*This is one in a series on the topic of Process in Product Development
In the E for Electric (vlog) episode released on December 29, 2019, Alex Guberman interviewed guest Carsten Breitfeld, who at some point ran the BMW i8 program, former CEO of BYTON and at the time of the interview, CEO of FARADAY FUTURE. The title of the episode was “Why Best Electric Cars are Made by Start-ups”. In the conversation, Mr. Breitfeld highlighted a flaw in the way big auto legacy manufacturers approach development: “Big companies try to build everything into a perfect process… to some degree they try to get independent from people”. In short, minimize human input, so as to minimize potential human errors and likely also other disturbances that can negatively impact innovation. So, what are the consequences, good or bad, of limiting human input, altogether? In Lean Manufacturing there is a tool that has a very similar goal of minimizing variation due to a set of inputs (multi-variable function), including the human factor, called Gage R&R. Anybody who has studied the Six Sigma methodology or Manufacturing Statistical Process Control (SPC) may know about Gauge Repeatability & Reproducibility:
R-epeatability: of measurement system
R-eproducibility: reproducibility between different users (the human factor)
From Manufacturing Principles to Development
Without implying that Mr. Breitfeld perceived that legacy automakers were in fact applying a tool like Gage R&R outside of the manufacturing context, let’s take a quick look at one of its formulations and its potential application to Product Development, as is (for more information see reference [1]).
There are multiple formulations and methodologies for Gage R&R studies, two of the most common ones are the Average and Range method and The ANOVA (Analysis of Variance) method. From [1]: “Formal analysis involves estimation of the components of variance...The measurement system variation is also referred to as total gauge R&R and is partitioned into a repeatability component (i.e. a component due to the gauge or measurement tool) and a reproducibility component (i.e. a component due to the operator, the users of the gauge or measurement tool)”
In the case of Gage R&R as applied to a measurement system, because the desired target is static, the ideal scenario results when σ2MS → 0, over time.
Figure 1: Measurement Process.
In the case of innovation, the desired target is not stationary. Let’s define the total innovation over one year as:
where for example iϵ[Q1, Q2, Q3, Q4], the quarters in a year. Interpreting individual data points in the innovation chart as contributions made by different teams or departments, and μ(i) as the moving average of I within a given quarter (Iμi=μ(i)), we see that the ideal case occurs when there is a positive (constant or increasing) pace of innovation, ΔIμi/ Δi > 0.
Figure 2: Innovation Chart displaying Ideal Pace of Innovation.
Or, in a more realistic case, it occurs when Iμi > ITarget, with ITarget probably still increasing at a lower pace, say year over year.
Figure 3: Innovation Chart displaying Realistic Pace of Innovation.
In either the ideal or the more realist case, and unlike a measurement system, innovation is almost fully dependent on human intervention, more specifically human creativity. And this will very likely continue to be the case, even when technologies like AI start to play a bigger role in technology, at which point combined human-AI contributions may simply take a larger dimension. Hence, removing or minimizing the human factor will inevitably also reduce the chances of achieving the very goal we were aiming for: to innovate. Note that the variance, σ2i, is still meaningful, as it can provide an insight on how the culture of innovation is spread throughout the different teams; a potential indicator of how systematic the innovation process really is within the organization.
So is our analogy valid? If anything, the analogy serves to point out key differences and specific variables that are dominant in a repeatable process, highly applicable to measurement systems or lean manufacturing, versus an evolving process such as development or even more so, research, for the sake of innovation. In the latter cases, unlike a measurement system, human intervention and intellectual contributions are extremely important and needed, in order to maintain an expected pace of innovation.
Can Structure and Innovation be both Part of the Design Process?
In good theory, Systematic Product Design should not go against Innovative Product Design. But maybe we can benefit from shaping the Design approach to inherently promote both. How can we do that? Let’s see.
The Design methodology should include innovation and critical thinking. But the Design process – as we know it – already promotes this. Follow a predefined structure and display repeatability in execution and administrative steps, minimizing human error and reducing unnecessary decision making, whenever possible. But, by default it already does, or should do. And carry a sense of urgency – time is of the essence! And yet again, it should already display this characteristic as well, even though in practice sometimes (many times!) it fails or comes short at it. Now, you may be thinking: none of this is really new. Agreed! So, is it possible that the sin is on the emphasis we put on each of these, or maybe the lack thereof? And also, could it be the fact that even if the perfect process existed, it may not be consistently and equally applied throughout the organization – be it large or small?
Generally speaking and following Mr. Breitfeld comments, large organizations display the following pattern:
High emphasis on structure
Mid-to-low emphasis on urgency, innovation
While, generally speaking, startups follow a different pattern:
High emphasis on innovation, urgency
Mid-to-low emphasis on structure
Innovation is not an easy task. Legacy automakers cannot just snap their fingers and replace ICE (Internal Combustion Engine) assembly lines, and shift the focus of thousands of workers and maybe hundreds of engineers, over night. That is one of the major challenges they face, in addition to striking the right structure-to-innovation balance. Are there organizations that have succeeded at applying the right dose of these parameters over time and have thereby achieved an optimal point? Very likely the answer is YES, not many since it’s a very difficult task – not impossible, but certainly difficult. Organizations that are in the lead of their respective technical fields and have managed to stay competitive, over a span of several years or decades and are still innovating. These organizations highly likely operate near or at an optimal point.
Reference:
Robin Henderson, G. Six Sigma, Quality Improvement with Minitab, Second Edition
Comments