LLenny's Podcast
← All episodes
Sri Batchu (Ramp, Instacart, Opendoor)25 June 2023

Lessons from scaling Ramp

8Frameworks
15Insights

Frameworks in this episode

Marketing4 steps

B2B Channel Sequencing Ladder

Turn on growth channels in a fixed order, cheapest-and-least-scalable first, because each later channel needs the customer knowledge the earlier ones produce.

Read →
Entrepreneurship3 steps

Cap Table as a Growth Strategy

Recruit influential founders and operators onto your cap table so they become your earliest, most credible customers and evangelists.

Read →
Finance3 steps

Contribution-Margin Payback Period over CAC

Judge acquisition ROI by how many months of contribution-margin profit it takes to repay CAC, not by CAC or LTV:CAC.

Read →
Productivity3 steps

Escape-Velocity Activation Metric

Identify the specific actions a user must take in their first window to become durably active, then make one team obsess over driving them.

Read →
Innovation3 steps

Maximize the Treatment Effect to Fail Conclusively

In low-sample B2B tests, throw every tactic at a hypothesis at once so a failure kills the idea for good instead of resurfacing for years.

Read →
Strategy3 steps

MECE Problem Decomposition

Break any problem into buckets that are mutually exclusive and collectively exhaustive so you never miss a possible cause or solution.

Read →
Strategy5 steps

North Star Translation Layer

Give every team a metric they can directly move, then convert each into a single company North Star using a shared translation factor.

Read →
Leadership3 steps

Two-Path Talent Sourcing (Network + Data)

Find the companies best at the skill you're hiring for, then find the best person there - via referral chains or via public performance data.

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Hot Take· 5

Hot Take10:30

Ramp's Edge Isn't a Magic Channel — It's Tech and Data

Sri argues Ramp's channel mix isn't meaningfully different from peers at its size; the real difference is that every channel investment is driven by technology and data. He points to sales as the example: a dedicated growth-engineering team makes the sales org unusually efficient, work they'd been doing for almost two years before AI became a buzzword.

  • Ramp's distribution across channels looks similar to comparable companies
  • The differentiator is making every investment technology- and data-driven
  • A growth-engineering team supports sales efficiency directly
  • They enrich third-party data and automate workflows with AI
  • This team predates the current AI hype by nearly two years

“I think what we've done differently is we've really focused on making all of those Investments very much driven by technology and data”

Sri Batchu · 11:00
#growth#data#sales#automation
Hot Take21:30

Hours Aren't the Problem — Autonomy and Meaning Are

Sri pushes back on the backlash against long hours, saying hard work is important for learning and impact, especially early in a career. His key learning on burnout: among the highly successful people he's worked with, hours aren't the issue — autonomy, flexibility, mission alignment, and meaningful work are. He also credits celebrating wins and building in public as counterweights.

  • Believes hard work and long hours are an important ingredient to success
  • For successful people, hours are not the source of burnout
  • Autonomy, flexibility, and mission alignment matter more than hours
  • Focus on quality of work and impact, not hours logged
  • Celebrating wins and building in public sustain the culture

“I believe in working really hard and working long hours”

Sri Batchu · 21:30

“hours are not the problem it's really you know autonomy flexibility”

Sri Batchu · 22:30
#culture#burnout#hiring#management
Hot Take46:30

Why Payback Period Beats CAC and LTV:CAC

Sri argues CAC is a flawed target because optimizing to reduce it pulls in cheaper, less valuable customers. LTV:CAC is better in theory but LTV is DCF-like, assumption-laden, and easy to over-estimate — leading you to overspend before you learn your real churn. He prefers payback period measured on contribution margin: how many months of profit it takes to recover a customer's acquisition cost, because those assumptions are recency-based and testable faster.

  • Reducing CAC can be damaging: it attracts less valuable customers
  • LTV is hard to predict and assumption-heavy, like a DCF
  • Over-trusting LTV:CAC can destroy value once real churn emerges
  • Payback period on contribution margin is the preferred metric
  • Its assumptions are recency-based and can be evaluated quickly

“when you focus on CAC and reducing CAC what tends to happen is you uh actually might be doing something very damaging where you're succeeding…”

Sri Batchu · 47:00
#metrics#unit-economics#cac#payback-period
Hot Take51:30

Don't Just Fail Fast — Fail Conclusively

Sri reframes experimentation: failure isn't the absence of revenue, it's the absence of learning, and you can only learn if you designed a test that fails conclusively. Since B2B rarely offers large sample sizes, he advises maximizing the treatment effect — throw every tactic and resource you think could move the needle so a null result is decisive. You can cost-rationalize and trim tactics later only if it works.

