There is an old joke about cadets in a tank warfare training program with three sessions, on mobility, communications and firepower.
The first instructor, an engine expert, concludes his session with the declaration, “a tank that can shoot and communicate, but not move, is useless.” The next instructor, a radio expert, concludes his session with a similar line, “a tank that can shoot and move, but not communicate, is useless.”
The last instructor, a gunnery expert, finishes his session with the line, “a tank that can move and communicate, but not shoot, is basically a 50-ton portable radio.”
The lesson I draw today from the joke (which I first heard 30 years ago) is this.
Complex problems contain three sub-problems: schlep, puzzle and package. For a tank, mobility represents the schlep sub-problem (building a vehicle for lugging a big gun around on rough terrain, using known technologies). Firepower represents the puzzle sub-problem (shooting accurately from a fast-moving, wobbling platform). Communication represents the packaging sub-problem (integrating the tank into a battle plan). It took decades to get the solution right, resulting in the modern main battle tank (MBT).
When you solve complex problems right, you are left with three corresponding intangible things of value: an asset, an insight and an aesthetic, which make the solutions both durable and generative (the solutions gradually and intelligently expand to occupy bigger problem spaces, realizing the potential of the original specific solution).
Understanding the interaction of these 3+3 input and output elements can make a big difference to how you attack complex problems. I am going to try and explain using the Iron Man movies.
Built to Last
My definition of a good solution to a complex problem is one that solves the immediate problem, is built to last and generates more potential than it uses.
The more I think about complex problems, the more I get convinced that the built to last part is critical. Almost all failures are caused by not aiming for durability. The three parts of the definition interact. Solves the problem is what enables the solution to survive in its infancy. Creates more potential than it uses is what allows it to keep going long term, as a grown-up idea capable of earning a living indefinitely. And attacking complex problems (even if small-scale) is also important. Simple problems can be solved in less powerful ways.
The three-way breakdown of complex problems is driven by durability logic.
- The schlep piece: the bit that takes the most dull/dangerous/dirty work, very little creativity, and a lot of energy, but leaves you with an essential strategic asset that will be useful in any solution to a broad class of problems.
- The puzzle piece: the part that requires an insight breakthrough; the part that’s going to take some luck, a lot of intelligence and genuine creativity. This piece of the solution is more vulnerable. Others can be lucky, insightful, intelligent and creative.
- The package piece: the bit that determines how the whole solution is put together to fit into the environment so it adds value gracefully. This is the least durable part of the solution, since the environment is not within your control and can change rapidly.
Good solutions have three parts, an asset that is the fruit of the schlep, an insight that is at the heart of how the puzzle is solved (which, in the best cases, will apply to a bigger class of similar puzzles that can be solved with sufficient imagination), and an aesthetic that determines how the solution is put together into a package.
In the case of our opening tank example, solving the schlep problem essentially creates a modern military-industrial complex for a country. Solving the puzzle piece creates a viral, generative solution with many applications (once you know how to aim a gun from a fast-moving, wobbly platform, you can attack many similar problems, ranging from other fire control problems, to image stabilization for cameras to vibration isolation for delicate telescopes and earthquake-proofing of buildings). The package piece creates a modern military capable of (for instance) operating by the aesthetic principles of Blitzkrieg.
Let’s see how these operate together in the Iron Man movies. I am going to try and extract lessons relevant to business competition in the technology sector, but you can extract similar lessons for other relevant domains (such as military capability, social work, governance or even private domains like writing or coding). Mahan’s Sea Power and its Influence Upon History is in many ways about the idea we’re talking about, but I’ll try to cover the gist more quickly.
Business Lessons from Iron Man
Ignoring the escapist and fantasy elements, the Iron Man movies are a great illustration of the fundamental idea here.
- The schlep was the montage-effort it took Tony Stark to build a fabrication shop in an Afghan cave (and later, a better one in his basement).
- The puzzle was the problem of miniaturizing the arc reactor to power the suit.
- The package was the idea of the suit itself.
Stark’s solutions to two complex problems (keeping shrapnel from his wounds from migrating to his heart and getting out of his kidnapped situation where he’s being forced to build weapons for terrorists) are based on a single creative insight that creates potential far beyond the immediate problem. This plays out over three movies.
