LLenny's Podcast
← All episodes
Mayur Kamat (CPO at N26, ex-Binance Head of Product)22 May 2025

Unconventional product lessons from Binance, N26, Google, more

8Frameworks
15Insights

Frameworks in this episode

Innovation4 steps

Cohort-and-Control Data Reading

Never trust pre/post dashboards — read behavior by cohort and judge changes by variant-vs-control so the macro can't fool you.

Read →
Strategy4 steps

Hypothesis-to-Data Over Strategy

For most PMs, strategy is overrated — your real job is to shrink the time from hypothesis to data.

Read →
Strategy3 steps

Naturally Hedged Product Portfolio

Build products that make money in opposite macro conditions so troughs in one are peaks in another.

Read →
Leadership4 steps

Running Products via the Daily Meeting

Assign one owner to an urgent problem, give them unlimited resources except time, and force a daily report until it's solved.

Read →
Mindset3 steps

Slow vs Fast Decisions

For any reversible decision there's no right or wrong — only slow and fast — so decide fast and correct fast.

Read →
Productivity4 steps

The 10x Leverage Desk

Spend your finite time only on problems with 10x positive or negative impact — and physically move yourself into the details until they're solved.

Read →
Self-Mastery4 steps

The Four Filters for Career Acceleration

Growth rate, strength-fit, deferred comp, and the C-suite question — pick jobs on these, not on salary.

Read →
Self-Mastery3 steps

The Superpower Overlap Test

Find your true strengths via a two-layer self-plus-peer assessment, then only take roles where your superpower meets the company's.

Read →

Insights & moments

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

Hot Take· 2

Hot Take22:00

Don't Optimize for Comp Early — 90% Comes in Your Last 5 Years

Mayur's contrarian advice: for the first 15 years of your career, don't choose between jobs based on compensation. If you're on track to become an executive or found a company, your comp is so backloaded that roughly 90% of it lands in the last five years — so optimizing for a 10-20% early raise is a mistake versus optimizing for learning, strengths fit, and fun.

  • Comp is heavily backloaded for executive/founder tracks
  • Roughly 90% of lifetime comp comes in the last 5 years
  • Chasing a 10-20% early raise is optimizing the wrong variable
  • Rank jobs on learning compounding, strengths overlap, and fun instead

“you will find that the compensation is so much backloaded uh that you would make 90% of your compensation in the last 5 years uh…”

Mayur Kamat · 22:30
#career-advice#compensation#product-management
Hot Take42:00

Strategy Is Overrated — Race From Hypothesis to Data

Mayur's hot take: 'product strategy' is nearly an oxymoron. If you can test a hypothesis cheaply, you don't need a strategy — strategy is just 'keep doing more of what works, stop what doesn't.' For most PMs the real discipline is how fast you can get from a hypothesis to data. He recalls Google CPO Jonathan Rosenberg's line that if you bring ideas instead of data, the boss's ideas win.

  • If a hypothesis is testable, you don't really need a strategy
  • Strategy reduces to 'do more of it' or 'don't' once you can experiment
  • Most PM 'strategy' is just packaged intuition, and the loudest title usually wins
  • The real edge is speed from hypothesis to data

“for most product managers your strategy should be how fast can I go from hypothesis to data”

Mayur Kamat · 44:00

“come to me with data. If you come to me with ideas, we'll go with mine.”

Mayur Kamat · 44:00
#strategy#experimentation#product-management

Explainer· 5

Explainer07:00

How Binance Ran on 55 Direct Reports and a Daily Leadership Call

Binance operated with an extremely flat structure — founder CZ had around 55 direct reports and for much of the company's history there was just one level between employees and the CEO. Because one-on-ones were impossible at that span, they built a culture of one-to-many communication and a daily leadership call, so no decision was ever blocked for more than 24 hours.

  • CZ had ~55 direct reports; often just one level between staff and CEO
  • One-on-ones are impossible at that span, so the culture became 'one to many'
  • The global leadership team met every single day, including weekends and holidays, at 11pm Singapore time
  • No decision stayed blocked more than 24 hours; urgent items got an owner reporting on the daily call

“for a large portion of the history of the company, there was just one level between the individual employees and the CEO”

Mayur Kamat · 07:00

“You cannot have one-on- ones when you have 55 direct reports.”

Mayur Kamat · 07:00
#binance#management#org-design#decision-making
Explainer34:00

Why Fintech Companies Produce the Strongest PMs

Fintech dominates the list of companies producing the best PMs because it uniquely has two customers — users and regulators — where making one happy usually makes the other less happy. Unlike most companies where trade-offs are routine, in fintech every trade-off is existential: a wrong decision can end the business, which conditions PMs to a whole different level of stakeholder juggling.

