Nov 15, 2024 · 1h 0m · news
Raman Malik: Inside Perplexity’s Growth Machine: What Worked, What Did Not Work | E1226 · 20VC with Harry Stebbings
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of 20VC, Harry Stebbings interviews Raman Malik, Head of Growth at Perplexity, to explore the tactical frameworks of scaling an AI-first product. Malik breaks down why organic user retention and activation must precede paid acquisition, while sharing his contrarian philosophies on hiring, A/B testing, and growth team organization.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 25.7% of the talking time here. How this is scored →
Harry as informed peer 4.0 Guest teaching 3.8 Guest disagreement 1.4 Harry pushing back 2.2
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 32:27 Paid acquisition is a total drug that cannibalizes organic
Raman forcefully criticizes industry reliance on paid acquisition, calling it a drug and presenting counter-intuitive findings from Lyft showing zero incrementality.
Hardest push from Harry ▶ 6:32 Harry challenges growth team cost and ARR prerequisitesHarry directly pushes back on Raman's timeline for building an independent growth function, arguing that multi-million dollar growth teams require $25M ARR rather than early PMF.
Biggest teaching moment ▶ 32:45 The zero-incrementality paid acquisition experiment at LyftRaman educates Harry on attribution illusions by detailing how turning off all paid passenger marketing at Lyft resulted in almost no drop in acquisition due to organic cannibalization.
Harry holds his own ▶ 15:51 Harry challenges session duration vs time-to-value metricsHarry demonstrates strong product domain expertise by challenging the metric of longer session duration, arguing that efficient search engines should optimize for minimal time-to-value.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| From Alabama to Silicon Valley | 2 | 2 | 1 | 1 | Harry sets an engaging tone, asking about Raman's journey from Alabama through an MBA program to becoming Head of Growth at Perplexity. Raman explains how MBA programs serve as pre-seed tinkering environments and how transitioning from founder to employee involves de-scoping focus. | |
| Defining Growth: Bridging Product and Marketing | 5 | 4 | 2 | 4 | Harry pushes back on Raman's definition of an independent growth team, arguing that dedicated growth product and marketing teams require tens of millions in ARR rather than early post-PMF stages. Raman defends early investment by focusing on retention metrics as kindling for growth. | |
| Flat Organizations and Mapping the Funnel | 3 | 4 | 1 | 2 | Raman discusses working in flat orgs and mapping user funnels before debating whether growth relies on micro-optimizations or big swings. He illustrates how small retention gains compound across active user bases while emphasizing quarterly high-risk swings. | |
| A/B Testing: The Good, the Bad, and the Ugly | 4 | 3 | 2 | 4 | Harry challenges whether great product managers actually need A/B testing or if testing becomes a crutch for performance reviews. Raman outlines the good, bad, and ugly of A/B testing, stressing that intuition must guide test hypotheses. | |
| Mixed Metrics and the Perils of Sign-In Gates | 4 | 4 | 1 | 1 | Harry references Facebook's classic milestone metric of getting 5 friends to drive retention and asks for Perplexity's equivalent. Raman shares that getting a user to three queries in their first session serves as Perplexity's key milestone metric. | |
| Gaming Metrics and Downstream Correlations | 6 | 4 | 2 | 4 | Harry cites Alex Schultz on metric gaming and challenges Raman's desire for longer session times, arguing search tools should minimize time-to-value. Raman explains that session duration in Perplexity reflects deep exploratory learning rather than friction. | |
| Mistakes in Acquisition: The Friction of Influencer Marketing | 3 | 4 | 1 | 2 | Raman details a failed influencer marketing push where managing long-tail creators created severe operational drag. He explains how analyzing funnel data revealed that early retention rather than top-of-funnel acquisition was Perplexity's core bottleneck. | |
| Defining Good Retention in Consumer Tech | 5 | 4 | 1 | 2 | Harry reacts with surprise to Raman's 45% month-6 consumer retention target and benchmarks it against Duolingo's world-class stats. Raman breaks down cohort mix shifts, explaining how targeting work and academic use cases drives stickier retention. | |
| The Homepage Battle: Minimalism vs. Value Messaging | 5 | 3 | 2 | 3 | Harry performs a live copy teardown on Perplexity's meta description, calling it overly wordy. Raman candidly rates Perplexity's current messaging a 6 out of 10 and discusses balancing SEO optimization with clean brand messaging. | |
