Kevin founded Obtiva, web dev company, five years ago
now has 30 employees
every startup has a limited runway
how do you validate a product and biz model before you run out of money?
who are your customers?
what problems do they have?
does your solution solve their problem?
can you do this before you run out of cash?
This presentation is based off of the book Four Steps to the Epiphany
leverage cheap infrastucture
leverage agile software development
iteratively develop software
- instead of spending three months coming up with a product plan/release cycle
- instead release as often as possible: daily, weekly.
- get functionality to users as quickly as possible.
customer discovery - will anyone put a dollar into my hand?
customer validation - how do i add lots of features to get lots of people to put dollars into my hand?
Only at that point do you hire sales and marketing
sales people/marketing campaign / build repeatable practices and systems
Things to remember
have problems so pressing that they're willing to take any solution that you throw at them
this problem i have is killing me, I will pay you anything - I don't care how rough it is
they are your friends when you are a startup
minimum viable product
the most basic expression of your product
may be a set of powerpoint slides that show the functionality of your product that you use to shop around to find a co-founder/angels
may be static HTML
MVP is only used to get to the next stage
eventually you want an MVP that earlyvangelists will pay you to use
changes in assumptions based on feedback from customers
"maybe my customers aren't CEOs but CFOs"
Two Companies - One that Took this Approach, One that Didn't
All the rage in silicon valley about 12 years ago
Raised $100 mil
built delivery fleet
for every piece of furniture we sell, we lose $19
that used up $100 mil real quick
they went out of business
design within reach
also sell furniture
going to print a catalog
going to sell through catalog to other furniture retailers
spent months talking to VCs they said...
"do you have an ecommerce site? warehouses?"
it didn't fit with the common biz model of ecommerce and distribution
found friend and family investing
running $180mil biz today (and today they have e-commerce)
steps of customer development
...launching new industry?
...recycling existing industry?
...or going toe-to-toe with established player?
start with a big vision
iterate over customer problem solution match
iterate over product/market match
what problem am I solving? will people give me any money for that solution?
when you realize your assumptions are wrong - come up with new assumptions - start over
must get out there and talk to early customers - this is the hard part
verify the value that your product adds to the customers
looking for people that will pay because they need your solution
Build the following matrix with your early customers:
Problem: students didn't pay attention in class and can't write term paper
Current situation: find a friend to help (unreliable)
Proposed solution: we will sell you a term paper
Problem: student needs passing grade in class, but is not a strong writer
current situation: hire tutor or pay $25 for a ghost writer
proposed solution: we sell you a term paper for $9
Want to answer question:
do we continue on, or do we pivot?
make small changes
be very specific about who your buyer is
if you're not getting reception on who your buyer is, make tweaks
often buyers can't articulate the exact problem they need solved
this is part of the reason it takes many tweaks to get there
Q: Are investors embracing lean startups?
A: (from the audience) "Hot buzzword in the valley right now is capital efficiency" - how much bang can I get for my buck?
Q: how do you get early customers to give you their time when you don't have a product
A: B2B - they will give you time if they have a huge pain point. B2C - use adwords etc.
Q: How do you protect against competition and IP theft?
A: You can still use NDAs (I'm not sure if I agree with this)
Q: A/B Testing for iteration?
A: Great way to test more hypothesis.