— Tony Ulwick (@Ulwick) January 3, 2013
…and a blog post is born.
By now, most of you have probably heard of the Lean Startup movement. To summarize as succinctly as possible, it’s an iterative process of hypothesis-driven experimentation. The term “lean” implies linkage to the continuous improvement concept coming from the Lean Manufacturing world. In the case of Lean Startup, however, the focal point is business models and/or products for startup companies.
The driver for this new way of thinking is the high percentage (70-90%) of product launch failures. Companies and entrepreneurs are desperately looking for ways improve the rate of success in the market; and as a result we see innovation experts coming out of the wood work. It appears they are all finally ready to share their secrets with the world!
In the Startup Owner’s Manual, Steve Blank marries his Customer Development concept to Agile engineering to create a better and repeatable approach for startups. In order to produce repeatable results, a process is introduced, and this is not the traditional process that has failed us for so long, it’s a new more iterative process that uses terms like pivot, learn, fail and learning more. It begins with a step that “first captures the founder’s vision and turns it into a series of business model hypotheses.” It then moves on to testing the hypotheses; and one must assume that failure is expected, or why else would we need to iterate?
Actually, failure is expected. It cannot be avoided according to this theory; so we must all embrace it, learn from it, and do this faster than our competitors. The Agile approach to process improvement has worked so well for the past 50 years that it simply must be the answer for startups and founders. But there is one significant problem…
As with most other innovation processes, the fail fast approach to product development is a process that employs the wrong sequence of steps, and is actually missing the most important step of all. If you begin innovation with the wrong set of assumptions, there is no guarantee you will be able to iterate to the right set of assumptions. Chaos theory explains it very well…
If we are to seek stable and predictable outcomes, we need to control those initial conditions. With respect to innovation (which is the main focus of startups), the high variability of initial conditions stems directly from the high variability of founders’ ideas; or as it’s called in the fail fast world, hypotheses. One could suppose that getting out of the building to test hypotheses with customers (not really customers yet) is better than letting your marketing organization run wild with a bad idea. However, how many hypotheses does it take until you get it right? Is there any guarantee that you started in the right Universe? Can you quantify the value of your idea? I’m sorry, I meant hypothesis.
How many times does it take to win the lottery?
You can find many experts (especially venture capitalists who are fine with a 10% success rate) cherry picking fat vs. lean stories showing that lean startups succeed at a much higher rate. They will use examples like DropBox and ignore the vast sea of unsuccessful lean startups that failed really, really fast. The missing (or out of sequence) step is understanding the customer need; or I should say needs. Needs are the metrics that describe success; plain and simple. How many founders can describe success for the customer?
It’s not just innovation that has failed as a process, other types of consulting are constantly lampooned as bottom-feeders.
The missing link in all of this is our inability to agree upon the definition of what a “need” is; plain and simple. Each and everyone one of you (me included) is guilty of latching on to mental models that make us feel good; whether they get the job done or not. And speaking of jobs, let’s turn our focus to what is really important here. No, not a founder’s hypothesis, but the job a “customer” is trying to get done. This is the one stable concept we can focus on to ensure that we have a solid target to shoot for.
A job, like “listening to music” doesn’t really change much (if at all) over time. It is a process that we go through, and each step of the process has metrics that we use to define its perfect execution. It’s the steps where execution isn’t perfect that we may find opportunities to innovate. To do this, however, we must agree on what these metrics are, and that understanding each and every one of them is critical in finding opportunities to co-create value with customers over time. Tony Ulwick, of Strategyn, uses a simple example of the process of listening to music to demonstrate this. Any of you, like me, who grew up in the turntable era should be able to appreciate it…
When MP3 players came along, the job didn’t change, but they helped get more steps done better…
…and when streaming services came along, the job still didn’t change. But they helped get more steps done better…
In order to provide innovative solutions to problems, we must view the problem as a process that a customer follows to accomplish something. The needs – that we seem to have such a hard time identifying – will describe the perfect execution of that process along dimensions like speed, stability and output (that’s lean thinking by the way). It’s the customers’ evaluation of these needs which provide the gaps (the problem) we must address to innovate. It may not be possible, today, to satisfy needs at certain steps as well as other steps, but understanding them, today, is critical as new capabilities and technology develop because it gives you the advantage over others who are still just guessing, and failing.
What does a need look like? It looks like a metric, not a feature.