I will not attempt to explain what Lean is, or ought to be according to any canonical definition, as there are already too many voices in this tangled confusion of jargon and I’m not sure that my contribution boosts the signal. Also, being frank, I don’t really care.
What I do care about is how people think, myself included, and the importance of rooting efforts in some kind of plausible reality rather than whimsical fancy or, worse still, institutional bubble. The latter is particularly hard to escape, as I’ll touch upon in a related post in this series of short posts about scientific realism.
Sadly, I have discovered that many in apparently educated circles do not really know what science is. For some, it is nothing deeper than a class of activities, or subjects, like physics and chemistry. As ever, the word “scientific” is used liberally to mean somehow more truthful or logical than alternatives. Lean is often referred to as scientific.
There is a general familiarity with the so-called scientific method, although I have expressed my skepticism about the concept (in terms of it’s colloquial use as meaning a kind of conveyer belt way of exploring truths free of human whim). Perhaps this is what is meant by “scientific” in this case.
But perhaps there is a philosophical body of work that might shed light on how scientific Lean really is: scientific realism.
Scientific realism is a set of beliefs, or a philosophy, that underpins how we might view the correspondence between scientific explanation, which is a construct of our minds, and reality. In short, the claim is that science really does explain reality.
For many, scientific realism is irrelevant because when we do science it gets the results we want, like aircraft that fly. This attitude is fine of course, although it does diminish human potential for discovery because it can easily lead to a kind of highly-functional non-reflective participant, as many “smart” folk are. (I call them clever idiots.) In this vein, one might argue that Lean is just common sense as it offers a promise of how to maximize business efficiency, at least in one way.
Three core elements of scientific realism might help us to think more clearly about Lean. From there on, the reader can decide the meaning and value of a contentious idea like “Minimally Viable Product”, if it means anything at all.
Firstly, scientific realism assumes that there is a reality “out there” (i.e. independent of our minds). This might seem obvious, but many schools of philosophy will disagree, as does the film The Matrix (based upon the “brain in a vat” hypothesis that is not without some validity). This feels correspondent with Lean as presumably there is a reality “out there” that is the actual event of people experiencing the proposed product. Of course, we don’t know that reality yet. It is unobserved.
This leads to the second element of scientific realism, which is that it assumes two classes of events in the world: observables and unobservables.
Observables are what we can grasp through direct experience (using our senses), like how a lump of hot coal burns.
Unobservables are what we cannot grasp, such as light bending under gravity and dark energy.
However, there are other classes of unobservable event, like the future. If we knew what products people would like and buy, we’d know the future. Clearly we don’t.
Another class of unobservable event is what I’d broadly call aesthetic response.
It can be highlighted by the common reaction to an unseen work of art. Outside of very narrow limits, it is often impossible to predict if someone will like a particular work of art. The viewer herself will seldom know that she will like it until she sees it and confirms a broad truism: “I know (or like) it when I see it.”
There are myriad reasons, most of them still unknown, as to why this is.
It is insufficient to merely predict that something might be of aesthetic interest to a viewer if the ultimate test is whether or not she will buy the object. Actually, these are almost two entirely separate realms of concern.
The cognitive mechanics of deciding to buy a product, per a large body of work called behavioral economics, appear to operate an a plane that is arguably often independent of the product itself.
These two factors alone, let’s call them the aesthetic response of a product and the economic response (broadly speaking) often combine in entirely unexpected and unpredictable ways, particularly when the product is novel and therefore fails to offer a familiar frame (or set of heuristics) that might guide decision making. Much of this cognitive reaction seems guided by how our predicting minds handle surprise (which any novel product will introduce).
What has this got to do with scientific realism?
When dealing with unobservable events, which includes novel product experiences in the future, the only known means to produce some kind of knowledge about the world (the epistemological aspect of scientific realism, which is the third element) is to carry out experiments that in some way confirm, or deny, expected outcomes.
One method of experimenting is to bring about the future event itself – i.e. to build a product and put it on sale. But this has a cost, which may or may not be affordable. Much of the time, the opportunity cost (of missing the market or wasting human resources) is the bigger cost, but, in pragmatic terms, the actual cost is what usually counts.
A cheaper way is to simulate the experience of the novel product for the customer.
The goal then is to maximize the correspondence between the simulation and the actual final outcome with the highest confidence at the lowest cost, or at an acceptable cost. Fulfilling this goal is what might be called Lean. Figuring out the least cost simulation that yields sufficient information to continue, or not, is actually the minimally viable product.
In the ways just discussed, Lean is semi-rooted in the philosophy of scientific realism. However, in practice it is seldom approached with the formal rigor that science might bring to the table. In fact, I will argue shortly that it isn’t really true to science at all and is more of a logistical optimization method than a scientific method, rooted in a kind of pragmatism that is only concerned with one type of knowledge: how to get to profitability at minimal cost.
