LLenny's Podcast
← All frameworks
StrategyMayur Kamat (CPO at N26, ex-Binance Head of Product)

Hypothesis-to-Data Over Strategy

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

Difficulty
Moderate
Time to result
~weeks to results
Steps
4
Confidence
90%

Kamat argues 'product strategy' is largely packaged intuition that defaults to the loudest voice in the room. Instead, a product manager's strategy should be measured by how fast they can go from a testable hypothesis to real data. If you can define a hypothesis (users, conversion, CAC, LTV) and test it cheaply in-market, you don't need a strategy — the results tell you to do more or stop. Strategy-heavy deliberation only earns its keep where decisions are irreversible or genuinely slow to validate (pricing in Europe, licenses, compliance).

Origin

Mayur Kamat's contrarian take, reinforced by Google CPO Jonathan Rosenberg's rule 'come to me with data; if you come to me with ideas, we'll go with mine.'

Core principles

  • 01Strategy is either 'do more of it' or 'don't do it' — the hard part is sorting which bucket
  • 02Un-tested strategy defaults to the loudest voice or biggest title
  • 03A testable hypothesis beats a strategy deck
  • 04Reserve deep strategic thinking for irreversible or slow-to-validate decisions
  • 05Speed of learning compounds like interest — daily beats yearly

How to run it

  1. 1

    Write the decision as a testable hypothesis

    State it in measurable terms: if we build X, it will add this many users in this time at this conversion rate, this CAC, and this LTV.

    Pro tip If you can state it as numbers you can test, you don't need a strategy debate at all.

  2. 2

    Check reversibility before choosing a path

    Decide whether the decision is cheaply reversible/testable or irreversible and slow (e.g. pricing tests are fine in the US but constrained in Europe; licenses and compliance can't be A/B tested).

    Pro tip For irreversible or slow-to-validate calls, do the deeper cohort work up front — you won't get a fast signal to correct with.

    Watch out Treating an irreversible decision like a fast experiment can be existential in fintech, healthcare, or compliance.

  3. 3

    Test in-market and read the result

    Run the experiment, watch metrics and p-values to statistical significance, and let the outcome decide whether to scale it.

    Pro tip Keep a light experimentation tool (Kamat uses Statsig) so PMs get a dopamine hit each time an experiment moves a metric.

  4. 4

    Scale winners, kill losers

    If it works, do more of it — that becomes your strategy. If not, stop. Repeat fast enough and 'strategy' largely solves itself.

In the wild

Rosenberg's data rule at Google

Jonathan Rosenberg, Google's product head, told PMs 'come to me with data; if you come to me with ideas, we'll go with mine' — meaning you can bring any idea, but without proof the highest title wins the call.

Reframed strategy debates as evidence contests, pushing PMs to arrive with data rather than opinions.

Common mistakes

Confusing motion with strategy

Treating Porter's Five Forces decks and data-slicing as 'strategy' when it's usually just packaged intuition — the loudest voice then wins, which is maddening for junior PMs who may actually be right.

Taking on 6-month-to-a-year bets

Long projects leave you exposed to macro shifts you don't control (a crypto crash, COVID). Prefer small things shipped fast, and get early signals even on big launches.

Is it for you?

Best for

Product managers and product leaders on growth-type problems where in-market experiments are cheap and fast

Not ideal for

Irreversible or heavily-regulated decisions (EU pricing, licenses, compliance) that can't be validated in days or weeks

From the transcript

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

44:30

strategy is always like keep doing more of it or don't do it right there that's all there is to strategy

42:30

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

44:00

those if they're are either irreversible decisions or they're just extremely time consuming to find out then do some pre-work

46:00

don't take on projects that are going to be 6 months a year. Uh because you just generally don't have control over the macro

1:21:30

From the episode

Unconventional product lessons from Binance, N26, Google, more

Mayur Kamat (CPO at N26, ex-Binance Head of Product)