Raman Malik on Why Growth Teams Should Fail More and Spend Less

The 25% Success Rate Doctrine

Most growth teams are lying to themselves. They run A/B tests to pad performance reviews, chase micro-wins that look good in standups, and avoid the career risk of shipping something that might actually fail. Raman Malik, head of growth at Perplexity, has a different religion: once a quarter, his team must take massive swings with a 25% success rate.

Not 60%. Not 80%. A quarter of the time. Maybe once a year, they'll ship "a banger campaign or banger new feature that is now really driving growth or really opened up a new audience." The rest of the time, they fail in public. This is not growth theater. It's growth as practiced by someone who spent three years getting his "ass kicked" as a founder before joining one of the fastest-growing consumer products in history.

Once a quarter, we have to be taking at least a couple massive swings. A big new feature, a big marketing campaign. It has to be high risk. It has to be high reward. I have to be able to stand in front of the company and say, I'm going to do something and be willing to fail.

— Raman Malik

The confidence to fail comes from understanding the alternative. Malik describes the "ugly" version of A/B testing as "running an A/B test so that I can put something in my performance review about a number I moved." He's seen it metastasize as companies scale: PMs running tests not to learn, but to manufacture legible wins. The antidote is reserving space for experiments where the primary outcome might be embarrassment.

The Water Level Theory of Micro-Optimizations

The irony is that Malik also believes micro-optimizations are "seriously underrated." Moving retention from 30% to 40% doesn't sound like a moon shot. But when you model it out—new users plus retained users plus resurrections minus churned users—that 10-point lift "raised the entire water level of the company." Weekly actives rise. Monthly actives compound. The growth model that lives in his Excel sheet proves that small changes to retention curves create durable lift.

The trick is knowing when to stop squeezing. Micro-optimizations have diminishing returns, and "at some point, I'm not going to be able to squeeze out an incremental 10-15% without a lot of wasted experimentation time." That's the forcing function for the quarterly big swings. You can't micro-optimize your way to a new audience or a new use case.

That 10% small change in retention or in new user activation, you see the entire water level of your active user base just increase. That small micro-optimization just raised the entire water level of the company.

— Raman Malik

Malik's framework splits growth into two disciplines that sit between product and marketing. Growth product is engineering, design, and data science obsessing over the user funnel: acquisition, activation, retention, monetization. Growth marketing uses channels, lifecycle comms, community, and campaigns to move the same funnel. Both teams cost millions. Both require "early signs in retention"—he cites 30% retention at month three or four—before the investment makes sense. Below that threshold, you're burning budget on a fire that hasn't caught.

Three Queries and the Milestone Metric

Perplexity's magic number is three. If a user completes three queries in their first session, Malik knows "I'm really onto something." It's enough time in-product for someone to understand the value prop, and it correlates with retention at 30 and 60 days. He calls these "milestone metrics": behavioral markers in early sessions that predict long-term engagement.

But every milestone can be gamed. As Alex Schultz from Meta told him, you could flood the interface with recommended searches and technically hit three queries without delivering real value. Malik's hedge is monitoring correlation over time and adopting a "rounded approach." He doesn't just track query count; he also watches first-session duration. Did the user spend 10 minutes exploring? Did one query lead to follow-ups, pulling them into a "rabbit hole" of learning?

If I can get a user on Perplexity specifically to 3 queries in that first session, now I know I'm really onto something because that is what a lot of great people did. It's enough time in product where that user now understands the value of Perplexity and there's a high likelihood that they're going to come back and retain.

— Raman Malik

The worst version of milestone metrics is when teams optimize for the metric instead of the outcome. Malik has shipped experiences that juiced activations—like showing a sign-in gate after five queries—only to discover it tanked query volume because users "hated it" and bounced. Mixed results like these force trade-offs: do you value the activated user or the lost session? He won't run a 60-day test to find out. He makes a call based on whether activated users will retain enough to offset the initial volume loss.

80% Organic and the Curiosity Tax

Perplexity's acquisition makeup is 80% organic word of mouth. That's the "best channel you can ever have," and it's a function of what Malik calls "curiosity traffic." AI companies that ship something magical get a wave of organic trial. Two years in, Perplexity is still riding it. The growth team's job isn't to replace that wave—it's to nurture it and find wedges into new audiences.

Partnerships are the wedge. LinkedIn, Xfinity, and Lenny's podcast all give away a year of Perplexity Pro to their members. Malik's partnerships team are "absolute sharks," and the unit economics work because the cost is variable: "If you don't use Pro, it's pretty variable. If you're not using Pro, we're not losing money. But if you are using it, well, now we're capturing you as a user." The real game is converting trial users into paying customers after the free year expires.

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. That's going to drive a lot more trial.

— Raman Malik

The partnership Malik most wants is access to students. He's thinking in audiences, not channels, and students represent a cohort that could become power users if Perplexity embeds itself in their research and learning workflows early. The biggest mistake he's made in acquisition? Influencer marketing. He "totally underestimated the amount of effort it is to work individually with creators on specific content," especially when "Perplexity is not a simple concept to explain." Saying "AI search engine means nothing." The experiments fell flat and consumed time. The pivot: work with fewer creators and build deeper relationships instead of managing long-tail volume.

From Founder to Function

Malik came to Perplexity from three years of grinding as a founder—"month after month, pivot after pivot, nothing is moving." When he shut down his startup and joined Perplexity, the transition was brutal. His ego was bruised. He de-scoped himself from steering the ship to driving one function. Founders enjoy variety; growth leaders hyper-focus.

He'd left his MBA at Stanford a year early to tinker full-time on ideas, funded by the school's pre-pre-seed program. He never took an interview. When the startup failed, he knew "how freaking hard startups are," and joining Perplexity—already moving "crazy fast"—felt like a gift. The first weeks were spent turning over stones, building a map of acquisition, engagement, and monetization. Most questions yielded "100 more questions." So he built infrastructure: logging, attribution, first-party cookies.

The growth function he leads now bridges product and marketing. It's a team that costs millions and requires retention metrics that justify the spend. But the mandate is the same as when he was solo: obsess over the user journey from acquisition to monetization, make big bets that might fail, and raise the water level a few percentage points at a time.