Post 011 is part of LE’s foundational series on diffusion theory. Here’s the key point: Speed of adoption is not a reliable guide for an innovation’s importance. In fact, competitive advantage is much more likely to lie among slower ideas where innovators focus on several key factors to accelerate the rate of adoption.
It is difficult to accept an insight this counterintuitive. Thus, we need an illustration.
A life and death illustration*
Consider the following real world example of two important innovations, both in the field of medical surgery.
The first is the use of ether as a surgical anesthetic. With patients unconscious and insensitive to pain, doctors were able complete their work without patients screaming in agony. Reports of initial trials were first published in the Boston Medical and Surgical Journal (now The New England Journal of Medicine) in 1846. Within a half year, similar experimentations rapidly spread throughout the U.S. and Europe. Within seven years, ether-based anesthesia had become standard operating procedure in hospitals throughout the world. By any standard, this innovation spread very rapidly.
The second innovation is the use of antiseptic procedures to eliminate microorganisms from the operating room. During the 19th century, post-surgical infections were the biggest source of mortality for major operations, claiming up to half of all patients. In 1867, the British surgeon Joseph Lister published his theory of antisepsis and promising clinical results in a series of reports in the The Lancet.
Since Lister’s innovation curbed patient death — likely the most important objective in the medical field — one would think that antiseptic procedures spread as quickly as ether. Yet, that is not what happened. In contrast to ether, antiseptic procedures were relatively difficult to try out. Hands needed to be washed. Surgical instruments needed to be sterilized. Surgical sponges couldn’t be reused. Gowns and gloves had to be changed after each procedure. Thus, despite the incredibly high stakes for patients, it took a full generation before Lister’s recommendations took hold.
* These vignettes are drawn from Atul Gawande’s 2013 essay, “Slow Ideas,” The New Yorker (July 29, 2013).
The above vignettes illustrate that speed of adoption is not synonymous with what is socially and economically valuable. In fact, these are distinct constructs. We confuse the two because innovations that diffuse faster are generally easier to try out and understand. Thus, with a small investment, we can experience their value. Not surprisingly, we favor fast ideas.
Empirically grounded principles
This difference between speed of adoption and an innovation’s benefits can be seen in the prediction model discussed in Post 008 (summarized in figure to the right). Recall that for most innovations, between “49 and 87 percent” of the variance in rate of adoption is explained by just five factors: (1) relative advantage, (2) compatibility, (3) complexity, (4) trialability, and (5) observability. See Rogers, Diffusion of Innovations 221 (5th ed. 2003).
In Rogers’ model, rate of adoption is the attribute we want to influence (the “dependent” variable). The social and economic value of the innovation is the “relative advantage” factor that influences the rate of adoption (an “independent”, or predictor, variable). Note that relative advantage is but one factor that influences the rate of adoption. The innovator / early adopter / change agent who fails to understand this is flying blind and faces enormous risk of failure. See Post 007 (describing adopter types); Post 008 (discussing sources of adoption failure).
To illustrate the power of the first five factors of the model, let’s apply them to the case of ether and antisepsis. For ease of calculation, each factor is measured along a seven-point scale that runs -3 to +3. The following interpretative rules apply:
- Positive numbers (+1 to +3) speed up the adoption rate
- Negative numbers (-1 to -3) slow it down
- Mild effect = -1 or 1; moderately strong = -2 or 2; very strong = -3 or 3
- No effect on rate of adoption = 0
We then total the scores for all the variables. Innovations with a total score > 0 will tend to diffuse faster, in a relative sense, than innovations with a total score < 0.*
The table below summarizes the scores for ether and antisepsis. Note I divided relative advantage into two scales to reflect the differences in benefits flowing to doctors versus patients.
