We tend to think about things in terms of evolution and revolution, of steady continuous changes or else monumental discontinuity. We reach for examples that allow us to identify these watershed moments, and we are often blindsided by changes that look revolutionary after the fact, but which look pretty gradual while they are taking place.
At issue is that the analytic tools used to understand institutional change relies on a model of stability punctuated by disruption. This is too conservative, since it underestimates gradual transformational change and leaves us arguing about ‘disruption’ vs. ‘continuation,’ as two opposite poles. To correct for this either/or problem, we need to distinguish processes of change from outcomes of change (e.g., Streeck and Thelen Beyond Continuity). Consider this 2X2:
The interesting box is that top-right one, ‘gradual transformation.’ These kinds of transformative changes are game-changing, but they happen as a result of an accumulation of gradual and incremental changes. They are often endogenous, and they come as a result of the gaps between formal institutions and their actual implementation and enforcement.
Gradual Transformation
There are 5 broad categories of gradual institutional change.
- Displacement. New models for action emerge and diffuse which call into question existing, previously taken-for-granted organizational forms and practices. For example: electronic trading in finance, took the assumptions of markets to extremes, and then ended up displacing face-to-face marketplaces.
- Layering. New forms are enacted atop existing forms, and then the process of differential growth creates systematic changes. Compromise between old and new slowly turns into defeat of the old. For example: charter schools become a compromise towards school reform, and then slowly siphon resources from public education.
- Drift. Changes to the economic/political/cultural environment change, leaving a seemingly-stable institution to lose salience. Institutions retain their formal integrity even as they increasingly become inattentive to social reality. For example: Without changes (which, to be fair, they are seeking to make) I would gently suggest that Intel may be in this position. Or at least it’s historical attention to cadences and user practices may put it there.
- Conversion. Redeployment of old institutions to new purposes. Or rather, new purposes are attached to old functions. For example: interest group politics (by marginal actors, like women who were unable to be part of political parties because of disenfranchisement) at first cross-cut, but then reorganized party politics in the US
- Exhaustion. Gradual breakdown and withering away of an institution over time. The normal working of an institution undermines itself. For example: (a cynic would suggest) the world power model of democracy, combined with the workings of a modern security state. We need security state to make us safe so that we can be a democracy, which requires security state, which undermines democracy, etc.
For large corporations interested in innovation, this 2x2 offers a couple of lessons. The first is negative – drift and exhaustion are both things that tend to happen ‘to’ companies. Drift is when you are good at something that no longer matters, as the underlying conditions in your external environment change. At Intel, this was the story around mobile. Competing along vectors of ‘computational power’ when the world has turned to ‘low-energy’ devices means that increasingly your position in the ecosystem gets eroded. We are seeing some dangerous signs of similar activities around GPU cycles, as machine learning turns to non-CPU cycles to run efficiently. We have highly competitive products in these spaces, but they are now competing along different vectors.
Exhaustion, by contrast, happens in the form of something like an increasingly complex and onerous PLC process. As a highly integrated manufacturer, our central values are around integration in the compute ecosystem, and so our core products become more complex over time, and more difficult to integrate. Or, at least left unchecked, more expensive. This is the horizontalization story – and we have done it to other competitors, even as it happens to us.
Both of these are not inevitable, but they tend to be more about losing positions in ecosystems.
More useful for understanding how to drive transformational innovations into the company, are displacement and layering. The way I would characterize displacement is closer to Moneyball in baseball – a new set of analytics (on-base percentage, eg) replaced long-standing ones. And their diffusion into the game actually changes quite a bit, from draft to player management. It’s not that the more conventional measures go away (people still, I’m sure, talk about someone having a good ‘baseball body’ or a 5-tool whatever), it’s that the new measures have more or less replaced old ones. The clubs that have done it more readily have been more successful.
So what would one do at a larger corporation? Start small, and measure the hell out of everything. The process, the financial models, the business models, you just have to be sure to capture as many of the things Intel thinks it already knows. Because, my argument is, the existing measures (market forecasts, consulting firm strategies) are meant to capture a marketplace ecosystem that no longer obtains for us.
Second is layering. For our purposes, this is to create a new offering (product or service, or a new business model), in a niche market that is underserved by existing offerings – and which niche has the properties that it is ‘like’ other niches, or that is likely to grow over time. This is Venmo’s strategy with millennials in digital payments; it is Snapchat’s strategy with teenagers. Usage-based insurance starts with teenagers who otherwise can’t get insurance (because of their age group), and this increases chances they’ll put a tracking dongle in their cars – and then UBI spreads from there.
Absent a better map for driving innovations into large organizations, this is the high-level map I would use. The next level of detail is about what kinds of specific projects, and where in the organization it would be useful to drive those changes.