| Intentionally Deferring CAC to LTV Metrics | 4 | 4 | 1 | 2 | Harry highlights CEO Aravind Srinivas's comment that Raman deferred paid acquisition in favor of fixing activation and retention. Raman explains why early-stage startups should intentionally ignore CAC to LTV metrics until core product retention reaches diminishing returns. | |
| Paid Acquisition as a Drug: The Lyft Experiment | 5 | 5 | 2 | 3 | Harry asks why Perplexity touches paid ads given its massive organic momentum and endorsement from NVIDIA's Jensen Huang. Raman labels paid acquisition a drug, detailing a Lyft experiment where turning off passenger ads revealed zero incrementality as paid dollars merely cannibalized organic growth. | |
| Unlocking Organic Virality in Brand Marketing | 5 | 4 | 1 | 1 | Raman discusses brand marketing and virality, citing Perplexity's Jim Harbaugh TV ad campaign. Harry adds a case study of a portfolio company placing a provocative guerrilla billboard outside Goldman Sachs headquarters to drive massive organic PR. | |
| Unpacking Growth Alpha: Be First or Be Better | 3 | 4 | 1 | 2 | Raman breaks down his growth philosophy: either be first to an untapped channel or execute significantly better than competitors. He argues that community programs on college campuses represent an underappreciated channel for building local density. | |
| Unlocking Tastemakers and Content Sharing Loops | 4 | 3 | 1 | 2 | Harry asserts that marketing content must strictly educate or entertain, dismissing corporate announcement posts as useless. Raman agrees and reflects on his regret over ineffective newsletter sponsorships while discussing product-level sharing loops. | |
| Hiring a Growth Team: The Case for Founders | 4 | 4 | 1 | 1 | Harry asks about first growth hires and references Rippling's heavy hiring of former founders. Raman cites Mike Maples' jazz band analogy to argue that former founders excel in early growth roles because they comfortably take big swings and tolerate failure. | |
| Hiring for Taste and Take-Home Assignments | 4 | 4 | 2 | 3 | Raman defines taste in hiring as the ability to recognize excellence and reverse engineer it. Harry challenges whether company-specific take-home assignments create an unfair information asymmetry for candidates, which Raman defends as a collaborative gauge. | |
| Onboarding Rules and Slow Hiring Philosophies | 4 | 4 | 2 | 3 | Raman asserts that if a senior growth hire requires a 30-day onboarding checklist, it is a red flag. Harry offers a counter-perspective to Raman's self-professed slow hiring, noting that most podcast guests regret hiring too fast rather than waiting. | |
| Quick Fire: Expensive Mistakes and AI Application Value | 4 | 4 | 1 | 1 | In a quick-fire round, Raman highlights dashboard data mastery as essential for new growth heads and reflects on his changing perspective regarding AI application layer value. Harry echoes the broader tech ecosystem shift from viewing AI apps as wrappers to durable businesses. | |
| Quick Fire: CEO Strengths and Perplexity Wrapped | 3 | 4 | 1 | 1 | Raman praises CEO Aravind Srinivas's ability to switch seamlessly between macro company strategy and micro user funnel details. Harry commends Raman's balanced answer about his CEO, and Raman closes by hinting at an upcoming Perplexity Wrapped feature. |
Statements from this episode (52)
Malik: Third-party endorsement and product bundling beat direct marketing
“We can go out and scream from the mountaintops about perplexity, but it is a lot better when someone else is telling you to use it, or it is being bundled into an existing product of yours.”
Malik: Users need three to seven product exposures before trying it
“You need someone to hear about your product three to seven times before they're going to give it an honest trial.”
Perplexity's key activation metric is three queries in a user's first session
“If I can get a user from perplexity specifically to three queries in that first session. Well, now I know I'm really onto something.”
Malik: Most people in the tech industry absolutely hate MBAs
“MBAs have such an interesting reputation in tech. Most people just absolutely hate them.”
Malik: MBAs make sense primarily for entering structured corporate paths
“I think MBAs make a lot of sense for people who are trying to break into a specific industry. Like if you are trying to break into private equity or you want to be a PM at Amazon or go into McKinsey, then MBAs have like your preset path and you just follow you…”
Malik: Moving from founder to growth lead requires deliberate functional de-scoping
“Founders enjoy the variety of, I got to go work on a hundred things and I got to steer the ship, but now you're hyper-focused on, okay, I'm here to drive growth. I am de-scoping myself just to go do an excellent job at this, and that's an adjustment as well.”
Malik: Growth is user retention and power-user development, not acquisition
“Growth is not just about, let me fill the top of the funnel with new users. It's about getting those users to become retained, to develop into power users so that you can eventually monetize them.”