|Factor affecting adoption rate||Ether||Score||Antisepsis||Score|
|1a. Relative Advantage (for patients)||Patients are spared excruciating pain||3||Patients spared death caused by post-operative infections||3|
|1b. Relative Advantage (for doctors)||Doctors can methodically complete their work without patient screaming||3||Virtually all procedures altered; surgeons must operate in a mist of carbolic acid||-3|
|2. Compatibility||Completely new and foreign||-2||Completely new, foreign, and intrusive||-3|
|3. Complexity||Obtain right compound, construct inhaler, administer right dosage||-2||Extensive changes in all procedures||-3|
|4. Trialability||Assemble right equipment and conduct one or two operations||1||Change everything in operating room for weeks; calculate improved survival rate against historical data||-3|
|5. Observability||Benefits could be immediately observed and experienced by doctor and patient||3||Benefits only seen through data calculated weeks or months later||-3|
|Totals||Fast diffusion||+6||Very slow diffusion||-12|
In the above table, the factors with the largest numerical gaps are the factors that are exerting the largest impact on diffusion rates. Why did ether diffuse so much faster than antisepsis? With the benefit of an empirically grounded model, we can boil it down to three key factors.
- Relative advantage (for doctors): 6 points (+3 to -3)
- Observability: 6 points (+3 to -3)
- Trialability: 4 points (+1 to -3)
In narrative form, antisepsis was slow to diffuse because it required an enormous imposition on doctors, both to try out and to implement. Further, its benefits could only be observed through data generated weeks or months later. As a result of this adoption failure, hundred of thousands if not millions of people died of post-operative infections.
* Readers trying to apply this model are encouraged to use simple scales, at least initially. Although simple scales like the – 3 to +3 entail some subjectivity and do not capture quantum effects (e.g., impact of 1-2 may be less, or more, than the impact of 2-3, etc.), the results can be very useful, particularly when making simple comparisons.
Slow but Valuable Innovations
Within the legal industry, there are many valuable innovations that have a very high relative advantage but will tend to diffuse slowly. Using Rogers’ model, we can isolate the reasons: The innovations are: (a) culturally incompatible with lawyer norms and practices; (b) technically complex in areas outside lawyers’ core technical training; (c) difficult to pilot; and/or (d) difficult to observe, often only through data that must be calculated over time.
Examples of slow but valuable innovations include:
- Project management and process improvement, including integration of AI into workflows
- AFAs that reward efficiency and innovation
- Enterprise-level workflow management
- Talent management based on behavioral science
- Construction of client-facing self-help solutions
Despite being slow innovations, their relative advantage is steadily increasing over time because more and more legal clients cannot afford legal services that rely upon artisan methods of production. See Post 001 (lagging productivity is driving up price of legal services relative to other goods and services, negatively impacting both individual and organizational clients); Post 006 (discussing connection between historically low levels of law school graduates and 75% of cases in state courts with an unrepresented plaintiff or defendant).
In the legal profession, we tend toward fatalism when it comes to change management. Is it possible that this fatalism is the product of pervasive underinvestment in planning and implementation? One fast innovation we have all lived through is the diffusion of the smartphone. It diffused so quickly because Apple engineers took a decade to eliminate any feature that would detract from the end user experience. See Merchant, The One Device: The Secret History of the iPhone (2017) (discussing hardships endured by Apple engineers as part of the creation of the iPhone). The development costs, both financial and human, were enormous. Yet, by making these investments, Apple obtained staggering returns on the backend.
Speeding things up
The key point of Post 011 is that rate of adoption is an attribute than can be altered through planning and effort. I would encourage readers to reflect on slow but valuable innovations. Thanks to years of research, this is no longer an unstructured exercise. We increase rate of adoption by increasing cultural compatibility, reducing complexity, running low cost pilots, and making the benefits of the innovation more observable.
From far away, the success of this approach will look like luck or genius. From up close, however, it is just a high quality empirical framework + hard work.
The idea for this essay came from Jordan Furlong Law is a Buyer’s Market 187 & n. 166 (2017) (discussing Atul Gawande’s “Slow Ideas” in the context of lawyer change management).
Other Legal Evolution foundational posts:
- What is Legal Evolution? (001)
- What is the Rogers Diffusion Curve? (004)
- Units of Analysis and Adopter Types (007)
- Variables Determining the Rate of Adoption of Innovations (008)