The document discusses the importance of testing minimum viable products (MVPs) and prototypes in order to validate business hypotheses and learn about customer needs before fully developing a product or service. It provides examples of different types of MVPs and prototypes that can be used, such as concierge-style delivery or landing pages. The key lessons are that MVPs and prototypes should be rough, rapid, and focus on collecting user feedback to inform an iterative development process. The overall goal is to fail quickly and cheaply in order to learn and succeed sooner.
NYU Startup School: Testing your MVPs and Prototypes
1. @NYUEntrepreneur
Startup School:
Testing your MVPs and
Prototypes
Lindsey Gray
Senior Director, NYU Entrepreneurial Institute
Adjunct Faculty, Tandon School of Engineering
Instructor, NSF Innovation Corps (I-Corps)
October 18, 2016
2. @NYUEntrepreneur2
“Starting a startup is
something you !
can only learn !
intrinsically by !
doing it. What you !
need to learn is !
the needs of your !
own users.”
4. @NYUEntrepreneur
Your job is not to validate
your product…
Your job is not to validate
your product…
It’s to validate the
problem, who has it and
then how best to solve it?
11. @NYUEntrepreneur
What is an MVP?
11
“The version of a new product which
allows a team to collect the
maximum amount of validated
learning about customers with the
least effort.”
-Eric Reis
12. @NYUEntrepreneur
What is an MVP?
12
Minimal
Nobody wants
to use it
Viable
Products build by companies
that have no financial limitations
MVP
Minimum feature set
13. @NYUEntrepreneur
What is an MVP?
13
u Typically distributed to customers
u Can be used with larger test groups
u Enough functionality for customer to
use on their own
17. @NYUEntrepreneur
Prototypes can be…
17
u Sketches on paper
u Digital mock-ups
u Handmade constructions
u 3D prints
u Videos
u Simulated experiences
u Virtual models
u …
35. @NYUEntrepreneur35
Prototyping Best Practices
1. Think with your hands
Build to think
2. Test early
Fail often to succeed sooner
3. Collect user feedback
Be ready to be surprised
4. Avoid emotional attachment
Don’t invest in an idea too long
before testing it
5. Rough, Rapid, Right
Quick, dirty, and ugly is ok
6. Iterate, Iterate, Iterate
Build often to evolve
36. @NYUEntrepreneur
What is an MVP?
36
“The version of a new product which
allows a team to collect the
maximum amount of validated
learning about customers with the
least effort.”
-Eric Reis
37. @NYUEntrepreneur
Using an MVP to test
37
1. Define customer and problem
2. Identify key assumption to test
3. Design a test
4. Execute the test (utilize high-fidelity prototype?)
5. Measure test outcomes (customer behavior)
6. Evaluate what worked and what didn’t
7. Pivot or persevere
Melissa Perri
@lissijean
38. @NYUEntrepreneur
Testing Method:
Concierge
38
Delivering the product/service manually
@lissijean
Pros:
• Close to the customer,
easy to learn
• Easy to adjust based on
feedback and learning
• Easy and cheap to get
started
Cons:
• Not scalable
• Time-consuming to run
40. @NYUEntrepreneur
Testing Method:
Wizard of Oz
40
Looks real but manual in the back end
@lissijean
Pros:
• Looks real to customer
• Customer is responding to
what experience would be
if you built it
Cons:
• Hard to get started, higher
fidelity required
• Not as close to the
customer, may not know
WHY behind behavior
42. @NYUEntrepreneur
Testing Method:
Landing Page+…
42
A pitch of your product to gauge idea
reception
@lissijean
Pros:
• Don’t need complete
concept to get started
• Can launch in minutes
• Really cheap
Cons:
• You still need to get
people to your page!...see
last week’s lecture
44. @NYUEntrepreneur
Testing Method:
Video
44
Demo what your product can do by hacking
it together for a video
@lissijean
Pros:
• Can demo complicated
things
• Show a prototype before
producing them all
Cons:
• Harder to put together
• Some parts may have to
be done to include in the
video
50. @NYUEntrepreneur
The process
50
1. Define customer and problem
2. Identify key assumption to test
3. Design a test
4. Execute the test (utilize high-fidelity prototype?)
5. Measure test outcomes (customer behavior)
6. Evaluate what worked and what didn’t
7. Pivot or persevere
Melissa Perri
@lissijean