Feedback on experiment #1 | Solving problems in student health education loans

At the milvalchal we are attempting to prove several things... the most important of which that Eric Ries' LeanStartup "works". We have defined what we mean by "works" in another post, here we will break down the first experiment run by each of our teams so that others may learn.

Jossy is a brilliant entrepreneur based in the Philippines. Jossy is trying to solve a very important problem in healthcare.

After talking to 100 students, it was time to state a testable "action hypothesis." The table below is an excerpt from their experiment design document.

target number of interviews conducted with this specific segment prior to this experiment will perform this action 
Students studying any health related course in Philippines and people that would like to study one but don't have the finance100Will signup and become the first badge of students to seek loans

MENTOR NOTES:
  1. The customer segment could be stronger and more refined. We try to avoid the use of the word "any." We also try to avoid generic "lack of the solution is my problem, and lack of time, money, etc." That said, the seed of a testable hypothesis does exist here. The entrepreneur believes, after 100 interviews that people actually want to study medicine but it is the structure of loans that is preventing them, not something else. 
  2. The founder also believes that the students themselves are frustrated and are not artificially attributing a societal problem to the students. 
  3. After the 100 interviews - the first thing we look for is the ACTION that the SPECIFIC CUSTOMER SEGMENT will take, that proves the initial hypothesis - or at least reasonably suggests that the hypothesis is true... in this case the entrepreneur believes the customer will:
  4. "signup and become the first badge of students to seek loans" - this is where we as mentors risk sounding pedantic, but specifics are EVERYTHING in this experiment. 
  • Will sign up - great call to action but what will they ACTUALLY do - what constitutes signing up? The higher the value of the trade, the more evidence that there is true demand. Will they just add an email address? A social security number or equivalent? Fill out a long form for a loan?
  • "To seek loans" -- again at the risk of sounding pedantic - "loans" is not specific enough to measure a UNIQUE value proposition. 
  • How many? The entrepreneur left out the most important part of our request for funds... How many people will fill out the form, how many will sign up during the one month period that the experiment is running? It is important to have an idea, even if you end up being way off base. 

In order to get funding to run this experiment, we ask the entrepreneur to tidy up loose ends and then we fund the necessary amount! We fund them at an automatic $1,000,000 valuation for agreeing to go through this rigor, for helping to train others, and most importantly because we believe they are worth it. Anyone - even you - can get money to run an experiment like this one. Click here to accept the challenge. 

MENTOR FEEDBACK, Take two


  1. ON YOUR CHANNEL (Facebook ads) - Paid ads are usually lazy and expensive in the beginning but they can be fast, and fast is a key principle of lean. Make sure they are laser focused. Do not say that you are COMING SOON, say that you are READY NOW to provide value. You are. If you have a big issue with saying that, read this
  2. ON YOUR LANDING PAGE (Unbounce) - do a better job of explaining the UNIQUE value proposition. Why should they get a loan from you? What are the terms? What is the benefit to them? People constantly think that their conversions are low because of "design" or because of "the URL" but usually (especially for the true early adopters, the people who told you in your 100 interviews that they had a huge problem they wanted you to solve), a clear mention of the value proposition is usually enough to intrigue them into leaving their email address and filling out a form. 




Post a Comment

Popular Posts