I see some parallels between these three layers and Jabe Bloom’s Three Economies.
The gist of three economies is that there’s a layer of stuff–usually infrastructure or close to it–that gets consumed by use. It’s finite and loses value as there are more consumers. Think “disk space” or “network bandwidth” here. The goal then is to control usage and optimize for best value returned from the finite resources. Marginal usage has decreasing returns here. This sounds like Zunger’s X layer.
At the opposite end of the spectrum is very-highly-scalable stuff. User-facing applications, the product that people pay for. Depending on the product this doesn’t inherently gain or lose value as you increase the user population. (Leaving aside network effects via Metcalf’s Law for the moment.) The goal here is to differentiate and compete. Marginal usage has flat returns here. I think this is similar to the Z layer.
In the middle, Bloom posits there should be an “economy of scope” acting as a clutch between these two. It consumes finite resources but provides services that gain value from reuse. He uses the examples of a really good customer record or login function. Done right, marginal usage has declining cost for the “economy of scope” services and reduces cost in the differentiation layer. This sounds like the W layer.
I’ve just taken on a leadership role at my company and we’re looking to refactor software systems and services that are entering their 12th or 13th year of running. This kind of philosophy for looking at teams and systems is extremely beneficial to me. Thank you!
Thank you for the related link! That does indeed sound closely related. I wonder if there’s some shared ancestry to these ideas, or if they developed them independently — that would be a good sign for their accuracy!