Iron Man I
In this movie, once he escapes Afghanistan, Tony Stark competes against a villain who is able to imitate the package crudely without the aesthetic, using off-the-shelf assets and an unimaginative scaling (“bigger is better” substituting for a real aesthetic sense). But he lacks the solution to the puzzle. It’s almost a form of cargo-cult imitation that captures everything but the important part.
He manages to steal one of Stark’s miniature arc reactors, but under the stress of the climactic battle, his rip-off suit reveals its weaknesses against Stark’s more aesthetically coherent suit.
The movie resembles the Apple vs. Samsung story, though in a very caricatured form.
Iron Man II
In the second movie, Stark faces off against an alliance between a Russian mad scientist (Ivan Vanko) who cracks the arc-reactor miniaturization puzzle independently of Stark, and another competitor (Justin Hammer) who brings a dumb asset (money) and an unimaginative “more is better” package aesthetic to the party.
This is a more interesting contest, reminiscent of Apple versus Microsoft, but ultimately what fails Stark’s competitors is the weakness of the integration between individual puzzle solutions and the package. The result seems indiscriminate rather than eclectic and intricate. The differentiation (among the Army, Navy, Air Force and Marine variants of the Hammer-Vanko suit) looks like lazy and hasty variation driven by a land-grab on a market defined a priori, rather than uncovered via intelligent discovery.
The Russian is impatient for revenge against Stark for past wrongs done to his father, and pays for using what in startup lingo is called “other people’s money” (OPM) in order to act faster. Speed comes with a cost. Though he is able to dominate his partner and pursue his own revenge agenda, the alliance limits his creative control of the technology and forces him to scale in unimaginative ways. The marriage of convenience unravels under the stress of serious competition.
Many acquire-to-get-in-the-game scenarios seem to play out this way.
Iron Man III
In the third movie, Stark faces off against a totally new threat: competition from a package based on a completely different puzzle. Instead of an arc reactor, his mad-scientist adversary has figured out a bio-engineering technique to turn people into superheros, if they don’t explode first.
This is a business disruption scenario.
Unlike the villains of the previous movies, this competitor is truly a match for Stark: he has a puzzle solution based on an insight, the fruits of a schlep (an elaborate organization built up over as long a period as Stark’s) and an imaginative package based on an original aesthetic (a bin-Ladenesque theater of terrorism) rather than imitation of Stark’s aesthetic or the “bigger is better” and “more is better” aesthetics on display in the first two movies.
Here, Stark wins, after his assets are seemingly destroyed, by bringing out a hidden reserve of deep assets: an entire menagerie of drone suits controlled by him via an AI. All are based on the highly generative and continuing cascade of insights that started with the original suit. Unlike the dumb villain cohorts in the second movie, Stark’s army of suits is an intelligent and highly autonomous swarm with varied capabilities. It is able to put immense pressure on the competitor on all fronts.
This victory can be attributed to Stark having taken on and defeated many rivals over a long period, which has allowed him to build a depth of battle-tested operational capability newcomers cannot match, even if their capabilities have equal or greater potential.
Not many businesses today demonstrate the kind of capability that Stark models in the third movie. Amazon and Google come close, but seem incomplete (note: I am comparing Stark to companies rather than individuals, because he really portrays the capabilities of an entire organization rather than an individual).
Amazon has the package aesthetic down, but not a reliable source of generativity. A consequence, I suspect, of autocratic senior management creating a disempowered rank-and-file via hard-driving pressure, which also fuels a dangerously arrogant PR posture. As I have argued elsewhere previously, this is the same flaw that many military historians see as the cause of Napoleon’s downfall.
Google has the generativity, and a relatively empowered rank-and-file, but seems to lack a strong packaging aesthetic. So fascinating point solutions like Google Glass, driverless cars, maps a generation ahead of the competition, and Hangout aren’t coming together coherently to enable a high-potential assault on the so-called “so-co-mo” (social, collaboration, mobile) market. The symphony of search and advertising that came together in a young Google does not seem to be extending its aesthetic logic into the company’s recent wars.
Elon Musk, the model for Iron Man, is an interesting question mark at the moment. It is unclear whether his spectacular initial moves (Tesla, SpaceX and his solar technology) are going to create a vast and generative empire of organically evolving potential comparable to Google or Amazon.