Stebbings: Startups need $25M ARR before hiring dedicated growth teams
“If you were at like two, three million in ARR, and got product market fit, doth bust a no fucking parsnips. You need twenty five million in ARR for that team.”
Malik: Invest in growth teams after reaching 30% month 3-4 retention
“As soon as you see that you have early signs in retention, like we're seeing like, you know, if you see 30% retention in month three or four, you have the kindling in your fire where it is now worth it to go put gasoline on it, go make that investment.”
Malik: Growth teams must take high-risk swings at least quarterly
“My rule of thumb is once a quarter, we have to be taking at least a couple massive swings, right? A big new feature, a big marketing campaign. It has to be high risk. It has to be high reward.”
Malik: A 25% success rate on major growth swings is exceptional
“I'd say 25% would be pretty great on these big swings. Maybe like one, once a year, we have just a banger campaign or banger new feature that is now really driving growth or really opened up a new audience for us.”
Malik: Big tech runs A/B tests primarily to justify performance reviews
“And then the ugly is I'm running an AB test. So that I can put something in my performance review about a number I moved, or I can like say something to my manager. And you see that as companies get bigger, right? We're running A-B tests just to like be able t…”
Malik: Teams running A/B tests without intuition are flying blind
“You should go into an AB test with a pretty strong confidence of what's going to happen. And this is where like intuition comes in of like, We're running this, and here's what I think is going to happen. I think we're really going to move this metric, but we m…”
Malik: Growth teams must check if short-term metrics still drive retention
“One is monitoring. Does this still correlate down the road once we're done with gaming it, once we're done with tinkering with it? And if we're seeing it, that correlations start to break, well, okay, I'm not being honest with myself around if we're actually h…”
Malik: Perplexity plans student-focused partnerships for distribution at scale
“So I think a lot about the student audience. And so there's some partnerships in that student world that I really want to break into to help drive distribution within students at scale.”
Malik: Perplexity's long-tail influencer marketing strategy failed and wasted time
“Totally underestimated the amount of effort it is to work individually with creators on specific content and making sure they get the right messaging and perplexity is not a simple concept to explain. Like me saying AI search engine means nothing. And so to ma…”
Malik: Perplexity shifts strategy to partner with fewer high-relationship creators
“And so I think we're moving very quickly as we'd rather work with fewer creators and just develop a really good relationship with them.”
Malik targets 30% conversion rate for logged-out Perplexity visitors
“From that logged out experience, we, I'd love to push that into, like, the 30% of users are hitting my milestone metrics, and we're converting them to signed in users.”
Malik: Non-social consumer apps should aim for 45% month-six retention
“I think it was like by month six, you want 45% is really, really, really good. That's the target, right? That's where I want to push to, especially when you're not a social product that has more of like that smile retention curve.”
Malik: Duolingo achieves a world-class 50% 12-month user retention rate
“Yeah, that's 12 month retention at 50%. World-class. And the engine they have built is incredible.”
Malik: Organic channels always outperform paid acquisition in retention
“Organic and referrals is always going to have higher retention than paid acquisition and partnerships.”
Malik: Knowledge-gap search intent drives sticky user retention over curiosity
“You see a lot of trial users that like are coming in from like curiosity traffic, like the weekenders that are putting in, you know, how many times does someone fart in a day? That's a little bit lower retention. If I'm here to solve a knowledge gap, like I'm …”
Malik: Cross-device app conversion is a key retention driver
“Cross device conversion is really, really strong for retention because now you're using it on weekends. You're using it during the day.”
Malik: Perplexity is moving toward simpler landing page messaging
“Where we're moving to is really simple language around ask anything or what do you want to know and search like never before.”
Malik: Perplexity is completely unoptimized for driving user subscriptions
“Like you look at perplexity, it is completely unoptimized for driving subscriptions.”
Raman Malik: Only optimize CAC-to-LTV after retention hits diminishing returns
“Do they have, are they at the scaling period where it is now turn on these channels and drive growth because we feel Really good about where we are in terms of retention or any incremental work we're going to do on trying to move activation or move early reten…”
Malik: Perplexity barely uses paid acquisition channels
“We, I mean, we don't really. We barely touch paid.”
Malik: Paid acquisition is a drug that underperforms on user retention
“Paid acquisition is a total drug. It makes everyone feel good because top of funnel numbers move up, and once you see those, you know, weekly activation rates, well, no one wants to see that go down. But it's going to underperform on retention, and maybe most …”
Malik: Turning off Lyft's paid channels reduced signups by only 5-10%
“And so the growth marketing team turns off every single paid channel and installs and signups barely move. They go down by like maybe five, 10%. I don't remember what it was and what did go down. It was all Low quality. So every marketing dollar that was being…”
Malik: Perplexity's only paid acquisition is lightweight TikTok ads for young users
“So right now, like the only paid acquisition we've ever touched and it's so lightweight are some just like simple TikTok ads testing different value propositions with 18 to 24 year olds.”