In my experience, product teams do not think of the Lean process as attempting to simulate the final outcome. (This is leaving aside the discussion of how far the “I like it when I see it” effect dominates how we think about products and thereby effectively renders many simulation approaches useless.)
What often happens is not an attempt to simulate the experience, but rather an attempt to simulate a theory. For example, we can easily believe that by merely presenting something to potential customers, we are doing some kind of statistical sampling that has the power of confirming a theory. We put emphasis on the imagined power of sampling rather than the effectiveness (and design) of the simulation in its ability to simulate realism.
What I have seen in many of the discussion groups about Lean is not a desire to achieve cost-effective simulated realism, but a desire merely to “talk to customers” per Steve Blank’s advice to “get out of the building.”
The statistical illusion occurs by a arbitrary belief in numbers. Asking one person is obviously dumb, as is probably three, five or maybe ten. But asking 100 people? Well, that seems plausible doesn’t it.
This belief in numbers is so prevalent that it leads to quite ridiculous simulations, mostly in service of another false god: brevity. It is almost considered heroic in some Lean circles to have the verve to stand on a street corner and talk with one hundred people merely with a clipboard and scrap of paper. How “minimally viable” is that?
As I mentioned art earlier I feel obliged to draw the parallel of testing the Mona Lisa merely by presenting a sad stick figure on a scrap of paper. If you think this absurdly reductionist, then please do stop by the Lean talk shops to hear the stories folks often swap.
The problem here is a confusion of ideas. Testing ideas or entrepreneurial hunches is not the same concept as testing product experiences, never mind the actual problem of testing economic commitment. In fact, we should dispense with the term test altogether as it is misleading. What we should be concerned with is the design of the simulation.
However, this is not the dominant approach because there is no real scientific realism here. The Lean methodology is guided by a different philosophy. The approach is to build your way towards a certain business state in sufficiently small increments so as to achieve a kind of convergence (via approximation) on a stable state.
In this regard, the Lean method believes that there is a kind of landscape, or state space, with various local maxima that each represent a viable business in some fashion, mostly measured by a satisfactory return on investment (via product demand). The goal then is to follow some kind of gradient ascent process to find the nearest local maximum from where the entrepreneur started. Course corrections are called pivots.
Lean is therefore better described as a “statistical optimization” method or even “control” method. As an engineer, I find the control metaphor more apt because control engineering is about using feedback loops to create stability. The concepts of overshoot and undershoot are directly applicable, as in spend too much money and miss the mark or spend too little money and also miss the mark.
To me, this is quite divergent from what scientists generally do.
My point is that if you call something a scientific method, as Lean advocates do, then it implies a certain philosophy, like scientific realism. Or, if you prefer, it implies a certain framework of understanding.
And, as we know, frameworks can affect how we think.
The problem with calling anything scientific is that the term is often understood as a gloss for truthful, logical, rational, or other concepts. In doing so, it tends to mask the reality of the method’s natures and claims and so easily leads to faulty execution, like confusing statistical sampling with reality. We do irrational things that are masked by the purported rationality of the method.
If we explore other frames, like the simulation frame, then we might decide that other methods could help us get where we want.
What strikes me as odd about many of the Lean fora is that there is seldom any discussion about using existing knowledge or other processes to design a product. For example, many (and perhaps most) of the greatest products out there were simply designed by an apparent “visionary” using insight rather than via a statistical optimization technique.
Often these designers employ very different types of knowledge, more anthropological and psychological in nature whereby they formulate hypotheses about human reactions to products by understanding how humans work. This is very different to Lean, at least from my readings of it. It doesn’t seem to have much to say about where to start on the state-space of product demand nor how to actually determine the hypotheses that drive the pivoting process, nor much about how consumers work. It is a very crude gradient descent without any attempts to supervise or tune the process. The role of other techniques, like creativity, are somehow omitted. It is really a kind of blind optimization.
This is another pitfall of calling it scientific. Recall my earlier post about the scientific method wherein I spelled out the huge and apparently unsystematic role of insight and creativity in scientific exploration. Much of this comes from a certain mindset and outlook on the world that has more to do with curiosity and being prepared, per Chomsky’s advice, to be puzzled by simple things (i.e. things that seem obvious, but are only obvious by convention).
The kind of outlook I’m talking about is closer to artistic sensibility and a deep understanding of aesthetics (in the broadest sense). It is more like what Verganti describes in his book Design-Driven Innovation as the process of “Interpretation”.
In the world of art, there is a deeper understanding of the concept of metaphor. Indeed, this is probably also true of cultural taste makers.
There have been various attempts to understand the role that metaphor plays in product interpretation by users to the extent that one technique for simulating product response is to test applicable metaphors. Going further, Zalman has described a method that attempts to elicit metaphorical understanding from the customer’s mind, at least as it might apply to a set of states in the product state space.
Is Lean scientific? Kind of, but I think that this claim causes us to lose sight of other ways of thinking more clearly about product, ways that would serve the product designer and entrepreneur well.