How to Structure Good Hypotheses for Your Lean Startup

Crafting good hypotheses for your startup is hard. Most people focus on solutions rather than problems. That leads to a ton of products getting launched with zero traction.; the all-too-common “solutions looking for problems.” A good hypothesis is important because it leads to good experimental design. Good experimental design is important because you need it to properly validate or invalidate what you’re doing.

A good hypothesis needs to be written down. It’s amazing how few people to do it, but the simple exercise of writing things down is significant. Just do it.

Try structuring your hypotheses this way:

“I (or We) believe…”

Finish that statement and see what comes out of it. Each key element in that sentence is a variable in your experiment, and potential feature/component of your MVP. Each variable in your experiment has to be properly tested. If a variable passes a test it may very well become a cornerstone of your value proposition. Remember: The statement has to be testable, and it has to have the potential of failing.

Here’s an example of a hypothesis I might have for NextMontreal (which is primarily a content site on startup news for Montreal):

“I believe that U.S.-based investors will read NextMontreal on a regular basis because they have a growing interest in the Montreal startup ecosystem.”

The basic structure is this:

I believe [target market] will [do this action / use this solution] for [this reason].

You can go more high-level on the hypothesis. Doing so can often save you a ton of time, money and heartache. Here’s an example:

“I believe that U.S.-based investors have a growing interest in the Montreal startup ecosystem.”

It would have definitely been possible to test this before launching NextMontreal. I could also go more granular on this statement and segment the market of “U.S.-based investors.” I might hypothesize that East Coast investors are the right target market because of their geographic proximity to Montreal.

Now we get to the solution: NextMontreal.

The hypothesis is that the target market will “read NextMontreal on a regular basis” but what does that mean? Daily? Weekly? Monthly? It’s important to set a baseline goal that you’re looking to achieve, which in your opinion validates the solution and value you’re creating. This is very hard for a lot of startups to do because they don’t know what goal to set, but it’s an important exercise to go through. Draw a line in the sand.

So putting it together, we have:

“I believe that East Coast U.S.-based investors will read at least one article per week on NextMontreal because they have a growing interest in the Montreal startup ecosystem.”

That’s a fairly well-structured, reasonable and testable hypothesis.

Putting “I believe” up front helps get the sentence rolling. It’s also emphatic. It’s a bold way of making a statement – and hypotheses should be bold. You want to believe your hypotheses are true and have the conviction to define them clearly. But you also need the awareness, intellectual honesty and diligence to test your hypotheses vigorously and accept when they fail.

May 5, 2011 Posted in Customer Development by

Ben Yoskovitz
I'm VP Product at Codified (makers of VarageSale).

I'm also a Founding Partner at Year One Labs, an early stage accelerator in Montreal. Previously I founded Standout Jobs (and sold it).

My bio »

or

Follow me on Twitter

Get updates and special content
When I publish new content, get it directly in your inbox. Subscribers will get special stuff as well not available on the blog (but I promise it will be infrequent + high quality.)
Get the Lean Analytics Book!
Awesome Jobs
Check out the job opportunities at my portfolio companies.
Startup & Investor Resources
Find Stuff
My Photos