Malik: AI companies must focus on concrete utility rather than AI buzzwords
“You need to abstract away this whole AI piece and just go straight to the value prop of get answers immediately backed by sources.”
Malik: Perplexity caps ad testing spend at $10,000 per week
“Not much at all. You don't need to spend much to really understand what that retention is going to look like. 10 K a week.”
Malik: Perplexity does not have a dedicated marketing department
“What's cool about Perplexi is we don't have a marketing department.”
Malik: Perplexity found newsletter sponsorships ineffective for paid growth
“A good example of another area I got burned was newsletter sponsorships. It's like, okay, like there, there's some interesting audiences here. I want to try this out, but I think it's really hard unless you have a really crisp product that you can put into a r…”
Malik: Early adoption of unique channels drove growth for Dropbox and Pinterest
“Dropbox, Dropbox, like with the referral program, Groupon and email, Pinterest and SEO, Duolingo, and just like a totally unhinged social media, right? All of them were kind of first to this concept and it really, really worked.”
Malik: Growth strategy requires being first to a channel or significantly better
“There's no middle ground. If you're not first, it literally just means your only other option is to just be a lot better than everyone else. And that's like where we need to hold ourselves to.”
Malik: Community marketing via power users is an underappreciated growth lever
“I think community and how you can unlock your power users to share the value of your product to their own miniature communities is incredibly powerful.”
Malik: Perplexity under-indexes in brand awareness among students vs ChatGPT
“There are students using perplexity and getting tons and tons of value, but we under index and awareness with students relative to chat GPT.”
Malik: Perplexity's ultimate target is competing directly against Google
“I think search is a massive market to go after, and so if there's somewhere we want to live, it's going up against the big guy.”
Harry Stebbings: Effective marketing content must either educate or entertain
“A piece of content has to do one of two things. It has to educate or it has to entertain. And if it doesn't do that, then it doesn't have any place in any of our things.”
Malik: Perplexity drives substantial traffic from user sharing without viral loops
“I don't think sharing is inherent to the perplexity user journey or user flow. Like, I don't think that we have an inherent viral loop where someone does some research and then shares a thread or screenshot to another person. But there's a lot of sharing happe…”
Malik: Former founders make better growth hires due to high risk tolerance
“We have a lot of founders here, and I think the best part about hiring founders is that They are much more comfortable at taking very big swings and being willing to fail. Like it is by nature. If you're going down the founder path, you have to be a little bit…”
Malik: Removing experiment guardrails drives company-wide bold thinking
“I mean, you don't want to put guardrails over their ability to experiment and try different things. Guardrails is going to hurt everyone. It's going to make everyone think a little bit smaller. You remove those guardrails. Everyone's thinking big.”
Malik: Perplexity requires take-home assignments for every growth role candidate
“I do for every, every role. It's a big ask, so I am cognizant of trying to keep it well scoped.”
Malik: Needing a detailed onboarding document signals a bad hire
“My rule of thumb is if they truly need an onboarding doc, we're in trouble.”
Malik: Employees quit bad teams, not bad managers
“People say that, you know, they quit the manager, not the company. Like, I think people quit the team.”
Malik: Apple privacy changes killed Meta ads as a core scaling channel
“I mean, I think especially with all the latest Apple changes, I don't know how anyone is going to make Instagram or Facebook ads efficiently work at scale as like a core growth channel.”
Malik: Lyft overestimated product iteration potential and hit diminishing returns early
“We overestimated how much better we could make the ride sharing product. I think we started to hit, like when you look at ride sharing four or five years ago versus today, how much has the product changed? Not that much. So if you don't believe, well, sorry, i…”
Malik: The majority of AI value will emerge at the application layer
“When I joined Perplexity and was writing out my pro-con list, I think the one that was, like, up there was, like, can the application layer capture and monetize value? And, or is it all just, you know, one big pass through cost? And now, you know, fast forward…”
Malik: Perplexity CEO Aravind Srinivas uniquely bridges strategy and granular details
“He can go from a very high level company strategy of what direction do we need to go very quickly, two seconds later, down into incredibly specific weeds of like what a user funnel should be. He can jump at these layers like no one I've ever seen. And this is …”