Like Stark, Musk used a highly constrained “Afghan Cave” breakthrough (Paypal) to bootstrap into an unconstrained creative journey that’s just starting. We’ll see the importance of an Afghan Cave chapter in a moment.
The Interplay of Asset, Aesthetic and Insight
The Iron Man example illustrates how the three elements of good solutions interact.
Assets are the most durable part of a solution. Schleps have the advantage of creating unique advantages that do not depend on ideas or intelligence in a deep way. The only way for a competitor to match the advantage of a true asset is to undertake a schlep of equal magnitude, which requires a lot of energy and time.
Insights are less durable, since others can also stumble upon them. But insights are also the generative core of deep capability, the ability of resilient roots to sprout new shoots (rather literally imagined in Iron Man 3) even after a scorched-earth attack by an adversary (insert hydra/antifragile reference here if you like, though that’s farther than we need to go).
Companies differ in how imaginatively they are able to fill out the design space of possibilities created by a generative insight. Unimaginative companies fill out spaces in quick but dumb and homogeneous ways, with a great deal of repetition. They land-grab rather than truly occupy markets, and are easily destroyed by scorched-earth tactics.
By contrast imaginative companies generate highly varied ecosystems of interacting elements, all drawing on the same fertile and seminal insights. They civilize entire large continents in deeper ways, rather than just occupying them.
Packages are the least durable, since the environment is volatile. But they can be adapted and rebuilt surprisingly quickly if the aesthetic informing them remains alive in people’s minds, and generative insights and assets remain secure.
Bootstrapping vs. OPM
Let’s talk a little bit about Stark’s Afghan-cave experience. It sheds interesting light on the bootstrapping vs. OPM debate.
In business, there is a reason entrepreneurs prefer, as far as possible, to bootstrap rather than rely on other people’s money (OPM). Money (from investment, debt or non-core activities undertaken purely to generate cash) is a resource that speeds things up, but brings with it two contaminants:
- Other people’s motivations for solving different, unrelated problems (usually the problem of generating specific returns over specific time-scales)
- Tempting “assets” cannibalized from solutions to other problems that look “synergistic” but contain alien DNA.
What happens when you contaminate a good solution to a complex problem with other problems and solutions?
- Assets start to depreciate, due to loss of strong compounding effects from focus (solutions to various sub-problems not feeding off each other)
- Insights lose their potency, from being pointed in a less fertile direction, lowering generativity (alien DNA diminishing fertility or even rendering an insight-cascade sterile)
- Aesthetics decohere, leaking conceptual integrity and creating brittleness in packages (creating vulnerabilities outside-in, first in brands, then in products, creating openings for competitors to make inroads)
All three are forms of entropy that waste potential, which could have been realized more creatively in a bootstrapped story.
I am not being a bootstrapping vs. OPM ideologue, but I tend to view uncritical acceptance of OPM as responsible for killing the unused potential of creative solutions to important problems. In other words, OPM carelessly taken kills golden geese.
So the key to a successful OPM/bootstrap decision is to recognize why you are accepting contamination and how much you are really paying for it. Interest rates on debt, equity dilution from investment, and switching costs incurred from non-core activities are incomplete measures of the cost. The biggest cost is via intangible losses to assets, insights and aesthetics in the core.
Some good reasons to accept OPM anyway are:
- There is a strong window-of-opportunity constraint that forces you to operate at a certain minimum tempo
- There are forces in the picture that can destroy your limited capabilities with brute force while your roots are not deep enough
- There is a realistic expectation that there will be time later to “do it right” (pay off the “technical debt” of being forced to do it wrong initially)
Stark’s original Afghan-cave version of the suit actually displays sound handling of all three issues.
- He is working with cannibalized assets (parts of a missile system built by his company) and OPM (the terrorists money).
- He has limited time to manufacture an exit. He is facing an adversary willing and able to simply kill him if he interferes with their agenda.
- He has a realistic expectation that once he escapes, he’ll be able to “do it right” and pay off the technical debt incurred in the first suit (which is not a prototype; it is a deployed solution to the escape problem). When he escapes, he retains the generative core of insight (the arc-reactor in his chest, the core IP retained as a trade secret rather than patents).
