Larry Smith: Fail Fast, Fail Often, and Die

Written by Sean Murphy. Posted in 3 Early Customer Stage, Design of Experiments, skmurphy, Video

Larry Smith is an Economics Professor of Economics at the University of Waterloo who writes and lectures on Entrepreneurship, innovation, and Technology markets. What follows is part of a conversation he had with Alan Quarry in the AQ’s Blog & Grill series of interviews with entrepreneurs. His key point, that he makes in a somewhat cranky fashion, is that technology entrepreneurship is a complex undertaking that requires patience, careful analysis, and planning.

Fail Fast, Fail Often, Die

Entrepreneurship is an intellectual endeavor: you must juggle diverse information and manage the sequence and timing of your actions. It’s like playing multi-dimensional chess.

An innovation is a well designed experiment that may fail. You can do a wretchedly designed experiment that is doomed to fail and that is a total waste of time. Entrepreneurs need to think very carefully, both strategically and tactically, about what they wish to do.

Larry Smith in “Fail Fast, Fail Often, Die

Time spent designing experiments that explore feasibility and desirability will allow you to go faster in the long run. You may not fail less often but you will likely extract much more learning from each result and that’s what you have to pay attention to: are you learning faster than your competitors?

It is a complex matter to give birth to something new, it requires the practice of creativity. You have to have passion, an intellectual understanding of the difficulties, and patience.

Patience to wait for ideas to be formed and patience to test ideas before you launch prematurely.

You cannot just try anything. A try anything innovation is a trivial innovation, and while you occasionally get lucky you are not leveraging you intellectual understanding.

Larry Smith in “Fail Fast, Fail Often, Die

Bootstrappers in particular only have limited resources and need to extract the maximum amount of learning and risk reduction from each experiment that they run.

If you are patient you tilt the odds in your favor.

I am not a fan of “Fail often, fail fast” because it leaves one thing out: “fail often, fail fast, die.”

Because that is what happens when you don’t have unlimited resources to fail repeatedly.

If you’re failing fast, how much thought did you put into the endeavor that failed? Probably not a lot.

Larry Smith in “Fail Fast, Fail Often, Die

I think some of this is a side effect of the behavior that is encouraged by “Startup Fantasy Camps” like Startup Weekend and Lean Startup Machine which encourage running may small experiments or minimizing the time between ethnographic observation or conversations with prospects. What gets compressed out is time for reflection and therefore learning.

Related Blog Posts and Talks

Comment by David Telleen Lawton

David Telleen Lawton left a great comment I wanted to highlight:

Thank you for highlighting the messages in “Fail Fast, Fail Often, Die”. Larry Smith’s apropos distinction between thinking, then doing versus just doing is an important one that is often missed on many customer discovery fans…but the entrepreneurs do figure it out…eventually…usually the hard way.
David

Working Without a Plan is Waste

I think there is a growing realization that the original Shewhart cycle of Plan Do Check Act is superior–even for startups–to Build Measure Learn because it creates more waste to leave out the planning step.

Office Hours ButtonWant Help Designing Experiments? Schedule Office Hours
We help you make sense of what you have observed, the stories you have collected, and data you have gathered. We help you form hypotheses and design experiments that tinker with your initial product concept so that you can explore the boundaries and depth of a prospect’s problem.

Trackback from your site.

Comments (2)

  • David Telleen-Lawton

    |

    Sean:
    Thank you for highlighting the messages in “Fail Fast, Fail Often, Die”. Larry Smith’s apropos distinction between thinking, then doing versus just doing is an important one that is often missed on many customer discovery fans…but the entrepreneurs do figure it out…eventually…usually the hard way.
    David

    Reply

  • Tristan Kromer (@TriKro)

    |

    I like this one a lot. A badly designed experiment doesn’t generate data. Experiments that aren’t informing one another don’t make progress. Fail fast shouldn’t mean blind action.

    Reply

Leave a comment

Quick Links

Bootstrappers Breakfast Link Startup Stages Clients In the News Upcoming Events Office Hours Button Newsletter SignUp