  • Fintech PMs serve two customers at once: users and regulators
  • What pleases one side usually displeases the other
  • Trade-offs in fintech are existential, not routine
  • This constant high-stakes juggling builds unusually strong PMs

“fintech you have two customers. You have your usual customers and you have your regulated and you need to keep both of them happy”

Mayur Kamat · 34:00

“in fintech like every trade-off is existential”

Mayur Kamat · 34:30
#fintech#product-management#career-advice
Explainer48:00

Experimentation Turns Product Management Into a Science

Mayur's favorite thing he tells PMs: everyone thinks they can do a PM's job, so the discipline needs a science of its own. Building an experimentation culture provides it — you can tell someone their idea is bad because it failed six times across specific user groups at a measurable level. It takes a long time to move a team there, but once you do, every experiment becomes a dopamine hit and PMs feel like specialists rather than general-purpose technicians.

  • Everyone thinks they can do a PM's job, unlike accounting or engineering
  • Experimentation gives product management a defensible 'science'
  • You can reject ideas with evidence: 'we ran this six times and it failed'
  • Once adopted, experimentation is empowering and self-reinforcing for PMs

“the moment you build experimentation you now made it scientific”

Mayur Kamat · 48:30
#experimentation#product-management#culture
Explainer53:30

Why Banking Is a Naturally Self-Hedged, Unbounded Market

Mayur explains banking's structural advantages: there are more bank accounts than humans, so the addressable market never runs out. It's also self-hedged — when rates rise, spending falls but deposit income rises; when rates fall, spending and interchange income rise. N26 leaned into this by expanding from a bank-with-a-card into savings when rates were high and into lending as rates fell.

  • There are more bank accounts than people — the market is effectively unbounded
  • Banking is self-hedged across the rate cycle (deposits vs. interchange/spending)
  • High rates favor savings products; falling rates favor lending
  • N26 built a portfolio (card, savings, lending) that complements the macro

“there are more bank accounts than human beings right because a lot of people have more than one bank account so you never run out…”

Mayur Kamat · 53:30
#banking#fintech#n26#market-strategy
Explainer1:14:30

The Three Places AI Is a Real Game-Changer (and the 'Reverse Zip')

Despite being personally unimpressed by AI for his own CPO work, Mayur names three areas where AI is a company-level game-changer: developer productivity (an 18-25% boost from repo-integrated copilots on large codebases), customer support (deflecting 60-70% of the easiest queries), and fraud detection (spotting fraudster language and transaction patterns). He's jaded about consumer AI that inflates a few words into an essay only for another tool to re-compress it — a 'reverse zip.'

  • Developer productivity: ~18-25% boost from repo-integrated copilots on large codebases
  • Customer support: AI deflects roughly 60-70% of the easiest queries
  • Fraud: AI reads fraudster language and transaction patterns at scale
  • Skeptical of consumer AI that expands then re-compresses text — a wasteful 'reverse zip'

“there are three areas where AI is complete game changer”

Mayur Kamat · 1:15:30

“you have tools now that you write few things and they make like a long essay for it and then you have tools that compress…”

Mayur Kamat · 1:18:00
#ai#product-management#fraud#developer-productivity

Story· 5

Story05:30

Binance Went From Zero to a $400B Valuation in Five Years

Mayur describes the near-impossible scale of Binance: launched in 2017 as a crypto exchange when Coinbase and others already existed, it became number one within six months and reached a peak valuation of roughly $400 billion with only 2,000 employees inside five years. As head of product, his topline KPI was trading volume, measured monthly in the trillions of dollars.

  • Started in 2017, became the #1 crypto exchange within 6 months
  • Peak valuation around $400 billion in 5 years with only ~2,000 employees
  • Teams of 10-20 people ran multi-billion-dollar profit businesses
  • His monthly KPI (trading volume) was measured in trillions of dollars

“that's like launching a search engine today and in 6 months later beating Google”

Mayur Kamat · 05:30

“So that's 0 to 400 billion in 5 years.”

Mayur Kamat · 06:00
#binance#crypto#scale#growth
Story11:00

The 500-Cell KYC Spreadsheet and 'No User Left Behind'

When crypto suddenly became regulated, Binance's signup conversion collapsed from ~100% to ~2% because users now needed a full bank-style KYC flow across 200 countries. Rather than optimize only for the top 90% of users, CZ insisted every one of ~500 country/document-type cells mattered — even one user in Congo — treating financial inclusion as the point. They once scaled the KYC team from 20 people to a target of 500 for three months.

  • New regulation dropped conversion from ~100% to ~2%
  • They tracked a ~500-cell spreadsheet: top 50 countries × top 10 document types
  • CZ refused to ignore low-impact cells — every user counts as financial inclusion
  • The KYC team scaled from 20 people toward 500 for a three-month push

“It's not that customer is not a number. It's a person at the end of the the screen and we care about them.”

Mayur Kamat · 12:30
#binance#kyc#onboarding#customer-focus
Story14:30

Google Spun Up 1,000x Servers Overnight to Absorb an iOS Bug

On his first day at Google, Mayur was pulled from onboarding into a war room: an iOS 4 bug caused every user to re-sync their entire Gmail inbox, hitting servers with a 1,000x load while the Wall Street Journal wrote about the outage. A single call to the head of infrastructure produced a thousand times the servers for two weeks — a defining 'that is scale' moment.