It is interesting to note that even in his high-stress situation, Stark does not succumb to the temptation to simply put together a simpler escape plan (which a genius like him certainly could have figured out).
That would have been what is known as a “point solution” in engineering, or almost equivalently, “feature, not product” in the startup world. The sort of limited, prematurely optimized solution to the simplest framing of the immediate problem. One that a short-term thinker (or equivalently, somebody with a low threshold for aesthetic pain) might put together. Solves the problem, but does not create the potential.
The only case where this might be a good idea is when you have no interest in realizing larger amounts of potential, have occupied a naturally protected niche that others will likely not want to invade, and want to retire early.
But Stark is the unleash-big-potential type. So he puts together a creative solution with a lot of long-term potential, despite the short-term pressures.
By analogy, quarterly earnings pressures are sort of like terrorist guns trained on CEO heads by Wall Street. Relentless pressure from the dumbest kind of OPM available, driven by implacable short-term pressures and zero tolerance for bad short-term aesthetics (as in, temporarily non-pretty balance sheets).
Taking a company private is like escaping from terrorists in some ways.
Interestingly enough, the villain of the first movie copies the first suit rather than the “done right” one (which he can’t get at), and chooses to run with the expedient design towards bigger problems, without paying off the technical debt latent in it.
Iron Man vs. Platform-Product
I prefer the Iron Man framework (if you want an acronym, call it SAPIPA: schlep-asset-puzzle-insight-package-aesthetic) framework to the more usual way of thinking about solutions to complex problems: the platform-product model.
If you’re not familiar with it, the platform-centric model approaches complex problems as follows:
- You either solve a “first instance” problem in a design space (depth first) or develop a generically-capable “platform”, (breadth-first) depending on conditions. Both rely on an a priori mapping of an opportunity space. The latter requires a bigger capital stash.
- If you started with a “platform” you look for a “killer app” to realize its potential. This is a high-risk/high capital approach, but is sometimes justified.
- If you started with a “first instance” (better) you try to generalize the solution via a “bowling pin” strategy (applying it to a succession of similar problems until general principles emerge, and then trying to build out the “platform” under a collection of live point-solutions). This has a different risk profile — execution entropy and VUCA (volatility, uncertainty, complexity, ambiguity) getting in the way of your ever building a general solution. This is especially risky if your first few “bowling pins” were meant to be loss-leaders, with profitability dependent on your eventually building a lower-cost platform under them. When you fail in such a case, you bleed to death as you grow.
There’s more to it, but that’s the thumbnail sketch.
The problem with the platform-product approach is that it is all schlep (instead of a strategic “subset schlep”), with no puzzle or aesthetic components. So while it can work, the solution lacks durability and reserves of potential that can slowly be drawn upon to imaginatively fill out an entire design space, gradually transforming potential into deep operational capability.
What’s more, the platform-product mindset, by relying on an a priori mapping, loses the potential for discovery latent in a more generative approach. This manifests in a sharp distinction between “fit” and “scale” phases. Finding “product market fit” in lean startup lingo, is the intelligent, ready-fire-steer phase. Scaling is where you use OPM to grow as fast as you can to execute a big land grab. In fact, in today’s venture capital industry, investors openly agree that almost all the money, outside of a small “seed” bucket, is “growth capital” earmarked for use after a clear “product-market-fit” moment.
With the Iron Man approach, there is no a priori map. There is no sharp fit-t0-scale transition. Instead, by prioritizing survivability, you buy the ability to grow at whatever natural pace is needed, with low reliance on sharp transitions, and freedom from artificial pressures. So you have the following pattern:
- Raw-schlep assets gradually get locked down with increasing security
- Generative insights are unleashed and operate at an accelerating pace, creating a cascade of fractal SPP/AIA structures
- Packaging aesthetics are driven towards continuous refinement, you get increasing, eventually overwhelming dominance.
It’s what I’ve called exponential breakthrough in the past. It’s the difference between growing an intelligent entity that grows smarter as it scales, versus spreading a small initial amount of intelligence thinner and thinner.
But then, our technology leaders are mostly human. Not superheros. That is why our corporations are mostly linear, rather than superlinear.
But we seem to be getting tantalizingly close.