Why is this startup searching for a business model? Shouldn't you start with that and build your company around it?
Again, apologies if this is a naive POV.
He espouses that you should build a minimal product based on your idea. Then iterate over the following points seeking traction and revenue
1) "Get out of the building" and go talk to a lot of your potential customers
2) Based on what you learn from those people: Reevaluate your idea and decide if you should iterate more, or start a new direction
3) Develop #2
4) goto #1
Think of it as the scientific hypothesis/experiment cycle for startups
[edited to add the below]
As usual, Steve says it more succinctly and powerfully than I. I guess that's one reason why he gets to teach at Stanford :)
Here's how he puts it [1]
Your startup is essentially an organization built to
search for a repeatable and scalable business model. As
a founder you start out with:
1) a vision of a product with a set of features,
2) a series of hypotheses about all the pieces of the
business model: Who are the customers/users? What’s
the distribution channel. How do we price and position
the product? How do we create end user demand? Who are
our partners? Where/how do we build the product? How
do we finance the company, etc.
Your job as a founder is to quickly validate whether the
model is correct by seeing if customers behave as your
model predicts. Most of the time the darn customers
don’t behave as you predicted.
...
How do you know your business model is the right one?
When revenue, users, traffic, etc., start increasing in a
repeatable way
No, he doesn't. He really really doesn't. In fact, that's the exact opposite of what he advocates and what his customer development methodology attacks - the old school product development approach of building the product first and then collecting feedback to refine it, figuring out how to market it, finding out who the target customer is, etc.
Developing your customer base (rather than developing your product) means that you talk to customers and learn their existing pain points before you design or build a single thing. His Four Steps To The Epiphany MBA textbook calls it Customer Discovery, or step one. Everything in the #1-4 loop you described is step two - Customer Validation - where you then try to validate your hypothesis about the customer pain points with your product ideas and/or prototypes.
I'm ripping off this giant rant because for some reason, I see this on HN a lot. A lot. It's really common for people to conflate MVP with customer development and to also reverse the order that you do it. I think you see this on HN in particular because it's filled with coders that prototype in code the way designers prototype on paper, so they think "hmm, how hard could it be for me to throw something together in a week or two and see if anyone will buy it?" instead of "how hard could it be for me to go talk to people and see if there's a potential customer base for this idea I have?" It's not just that the second path is easier, it means that the product you build afterwards is way better informed. On the other hand, MVP seems to appeal to HNers because it's like the Nike message of startups - Just Do It. Just go ahead and build something already because all you need to do afterwards is refinement. A little bit of a/b testing magic on landing pages here, a bit of keyword research on Adwords there and then bam, your app is paying the rent while you drink margaritas on the beach.
Off the top of my head, one of the more recent examples I can think of was just this last month when Zach Burt (zackattack) unveiled Awesomeness Reminders and someone asked him about all the stuff he's built this year, so he wrote about his personal process for throwing tons of little MVP apps against the wall and then why he'd ditch each one before moving on to the next:
http://www.zacharyburt.com/2010/08/an-open-discussion-of-my-...
Not to pick on the guy, but this was a particularly egregious example. Explaining why he abandoned one of his apps (CustomerFind), he wrote:
"I did do some Customer Development meetings related to CustomerFind, and pivots suggested room for a potential enterprise-y Social Media dashboard product."
No, you didn't! You didn't do Customer Development - you did Customer Validation, where you did a little bit of research after you built the product to see if the product you already built solved the problems that you just plain guessed your customers had. In other words, you tried to skip a step and validate your product against customer base that you didn't create in advance and then surprise! There wasn't a fit and the app wasn't monetizable. An interest list based on email signups is not the same as purchase orders. Rinse and repeat for 4 or 5 apps and there went his 2010 so far.
So when do you do MVP and when do you do Customer Development?
Easy - it depends on whether you're building something where you're also the target audience, scratching your own itch, eating your own dogfood, Tim O'Reilly's "fishing with strawberries", etc, or if you're building something for other people.
Both approaches are still just different versions of PG's "make stuff people want" but you need some kind of internal compass to guide the way. If you know the phone in your pocket sucks and you can build a better experience, then Apple can go make something that's clearly better without having to do a lot of market research. So they do. And if you think you'd like to get a call every once in a while and hear a real human being tell you that you're awesome, then there you go. But if you're building something for other people to use that have problems that you don't experience on a daily basis, like say, an app for bloggers that get a good amount of traffic and are looking to get into direct sales rather than relying solely on Adsense's anemic payouts, then you better do your homework and get on the phone.
from: http://steveblank.com/2010/04/08/no-plan-survives-first-cont...
The idea for a startup is usually just a hypothesis. Once you go out and talk to real customers about what you're offering you'll gain some insight and new information. Most of the time, your original hypothesis is off the mark. Your idea may be slightly off, or completely wrong together, but armed with your new knowledge you can alter your hypothesis.
Repeat the process and eventually your evolving hypothesis will become more refined until you have a scalable and repeatable business model. This is the "search" that he's referring to.