Raman Malik on Why Perplexity Spends Almost Nothing on Paid Acquisition

Paid Acquisition Is a Total Drug

Raman Malik spent years on Lyft's paid acquisition team, burning millions on new passenger signups. Then his head of growth told him to turn off every single channel.

Installs barely moved. Signups dropped maybe 5–10%. Every dollar spent was cannibalizing organic traffic.

Paid acquisition is a total drug. It makes everyone feel good because top-of-funnel numbers move up. And once you see those weekly activation rates, well, no one wants to see that go down, but it's going to underperform on retention. And maybe most importantly, it's most likely not gonna be incremental.

— Raman Malik

At Perplexity, 80% of growth is organic word of mouth. The lesson: if you have product magic and real retention, paid channels don't add users—they just pull forward the ones who'd find you anyway.

It's Not an Acquisition Problem, It's a Retention Problem

When Malik joined Perplexity, the company had traffic flooding in from partnerships with LinkedIn, Xfinity, and Lenny's podcast. Dmitry and the business team were already crushing distribution. The diagnosis was fast: this wasn't an acquisition problem.

He went straight to week 1 and week 2 retention. Not bad—decent consumer retention numbers. But that's where the leverage was.

If we can move that number early on at the top of the funnel, we're increasing the water level of all activated users and weekly actives. That's where I need to hang out.

— Raman Malik

Malik's target: getting logged-out visitors to hit 3 queries in their first session. That milestone metric correlates with long-term retention. Users who ask 3 questions understand the value. They come back. The rest churn fast.

Good consumer retention by month 6? Around 45%. Duolingo hits 50% at 12 months. That's world-class. Perplexity isn't there yet, but the engine is built around mix shift—optimizing platform (cross-device), audience (students, professionals), and eliminating curiosity traffic that dies after one fart-related query.

Micro-Optimizations Raise the Entire Water Level

Growth leaders love talking about big swings. Malik thinks micro-optimizations are seriously underrated.

Open a growth model: new users, retained users, resurrections, churned users. If you increase retention by 10%, the entire active user base rises. That's not incremental—it's structural.

That small micro-optimization just raised the entire water level of the company.

— Raman Malik

But diminishing returns kick in. So Malik's rule: once a quarter, take a couple massive swings. High risk, high reward. Success rate on those? About 25%. Maybe one banger feature or campaign per year that opens a new audience or drives real growth.

The rest of the time, squeeze value from the funnel you've got.

CAC:LTV Doesn't Matter Until Retention Hits Diminishing Returns

When Malik joined Perplexity, Aravind expected him to ask for marketing budgets and TikTok ads. He didn't. He wanted activation and retention metrics first.

The logic: if you nail sticky retention, you'll figure out monetization. Consumer subscriptions, ads, whatever. Don't optimize CAC:LTV when you don't even know if users stick around.

Perplexity is "completely unoptimized" for subscriptions. They could add more gates, limit Pro Search to 3 queries per day for free users, surface upsells in the moment. They haven't. Because retention comes first.

Only now, with retention strong and diminishing returns on activation work, is Malik starting to think about paid channel efficiency and CAC:LTV ratios. The sequencing matters. Growth isn't about filling the top of the funnel—it's about making sure users who enter actually stay.

Partnerships Are Acquisition on Easy Mode

Perplexity's partnership team is "absolute sharks." LinkedIn, Xfinity, Lenny's podcast—all give away a year of Pro to their members. Does it destroy unit economics?

Not really. If users don't activate, the cost is variable. If they do, Perplexity captures them as retained users who now understand the magic. The game becomes conversion post-trial.

It's better when someone else tells users to try your product than when you scream from the mountaintops. Bundling into existing products drives trial at scale. The biggest partnership Malik wants but doesn't have yet? Student audiences. He's hunting for distribution channels that unlock that cohort—students using Perplexity to find sources for papers have the stickiest retention of any segment.

Curiosity traffic—people asking how many times humans fart per day—churns fast. Knowledge gap solvers stay forever.