10 Rules To Help you Design Better Experiments
While it’s in the DNA of most entrepreneurs to dream big, Upperstory puts a strong focus on ensuring that we follow the critical steps to get there.
Most of the time these critical steps come in the form of experiments, which drive our decision making process. For us, there is no such thing such as “truths”, but a set working hypothesis that we are desperately trying to translate into facts.
This post outlines a set of rules for implementing effective hypothesis statements. You can use this method to validate your business model, streamline your sales channel, nurture your business culture or even develop your professional career. Once this tool is under your belt, there is no stopping you from applying it anywhere.
RULE 1
A hypothesis represents what you strongly believe about your internal and external customer.
RULE 2
It falsifies an underlying assumption of the business through a formal test.
RULE 3
It maps to whichever canvas you desire to use. For example, you may use the Lean Canvas, The Business Model Canvas or the UVP Canvas. Either way the mapping is required.
RULE 4
The test shouldn’t conduce to self-fulfilling or leading answers. There is nothing worse than validating what you already know, or asking customers what they think they know, but don’t.
RULE 5
It defines a success criteria based on key business metrics.
RULE 6
The success criteria is always an educated guess of what you think will be a desired outcome. It’s like drawing a line in the sand.
RULE 7
The bar for the success criteria should always be set high.
RULE 8
Validation or invalidation of hypothesis should start by the one that is risky enough to make the whole business fall apart.
RULE 9
New insights that emerge during testing reinforce previously valid/invalid hypothesis, or creates a new hypothesis altogether. During and after validation, record the learning and establish what is next in terms of experiments. New data might affect other assumptions in your Canvas.
RULE 10
All problems have a solution, but not all solutions have a problem. Don’t run a solution test if you don’t have a valid problem hypothesis first.
PROBLEM-SOLUTION FIT (EXAMPLE: AIRBNB)
Hypothesis 1
Travellers have bad experiences because photos in marketplace platforms look different than reality.
Lean Canvas:
Problem
Test Question:
When was the last time you had a bad experience when renting an Apartment or BnB service online? What were the top 3 causes?
Expected metric:
80% of respondents indicated that misleading pictures led to a major disappointment.
Hypothesis 2
- Customers sign-up to service due to a good quality of photos in inventory of houses.
Lean Canvas:
Solution
Test:
Take 10 beautiful images of apartments that accurately represent the rental experience. Make them visible from landing page.Push through a channel and A/B test.
Do you need a platform to better manage your experiments?
