to the north star
• Translation factors updated every six months based on new data; used for resource allocation and cross-team prioritization, not marginal decisions
• Long-term holdouts (small control groups kept on an older experience) validate cumulative impact of each team's work over time

Payback period over CAC

• CAC encourages cutting acquisition cost, which often means attracting lower-value customers
• LTV/CAC is hard to rely on for young companies — churn assumptions are too uncertain and errors compound
• Payback period on contribution margin is preferred: how many months of profit (after variable costs) to recover acquisition spend
• Sets a concrete, near-term target the whole team can rally around; executives and board set the acceptable threshold

B2B channel sequencing

1. Founder-led sales — team must learn to sell directly
2. First sales hires
3. Low-cost targeted marketing: content, community, small-scale events
4. PR
5. Paid and brand efforts
6. SEO (starts around the same time as paid — but launched later because domain authority takes time to build)

Channels become more expensive and more effective as you move down the list; sequencing matches cost to customer knowledge and scale.

Experimentation principles

• Growth experiments succeed roughly 30% of the time — failure is expected; the failure is not learning
• Fail conclusively: design tests so a negative result actually closes the question
• In B2B, sample sizes are small; compensate by maximising the treatment effect — deploy all the tactics you think could work in a single test
• If the maximally-resourced version doesn't work, the hypothesis is disproved; if it works, strip it back to find the efficient version
• "Fail conclusively" matters most for expensive, cross-functional tests; fast, cheap tests (email copy, web conversion) can iterate freely

Hiring and compensation

• Two approaches to sourcing: (1) network mapping — ask who's the best at X, get intros, build a target list of people one to two stages ahead; (2) data-driven sourcing — use tools like SimilarWeb to identify which companies actually perform well in the function you're hiring for, then recruit from those teams
• Target companies one to two growth stages ahead — they've seen your stage and beyond
• Strong opinion: compensation bands should not be equalised; 10x operators should earn 10x comp
• Talent density depends on retention and performance management as much as hiring — managing out poor fits quickly is how companies maintain their bar
• Interview false-positive rate is high; invest in people once they prove themselves internally

Prioritisation and problem-solving

• MECE framework (mutually exclusive, collectively exhaustive): when attacking a problem, decompose it at each layer until all possible causes are covered — prevents missing important variables and builds confidence that the solution set is complete
• Sprint planning consolidated across product and marketing teams eliminates misaligned cycles and speeds execution
• Use Airtable (or equivalent) for sprint scoring with translation-layer metrics; use external tools (e.g., Mutiny for website personalisation) rather than building in-house unless it's proprietary and strategic