  • Growth experiments succeed only ~30% of the time — celebrate learning, not just wins
  • Failure = not learning, not the absence of revenue
  • Poorly-designed tests get repeated for years because they never failed conclusively
  • Experiments win via large N or a very significant treatment
  • B2B lacks large N, so maximize the treatment effect to get a decisive result

“for me failure is not that you didn't drive Revenue failure is not learning”

Sri Batchu · 52:00

“to counteract that I recommend people like just trying to maximize the treatment effect”

Sri Batchu · 53:30
#experimentation#growth#testing#culture
Hot Take1:03:30

Pay Your 10x Operators 10x the Comp

Sri takes the contrarian side of the compensation debate: instead of narrowing pay bands, companies should widen them. He believes small teams of successful people drive far more impact than larger teams of mediocre ones, so you must design a system that rewards 10x operators with 10x the comp — a differentiation already common at the executive level that he thinks should extend to more levels.

  • The comp conversation is too focused on narrowing gaps and bands
  • Small teams of top people beat larger teams of mediocre people
  • Design comp systems to reward 10x operators with 10x the pay
  • This differentiation already exists at the executive level
  • Extending it down levels is a competitive advantage for retaining talent

“you have to design a system where you're able to reward Timex operators with 10x the the con”

Sri Batchu · 1:04:00
#hiring#compensation#talent#management

Explainer· 3

Explainer12:00

At Ramp, Growth Engineers Own the Quota

Sri describes an unusual structure where the growth-engineering team shares the same goal as sales — pipeline driven and channel payback period — rather than owning product metrics. Because engineers feel accountable for quota and efficiency, they bottom-up generate the projects most likely to move topline. What they actually do: find the right prospects, craft messaging, prioritize responses, and even draft replies for reps.

  • Growth engineers share the sales goal: pipeline and payback period
  • They feel ownership of the quota, not just product metrics
  • Shared accountability drives bottom-up prioritization of high-impact projects
  • Concrete work: prospecting, messaging, response prioritization, drafting replies
  • It's one of Ramp's most efficient channels

“it's kind of unique to ramp where the engineers feel ownership of the quota”

Sri Batchu · 12:30
#growth#sales#team-structure#engineering
Explainer32:30

Rolling Every Team's Metric Into One North Star

Sri explains how Instacart unified around a single North Star — monthly active orders — while each sub-team optimized its own directly-controllable metric (e.g. app load time, searches per user). A finance/data translation layer converted each local metric into North Star impact, which made cross-team prioritization far easier by measuring where there's more 'mile per dollar' or per engineer. Translations were updated every six months.

  • Instacart's North Star was monthly active orders (i.e. monthly active users who order)
  • Each team owns a metric it can directly influence
  • A translation layer converts local metrics into North Star impact
  • This unifies cross-prioritization: where is there more mile per dollar/engineer
  • Translation factors are refreshed every six months

“instacart for example our North Star metric for for growth was monthly active orders and that's what we all rallied around and looked at every…”

Sri Batchu · 33:00
#metrics#north-star#prioritization#growth
Explainer38:00

Long-Term Holdouts to See a Team's True Cumulative Impact

Beyond regression-based translation factors, Instacart (like Facebook) ran long-term holdouts per surface area — e.g. the checkout team keeps a small holdout that never gets the new experience. This reveals the cumulative impact of all of a team's work on monthly active orders, effectively an A/B test with a permanent small control. Some Instacart users have never been shown ads because they've always been in the advertising holdout.

  • Each surface area keeps its own long-term holdout
  • Holdouts reveal cumulative impact, not just single-experiment lift
  • It's effectively an A/B test with a small permanent control group
  • Facebook used the same approach
  • Some users have never seen ads because they're permanently in the ads holdout

“we could also just see what the cumulative impact of all of the work is that Facebook did this too which is just long-term holdouts”

Sri Batchu · 38:00

“there is hold out for almost everything there's a holdout for ads for example”

Sri Batchu · 38:30
#experimentation#metrics#holdouts#ab-testing

Story· 3

Story07:00

Cap Table as a Growth Strategy

Sri explains a tactic Ramp's founders used that he'd never seen before: deliberately putting a large number of early-stage founders, operators, and advisors on the cap table, many of whom became Ramp's first customers. This seeded early word of mouth in the tech founder community. He adds that direct founder/operator investors mattered more than VC introductions, because peer customers who actually used the product carry more weight than an investor's recommendation.

  • Founders Eric and Kareem stacked the cap table with early-stage founders and influential operators
  • Many of those cap-table companies became Ramp's initial customers
  • Founder/executive customers drove more usage than VC introductions
  • A peer who has used the product beats an investor recommendation for conversion
  • Ramp's revenue today is mostly mid-market/Enterprise, not startups

“one of the things that they did is what I call kind of cap table as a growth strategy”

Sri Batchu · 07:00

“the investor opinion and recommendation matters but it turns out it doesn't matter maybe as much as uh as another customer who's actually used the…”

Sri Batchu · 08:30
#growth#fundraising#b2b#word-of-mouth
Story16:30

days.ramp.com and the Obsession with Cycle Time

Ramp runs an internal site, days.ramp.com, counting the days since the company's founding — 1,529 (with running decimals) when Lenny checked. CEO Eric cites the number at every board meeting and all-hands to reinforce that the team works in days, not weeks or years. The culture is a razor-sharp focus on reducing cycle time and a bias to action.