  • An iOS 4 bug forced every user to re-sync their whole Gmail inbox
  • The bug would take Apple 2-3 weeks to fix, causing ~1,000x server load
  • One call to the infra head provisioned 1,000x servers for two weeks
  • The lesson was less about having resources than being able to maneuver them instantly

“I slipped a button and there were thousandx servers that showed up for the next two weeks and I was like wow that is scale…”

Mayur Kamat · 15:30
#google#gmail#infrastructure#scale
Story1:07:00

The CEO Who Moved His Desk to the Highest-Leverage Problem

Mayur's rule is to work only on problems with a 10x (or 100x) positive or negative impact — usually a growth or compliance problem in fintech. He illustrates it with Whitepages founder Alex Algard, who had no fixed office but a movable desk he'd wheel into whichever department had the highest-leverage opportunity, sitting there and personally running it until it was solved. Doing this requires the humility to work in the details and a deliberately empty calendar.

  • Focus only on problems with ~10x (or 100x) positive or negative impact
  • In fintech that's almost always a growth or compliance problem
  • Alex Algard physically moved his desk into the highest-leverage department and ran it himself
  • High-leverage work requires humility to dig into details and an open calendar

“he would move his desk to one of the departments which you think had the highest leverage opportunity”

Mayur Kamat · 1:08:30

“a full calendar is a badge of shame, not a badge of honor.”

Mayur Kamat · 1:13:30
#leverage#founder-mode#prioritization#product-management
Story1:19:30

Why Hangouts Failed: Some Companies Can't Win Certain Products

Mayur calls being the first PM on Google Hangouts his most spectacular failure — thousands of people, the full power of Google, and Larry Page offering to change Chrome itself, yet they never built a great messaging product. His 15-year-later diagnosis: certain companies can never succeed at certain product types because the founder's DNA works against them — Microsoft with mobile, Google with social, Facebook with enterprise. The lesson: don't take on projects that will take six months to a year, because you can't control the macro.

  • Hangouts had thousands of people and Larry Page's full backing yet still failed
  • Premise: a founder's DNA can actively block a company from certain product types
  • Examples: Microsoft/mobile, Google/social, Facebook/enterprise
  • Takeaway: avoid 6-12 month projects where the macro can shift out of your control
  • A silver lining: the Hangouts team invented WebRTC, now used by every video-calling product

“Uh we had Larry literally sitting with us saying we can do anything we want Chrome to do and we still didn't manage to build…”

Mayur Kamat · 1:20:00

“I have this premise that certain companies can never succeed at certain type of products right like Microsoft with mobile or Google with social or…”

Mayur Kamat · 1:20:30
#google#failure#company-dna#product-management

Takeaway· 3

Takeaway19:00

Join Fast-Growing Companies So Your Learning Compounds

Mayur's number-one career advice: pick companies that are growing fast, because your learning compounds like interest — compounding daily versus yearly leaves you in a completely different place after two years. He contrasts Microsoft, where products he interned on hadn't shipped by the time he left years later, with Binance, where you shipped and learned every hour.

  • Growth speed determines how fast your learning compounds, like compound interest
  • At Microsoft, some products he worked on hadn't shipped 3-4 years later
  • At Binance you shipped and learned every day, hour, or minute
  • Faster compounding of learning means faster career growth

“the best thing you can do is find companies that are growing fast because it kind of compounds your learning at a much faster interval”

Mayur Kamat · 19:00
#career-advice#product-management#growth
Takeaway20:30

Optimize for Your Superpowers, Not Your Weaknesses

Mayur argues strengths are defined very early in life, so instead of chasing the standard career feedback of 'improve your weaknesses,' you should identify your superpowers and pick jobs where success is determined by how much you get to use them. Early in your career you can't build a team to cover your gaps, so pick a role that matches — creative, high-experimentation people belong on growth teams; structured, high-EQ people belong where complex stakeholder management wins.

  • Strengths get defined very early and are hard to fundamentally change
  • Most career feedback wrongly focuses on fixing weaknesses
  • Pick jobs where success is determined by how much of your superpowers you get to use
  • Match your profile to the team: creative/experimental → growth; structured/high-EQ → complex people & process roles

“Firmly in the camp that you need to know what you're great at, what are your superpowers, and you need to find jobs where success…”

Mayur Kamat · 20:30
#career-advice#strengths#product-management
Takeaway1:33:30

There Are No Right or Wrong Decisions, Only Slow and Fast Ones

Mayur's life motto, tied to his hypothesis-to-data philosophy: for any reversible decision, there's no right or wrong — only slow and fast. If you make a wrong decision fast, you'll discover it's wrong and correct it sooner than you'd have made the 'right' one after months of deliberation. The caveat: this doesn't apply where a mistake could kill your user or your company (healthcare, military, compliance).

  • For reversible decisions, speed beats deliberation
  • A fast wrong decision is caught and corrected faster than a slow right one
  • Exception: irreversible or catastrophic domains (healthcare, military, compliance)
  • Reframes decision-making as slow vs. fast rather than right vs. wrong

“there's no right or wrong decision. They're just slow and fast decisions.”

Mayur Kamat · 1:33:30

“for anything that's reversible, anything that's not going to get you in jail or kill your company, no right or wrong decisions, just slow or…”

Mayur Kamat · 1:34:00
#decision-making#life-motto#product-management