Iterative approach to making startups successful
Rafi Dudekula ,
06 / 28 / 2011 ,
lean, startup, success
Lean startups are about testing and avoiding incorrect market assumptions as early as possible. The questions to ask in this context are: how to reduce the work required to assess assumptions about the market and how do we decrease the time it takes the business to find market traction?. In the past I abandoned reasonably well developed startups of my own since I did not have better mental models or understanding of the odds of making a startup successful.
After reading this post on lean startups, I started thinking about a quantitative model to represent success percentage of a startup that has an assumed failure rate. Success rate depends on both personal and environmental factors. For example for a serial entrepreneur it may be higher, for a B2C company that has intense competition it may be lower. One way of modeling success rate is to use geometric progression as follows. If x is the assumed failure rate of your startup, then the success percentage at nth iteration of learning is 1 minus x to the power n or (1 - x**n) * 100%. The assumption here is that you are building on the learnings from all previous trials and not starting from scratch on an entirely new idea after every few trials.
Lets say you estimate your failure rate as 95% which makes x = 0.95, then using the formula above, your success percentage after 10 trials is 40%. After 18 trials, the success percentage is a good 60%, much higher than 5% success percentage you started with. When your startup gets to greater than 60% success rate you are able to acquire and keep more customers and hence may experience success. The beauty of this model is that given enough trials that build upon each learning, no matter how large your failure rate is, you will eventually succeed. To be practical though your hypothesis to market cycle should be small enough. That is you are following lean startup model of agile software development plus customer development. If each trial say takes 1 year then you may take 18 years to get to 60% chances of success. Lets say you decreased the scope and automated testing/deployment and somehow got to a 1 month trial period. Now with a trial period of 1 month your startup can succeed in just 18 months. This means you keep building minimal viable products every 1 month based on the customer feedback you get.
So another definition of a lean startup is, an entity that implements a systematic process of releasing features that customers value and learning from failures in a reasonably short time. Your startup is well on its way to success if it does intentionally focused trials while dutifully using analytics to measure key performance indicators and reducing cycle times to deliver features that customers want, before the money runs out. I will be testing this theory out myself on my own startup AgileSense.
Workflow applications and what makes them flexible
It is common among programmers or developer to create data driven applications using a group of forms with buttons. These group of forms can be accessed via menus or buttons. The application flow models specific business process flow.
Notes on what makes entrepreneurs entrepreneurial
How do entrepreneurs reason about specific problems in transforming an idea into an enduring firm? According to Professor Saras Sarasvathy, entrepreneurs engage in a distinct form of reasoning and thinking that she terms effectual reasoning as opposed to causal reasoning.
2 responses to "Iterative approach to making startups successful"
Leave your comment
Browser compatibility
AgileSense platform uses JSON data structures communicating with RESTful applications. The application screens are all web based and work with modern versions of the following browsers.
Lee
This is an interesting way to look at success from a more quantitative perspective. Let's just assume the function applied to experience is correct. My question would be, is there an upper bound how much it can contribute to the probability of success? If there is no upper bound, it would seem that with enough experience you could approach guaranteed success no? If there is an upper bound of it's contribution is that bound significant to any conclusions drawn?
July 06, 2011 10:41 am ReplyRafi
Thank you Lee for your comment. Assuming your definition of success is reasonable financial gains, I believe one can approach guaranteed success, by building upon or learning from failures. The problem is developing products/services that customers are willing to pay for and better yet become raving fans of, is very hard and time consuming. Many entrepreneurs give up after a few trials. My intention was to motivate myself and other entrepreneurs out there to keep learning from failures fast enough by shortening the customer development and agile product development cycle.
July 07, 2011 11:50 am Reply