A rat is a rodent that resembles a large mouse. But in the startup world, Riskiest Assumption Testing (RAT) is a feedback loop thats used primarily for working through your product development assumptions.
You start by 1/ Looking to learn something --> 2/ Building something to learn from --> 3/ Measuring things from the thing you built --> 4/ Rinse and repeat
Until you either support the assumption you had OR you learn something net-new so you make new assumption to learn from.
But beyond product development, I’m making the case that RATs can be applied to other functions like culture, management, investor relations, like I have. At its core its a vehicle to systematically kill your bad ideas while gaining validation (and momentum!) for the winners. And that’s something every person does at work.
RATs tend to do better than MVPs, especially for early stage companies because:
No Figma mockups. No no-code MVPs. Not before we all get aligned as a very small team with no time, no budget.
Even if they don’t come out and say it, the people you’re building with and for are asking these questions (or realistically, complaining about them behind your back). Do yourself a favor and get ahead of it. You’ll be much better off when you have data to support your decisions.
You probably already know the things you’re assuming about your market, your product, your competitors. So this is not a question of “What are the assumptions?” but one of “Which are the assumptions I should be focused on right now?”
Rather than prescribe an entire algorithm to figure that out, the rule of thumb I tend to use is “What can we NOT be wrong about?” That really helps boil down where you should spend your time validating assumptions quickly.
We all have amazing product instinct (right?). But RATS make you “prove” your gut instinct is on track. If not, you still can win because it’s a fantastic method for capturing structured learning with an applicable output (e.g., your product, your team, your fundraise, etc.).
RAT Examples: For example, here are some RATs for both real and fictional companies:
(Table here)
Note: The RATs above should be even more pointed, assessing 1 assumption at a time. All of the examples have multiple variables which can lead to confounding results.
Below is a small cheat sheet to get you started.
That’s it*! You’re an expert!
*The word “it” is holding a lot of weight in this sentence. But that’s enough to get you into trouble.
Here are a few RATS to inspire you. For ease of understanding, I've limited to brand name products that many have used or heard:
PRO-TIP: Don’t make your RATs convoluted. The more focused, the better. Think about it like a product/business-centered scientific hypotheses as a mental model. What are your variables? What are you testing for?