  • days.ramp.com displays days since founding, down to running decimals
  • The CEO references it at every board meeting and all-hands
  • The point: reduce cycle time and think in smaller units of decision-making
  • The growth team works in two-week sprints like a product team
  • Early board decks were labeled by day number (e.g. 'day 544')

“we we have like days.rank.com where we can see how many days it's been since the founding of ramp internally”

Sri Batchu · 16:30

“never put out something tomorrow that you can get done today”

Sri Batchu · 17:00
#velocity#culture#operations
Story27:30

Fundraising as a Growth Engine

Sri shares a non-obvious tactic: Ramp treats fundraising announcements as market moments that drive a non-trivial amount of top-of-funnel, not just capital. PR works, he says, but only when you genuinely have something interesting; Ramp bundles unique business color into funding announcements to make them valuable to a general audience.

  • Fundraising announcements are used as market moments that generate top-of-funnel
  • PR is a real growth lever, but only with a genuinely interesting story
  • Funding announcements are paired with extra 'color' to add reader value
  • Newsletters/podcasts help most with recruiting and reaching tech founders
  • Traditional PR still matters because it targets a different audience

“any time that we've been fundraising it we've been using that as an effective way of creating a market moment uh it's driven actually a…”

Sri Batchu · 28:00
#pr#fundraising#growth#marketing

Q&A· 1

Q&A1:11:30

His Favorite Interview Question: What Are You Bad At but Still Do?

Sri's favorite candidate question is 'what's something that you're really bad at but you still do and why.' Many people can't answer, which he treats as a yellow flag — a sign they only pursue things they succeed at and will bail at the first sign of trouble. He wants people who do things for intrinsic motivation and interest even when they aren't good at them.

  • Question: what are you bad at but still do, and why
  • Inability to answer is a yellow flag
  • It signals someone who only does what they're already good at
  • Such people tend to quit at the first sign of trouble
  • He wants intrinsic motivation beyond success

“favorite um interview question that I like to ask candidates actually is uh what's something that you're really bad at but you still do and…”

Sri Batchu · 1:11:30

“a lot of people actually struggle with that question and can't answer anything that they do that they're bad at”

Sri Batchu · 1:12:00
#hiring#interviewing#culture

Takeaway· 3

Takeaway50:00

How to Sequence Growth Channels in B2B

Sri lays out a general path most B2B companies should follow: founder-led sales first, then early salespeople plus low-cost targeted marketing (content, community, small events), then PR, then paid and brand, with SEO starting around the same time as paid. The logic: channels get more expensive and more effective the further along you go, as you understand your customers better. SEO is placed later because it takes time to build domain authority.

  • Start with founder-led sales — the early team must know how to sell
  • Then early salespeople plus low-cost targeted marketing (content, community, events)
  • Then PR, followed by paid and brand efforts
  • SEO starts around the same time as paid but pays off later
  • Channels get more expensive and more effective as you understand customers

“the channels get more expensive uh as you go farther along and they get more effective as you understand more about your customers”

Sri Batchu · 50:30
#growth#b2b#channels#marketing
Takeaway1:01:30

Sourcing Talent by Reverse-Engineering the Best Companies

Beyond the classic network search (ask who's the best, get intros, map the top teams), Sri shares a data-driven sourcing method most people overlook. Build a target list of companies one or two stages ahead of you, then use a tool like SimilarWeb to see which of them are genuinely best at the function you're hiring for — e.g. what share of traffic comes via email for a CRM/email hire — and source from those teams.

  • Classic method: network search, get intros, map the best teams and people
  • Target companies one to two growth stages ahead of you
  • Data method: use SimilarWeb to identify who's actually best at a function
  • Example: check which companies drive the most traffic via email for an email hire
  • Find the best company at the thing, then find the best person there

“you can go and see okay which ones are these are actually really effective at driving web traffic or app traffic or app downloads via…”

Sri Batchu · 1:02:30
#hiring#sourcing#tools#recruiting
Takeaway1:14:00

Everything Is Negotiable — Practical Money-Saving Tips

Closing the episode, Sri offers two money-saving tips. First, everything is negotiable: software contracts aren't as standardized as people assume, salespeople have quotas, so time your ask near quarter-end and offer to sign quickly for a discount. Second, hire slower — hire only when teams are truly stretched, which helps on both cost and impact.

  • Contracts aren't standardized — almost everything is negotiable
  • Salespeople have quotas and goals you can leverage
  • Time asks near quarter-end; offer to sign fast for a ~10% discount
  • Hire slower — only when people and teams are really stretched
  • Small teams accomplishing more helps on both cost and impact

“when it comes to contracts people think contracts are standardized for software and usually not”

Sri Batchu · 1:14:30

“be mindful of like quarter ends for sales people”

Sri Batchu · 1:14:30
#negotiation#cost#hiring#operations