# Hypothesis-Driven Thinking, Explained

> What hypothesis-driven thinking is, where it came from, when it speeds you up - and when it blinds you. With a worked example and a playable version.

Source: https://www.wavetable.net/frameworks/hypothesis-driven-thinking

[Frameworks](https://www.wavetable.net/frameworks)

StrategyAlso called Answer-first problem solving

# Hypothesis-Driven Thinking

Start with your best guess at the answer, then spend the time you have trying to prove it wrong.

## In one look

1.  1Best guess first.
2.  2Test what would kill it.
3.  3Revise and go again.

It's borrowed from science: state the claim, design the test, go where the evidence points. Strategy consulting added the deadline. A scientist can wait for certainty and a client project can't, so firms took to naming a provisional answer on day one and spending the weeks testing it.

The problem it solves

## Six weeks of research, and a finding everyone already knew.

Four analysts, forty charts, and a readout that says usage is down among younger members. Without a guess to test, every piece of analysis looks worth doing. With one, the risk is the opposite: you start collecting evidence that agrees with you.

## Try it yourself

No sign-up, a few minutes. You run a campaign's sign-up page, and the team is sure the video will win. You pick the test.

## The same six weeks, with a guess

A loyalty program in decline, and a team asked to find out why.

-   Researching the topicWe're researching the decline.becomes Day one: "Members earn points but can't spend them."
-   Researching the topicForty exhibits on who uses it.becomes Two weeks on redemption rates, abandoned redemptions and five member interviews.
-   Researching the topicUsage is down among younger members.becomes Half right. Redemption is hard only past a points threshold, and that's what they fix.

## Good times to run it

-   The first week of a diagnostic or strategy project
-   Evaluating a deal. Say why it works, then attack that.
-   Choosing what to analyze when the deadline can't move
-   A steering meeting that needs one claim to test

## When to skip it

-   When you like your guess too much. You'll stop testing and start collecting support.
-   Open exploration, like user research or a new market. An early guess narrows what you notice.
-   Clients who need to see an open mind. Some hear "hypothesis" as "you'd decided before you arrived."

Teams we've run sessions with:

-   ![Publicis Groupe](https://www.wavetable.net/logos/publicis-groupe.svg)
-   ![SoundCloud](https://www.wavetable.net/logos/soundcloud.svg)
-   ![WeTransfer](https://www.wavetable.net/logos/wetransfer.svg)
-   ![The New York Times](https://www.wavetable.net/logos/nyt.svg)
-   ![Columbia Business School](https://www.wavetable.net/logos/columbia.svg)
-   ![EY](https://www.wavetable.net/logos/ey.svg)
-   ![Havas](https://www.wavetable.net/logos/havas.svg)
-   ![NYC](https://www.wavetable.net/logos/nyc.svg)
-   ![MCI Group](https://www.wavetable.net/logos/mci-group.svg)
-   
-   
-   
-   
-   
-   
-   
-   
-   

## Run it with your team

That was someone else's campaign. With your own guess, one you've already said out loud to the client, it's harder. We run it as a live session on a question like yours, in small groups with a host.

[Consulting firms](https://www.wavetable.net/for/consulting-firms)[Strategy summits](https://www.wavetable.net/for/strategy-summits)

## More frameworks

-   [Structuring thinking**Issue Trees**Break one big question into smaller questions that don't overlap, until each is small enough to answer.](https://www.wavetable.net/frameworks/issue-trees)
-   [Structuring thinking**First Principles Thinking**Break a problem down to the facts you can check, then reason up from those instead of from what's usually done.](https://www.wavetable.net/frameworks/first-principles)
-   [Structuring thinking**MECE**A test for any list: no item overlaps another, and nothing important is missing.](https://www.wavetable.net/frameworks/mece)
