You know what I love about software? For any problem, there’s a solution that will automate doing a half-ass job on the easy parts of the problem, and a bubble based on claims that it will solve the whole problem and a couple of ancillary problems along the way. There’s the sordid history of AI of course, but also see business process automation, SIEM, and CMDBs. Systems that are important enough to attract automation investment and excitement involve people, and people are the hard part.
Arguably a hype bubble is harmless enough. Surely no one really believes in the flying car, right, and so misguided investors dropping some dollars or young engineers dropping a few years into that work are just learning about caveat emptor. When it happens on venture capital scale, or cultural phenomenon scale, that’s the same thing but bigger. Bubbles are just a fact of capitalism. However, there’s something interesting in that issue of scale: there are hundreds of thousands of people who participate in driving software bubbles, and they’re not all steely-eyed capitalists knowingly surfing the number-go-up waves. People make software to achieve these goals, they market and sell it, and they buy and use this software in their organizations.
Why does this happen? The easy answer is the Joe Isuzu answer, which is an extrapolation from the transparent tactics of low-end sales experiences in our daily lives. In this model the software product is known to be nothing special, but the salesperson lies and the consumer is fooled. If we assume this is the case, then software bubbles are also built on a small number of grifters tricking everyone else into compliance. Another, more charitable answer is more religious: instead of grifters, the core drivers of the bubble are people who want to believe and want you to believe with them. These two models both edge into conspiracy theory, sharing the conception of a central core of people who know what’s going on and are driving bubbly growth for their own motives. One can argue if those motives are nefarious or wholesome, one can wonder about the relationship of the bubble to grander forces, but at the end of the day it all feels more fictional and less chaotic than what I think is a more emergent reality.
A third model is more intriguing to me: those who work the software bubble had a plan to really do the amazing thing, but it hasn’t worked, and now what. Upton Sinclair wrote “it’s difficult to get a man to understand something, when his salary depends on his not understanding it.”
As a developer building a system, you’d have to ask yourself: is an observed failure to meet the customer’s goals because of a fundamental impedance mismatch between those goals and your product’s capabilities, or is it because the user isn’t providing the data that the algorithm would need to do a better job? Before you answer, consider that if it’s a fundamentally insoluble problem like human nature, you’ll need to go get another job and start your climb towards promotion over again. Alternatively, you could suggest greater investment in technical writing and professional services to help the customer hold it correctly, which means you don’t need to make new friends or alter your commute.
From another angle: as a sales or marketing person positioning that developer’s system, how do you differentiate it? In other words, what is the pitch that you will use to explain what your product is and why people should want it?
- It does a thing that customers need and haven’t been able to do before.
- It is better than the current answer for that thing, because features.
- It is better because of a principled stand and moral basis.
- It is better because it’s cheaper.
Those pitches are in order of goodness: a new ability like instant communication with distant family is better than a hair spray which doesn’t enlarge the ozone hole, which is better than “a sponge that lasts longer costs less.” So of course, our marketing friend is going to start with “it does a new thing” and maybe fall back to “it’s better than older tech” — like the religious bubble driver hypothesized above, they probably took this job because they believed the software could do the thing. It would take some disturbing data to shake that approach and drive one of the others. Still, sometimes that data does come along and force a reckoning.
“Better because cheaper” is really well-trodden ground, though sometimes incorrect (also note the existence of premium brands at all). The principled stand thing is a bit more foreign in enterprise software circles though: principles and morals are usually reserved for consumer facing brand washes. Exhibits A, B, C, D, E, and F are from Major League Baseball, but we can see similar on the home pages and blogs of many well-known B2B brands. What does it even mean to be a principled enterprise software vendor? The cynical moat-digging efforts of OpenAI and friends sure don’t meet the bar. Most open source vendors haven’t acted according to principles either, doing stuff like closing source and forking source as soon as possible. While they’ve failed, early and hard, I think it’s useful to look at the non-profit Mozilla.org: at least as an example of omission. There’s a manifesto, which completely fails as a mission statement (long, unclear, unmeasurable, non-guiding)… so I can’t really say what principle they’re upholding? By comparison, let’s look at Alameda County Food Bank: “food is a basic human right. Help us end hunger in the Bay Area.” It’s a little hard to argue that Mozilla’s actions don’t align with whatever that manifesto is saying, since it’s meaningless pablum… but if we go back in time to its origin in 2007, it was not much more focused. How owning Pocket or Anonym serves any of those goals is left as an exercise for the reader. So, marketing by principles: as something like a food bank or Habitat for Humanity, you stand for a position and you ask for whatever engagement in order to serve that position, which is not simply “make more money.” For our marketing friend, this is better than selling by cheapness because the consumer isn’t buying the least expensive option, they’re buying the best or possibly only option. In other words, it’s similar to doing a thing that wasn’t possible before or doing that thing better.
So, while not fully discounting the tinfoil hat paths, I think software bubbles mostly happen in this way: an attractive idea is attractive to software professionals and venture capitalists alike. “Sitting in traffic stinks, what if we could fly above it?” They start companies, the bureaucratic principle takes over, and they pivot development and marketing approaches through various challenges. The software changes in delivery (mainly adding more human time investment to the deliverable), and the marketing slides from “changing the world” to “but it’s cheaper”. Eventually the trough of disillusionment arrives and puts most of these companies out of their misery, while any actual value to the bubble lives on in the survivors.
Monkeynoodle.Org 
