Sunday, September 16, 2012

How a Startup Works


Do you have an idea for a business?  Are you doing a startup?  Here's how it works.

The Myth of Planning.  The widespread, fundamental, and persistent myth and misconception about entrepreneurship is that a startup is just like a larger company and needs to do the same things as, say, General Electric, but on a smaller scale.  The core competency of large companies is that they plan and execute well; smaller companies therefore should imitate this core competency and start out with a great and  detailed business plan and "Just Do It".  What sounds so right is so wrong and here's why:

The Reality of Uncertainty.  The history and accumulated statistics on startups show that a large number fail.  Among those that succeed, they DO NOT succeed with Plan A nor frequently with plan B.  They ride a rollercoaster of plans and ideas and strategies until, before running out of money, they find a plan that works.

Discovery and experimentation is Job One.  Therefore, however non-intuitive and however contrary to everything you may have read or learned, the hunt or search for an executable plan--and not the actual execution of one--is Job One.  Do not write a business plan!  In fact, perish the thought for the time being and let's talk about business models.

Address uncertainty with a business model.  An entrepreneur starts out with a vision of a product to be built, a service to be delivered, or a customer desire to be satisfied in some unique way.  Inside this big vision are one hundred question marks.  What customer problem should I solve?  Is there a market?  Can I compete?  If I build it, will customers really use it?  Will they pay for it?  Can I afford to make enough of them aware of my solution?  Can I make a profit?  Can I get buy-in from partners?  Do I have the resources to pull all of this off?  

A business model is a set of guesses.  A business model is a set of guesses that flow from your vision.  Unlike a school report card where four A's and a C is still pretty good, one bad mark, one NO answer from ANY of the above questions can spell failure for your business model.  What does this mean?  It means that we need to think like a scientist, tread our business model as a hypothesis, and then design and execute  experiments to test it.

Experiment.  Experiments drive the entrepreneurial process.  Many of our core assumptions need to be tested directly with customers.  Customers may tell us that our solution is right or that they don't care or that they care about a different problem we can solve.  

Iterate or Pivot.  After you have done an experiment, you have learned something about your business model that either suggests that your model is valid or it is invalid.  Based upon these results you have three choices.  If the data was not clear, you can get more data and talk to more customers.  Second, you can iterate--make small changes to your model in line with what you learned.  Or, third, if you learn that your guesses were completely wrong, you can throw out your business model and start with something very different.

Just Do It, But Lean.  You still need to "just do it."  However, you will probably need to just do it over and over again--in short iterations.  This is what we call lean--rapid cycling through the guessing and testing cycle so you can find the hard-to-find  good idea.


  • Sketch out a business model (plan A)
  • Test with customers using surveys or interviews or prototypes or other techniques
  • Review the results and determine which parts of your business were validated or invalidated.
  • Either update your business model or reject it and create a new model.
  • Plan and execute more experiments.
  • And then again, and then again
Good Luck!

1 comment: