Rational Control in Systems Design

Why persuade when you can engineer?

Do you prefer to persuade people to do what you want or compel them to do what you want? The answer to the question will say a lot about you, namely how you view questions of right and wrong, free choice, and authority. In the West, we’re intently focused on the individual and therefore are taught to use methods of persuasion. Among cultural heroes are not just sports heroes and cultural icons, but also include great orators and debaters: people who can persuade others to their way of thinking.

Limits to persuasion and rational control

In fact, there are several bookshelves of books published on the topic. Dale Carnegie’s “How to Win Friends and Influence People”. Robert Cialdini’s books “Influence” and “Pre-suasion”, Roger Fisher and William Ury’s “Getting to Yes” are all titles that readily come to mind and their bank accounts surely reflect the popularity of the topic. The central belief underlying the topic is one of changing human behavior: Namely, that through our own behaviors and conduct we can change the thoughts and actions of others. This assumption is flattering. It treats people as rational agents who respond to reason, or at least flatters us, because it gives us greater agency to act and influence.

I’m going to skip over the part where I evaluate whether or not this approach is right or wrong. But I will mention several known areas where we know this approach to changing behavior fails.

Here are just a few:

• People speed through parking lots even when they know the risks, and even when they resent others for doing it when they are pedestrians

• People eat poorly even after they understand human physiology, the needs of the body, and basic nutrition.

• People leave lights on in rooms they’re not using, even when they pay for the electric bills themselves.

• People don’t use their seatbelts, even though it saves lives and is against the law not to.

So as we see, persuasion has its limitations. This realization begets the question, what if I’m not interested in changing somebody’s mind about something? What if I only care about changing their behavior? Do you have other options? Do all these other options boil down to strong-armed domination, tyranny? And maybe most profoundly, are these options more or less successful than attempts as persuasion?

The alternative which we are after is something called Rational control. I’ve seen Rational Control defined as the power to get somebody to do something without relying on persuading them that they should, which I think it’s a fair enough definition. I might add a principle of de minimis - that is doing this in the least obtrusive way possible. Without this principle, Rational Control could always end in IBCMs.

Rational control It is the art of shaping behavior through the environment itself, rather than through the minds of individuals. It is about configuring, not convincing. And once you understand it, you’ll see it everywhere and realize you should be using it far more than you probably are.

A bit of a speed bump…

The speed bump is one of the cleanest illustrations of this principle. In a private parking lot where cars routinely drive too fast, endangering pedestrians, cyclists, and other drivers, the property manager has two options.

Option one is to persuade drivers to drive slower. He may put up a sign. “Please slow down. Children present.” Perhaps add a flyer in every windshield explaining the dangers of speeding in parking lots, citing accident statistics, appealing to community responsibility. He may get the city to post a police officer in the parking lot to catch speeders and deter others.

Option two is to install a speed bump. Drivers will slow down because the alternative is a jarring, potentially vehicle-damaging thud. The desired behavior is achieved not because the driver was convinced but because the environment made it the path of least resistance.

This is Rational Control in its purest form and it corresponds to many of the principles written about so far in this Substack. Namely: Goodness of Solutions and Human Engineering. The speed bump doesn’t care whether the driver agrees with it. It doesn’t require the driver to read anything, feel anything, or believe anything. It doesn’t require ongoing enforcement and deterrence. It simply produces the outcome with greater consistency at considerably less expense, time, or upkeep.

The persuasion approach has its place, and we shouldn’t dismiss it entirely. Sometimes people genuinely need information or moral clarity to act differently. But the speed bump reveals that when you design the right constraint into the environment, you don’t need consensus. You just need good design.

Nudge theory and choice architecture

The academic world has been developing this idea rigorously, perhaps most influentially through the work of behavioral economist Richard Thaler and legal scholar Cass Sunstein. In their landmark book Nudge, they introduced the concept of “choice architecture” which is the idea that every environment in which people make decisions has a structure, and that structure powerfully influences what people choose, often far more than the content of the choices themselves. It is a beautiful systems thinking applied on a micro scale.

Their canonical example is the cafeteria. If you want students to eat more vegetables, you could give them a nutrition lecture. Or you could put the vegetables at eye level at the front of the line, and relegate the cookies to the back. No one’s freedom is restricted and nobody is shamed for their selection, but nevertheless, the students eat more vegetables; the behavior changed because the environment changed. Rational choice.

Sunstein and Thaler call these interventions “nudges” small, low-cost changes to choice environments that steer people toward better outcomes while preserving their freedom to choose otherwise. Both the cafeteria and the speed bump operate on the same underlying logic that behavioral change is best effected through environmental design, not individual persuasion.

Human engineering

Rational Control isn’t just a quirky insight from behavioral economics. It has deep roots in the broader tradition of Human Engineering, of which I’ve written extensively. (Seatbelts, HUDs). Human Engineering can be boiled down to the systematic design of tools, systems, and environments to optimize human performance, safety, and behavior.

Human Engineering emerged most forcefully in the mid-20th century, partly through military necessity. During World War II, it became clear that many aircraft accidents weren’t caused by pilot error in the traditional sense but instead were caused by poorly designed cockpits. Controls were placed in confusing configurations, instruments were hard to read under stress, and the mental and physical load on pilots was simply too great. The solution wasn’t to train better pilots. It was to design better cockpits.

This realization shifted the locus of responsibility from the individual to the system. The question stopped being “how do we get people to perform better?” and started being “how do we design environments in which people naturally perform well?” The first few chapters of Ralph Nader’s Unsafe at Any Speed is a great encapsulation of this principle as well.

The principles that emerged from Human Engineering have since permeated industrial design, software UX, urban planning, and healthcare. A hospital that redesigns its medication dispensing system to make the right drug easier to grab than the wrong one is practicing Rational Control. An app that places a confirmation screen before a user deletes important data is practicing Rational Control. A city that designs crosswalks and sidewalk placement so that pedestrians naturally take the safest path is practicing Rational Control.

Not by persuasion alone

To fully appreciate Rational Control, it’s worth being clear-eyed about why persuasion so often fails even when it works in the short term.

First, persuasion is effortful on both sides. The persuader has to craft the right message, deliver it at the right moment, to the right audience, in the right way. The persuaded person has to receive it, process it, and act on it not just once, but consistently over time. This is an enormous coordination problem.

Second, persuasion is fragile. People who agree with a message today may forget it tomorrow, or override it when they’re tired, stressed, or tempted. A driver who sincerely agrees that they shouldn’t speed still might speed when they’re running late. Beliefs don’t always translate into behavior, especially when behavior requires ongoing active effort.

Third, persuasion scales poorly. If you want to change the behavior of a thousand people, you need to convince each one of them or at least enough of them for the cultural milieu to change. Environmental design, by contrast, operates on everyone in the environment simultaneously and automatically. A speed bump doesn’t care how many cars pass over it and the effort to affect one car is the same as the effort to affect ten thousand and one cars.

Fourth, and perhaps most importantly, persuasion depends on agreement. If the people you’re trying to influence don’t share your values, your evidence base, or your reasoning style, persuasion may simply not work. Environmental design sidesteps this problem entirely.

None of this means persuasion is useless. For building culture, for instilling values, for handling genuinely complex ethical tradeoffs, dialogue and persuasion are irreplaceable. But for getting reliable behavioral outcomes in designed systems, Rational Control is dramatically more effective.

Applying rational control

So what does this mean in practice? Whether you’re designing a product, a workspace, a policy, or a system of any kind, the principle of Rational Control suggests a consistent set of questions to ask.

What behavior do I actually want? This sounds obvious, but it’s frequently skipped. Designers often default to providing information and trusting users to act on it, without first asking what specific, observable behavior they’re trying to produce. Being precise here is essential.

What is the current default? In almost every environment, there’s a path of least resistance — a default behavior that people tend toward when they’re not thinking carefully. What is it? Is it the behavior you want? If not, can you change the default to the desired behavior?

What makes the desired behavior harder than it needs to be? Friction is the enemy of behavior change. If the desired action requires extra steps, extra thought, or extra effort, people will often skip it — not because they disagree with it, but because they’re busy and tired and human. Identify the friction points and eliminate them.

What makes the undesired behavior easier than it should be? Conversely, adding appropriate friction to undesired behaviors can be just as effective. The speed bump adds friction to speeding.

Can the environment enforce the behavior directly? Sometimes, as with the speed bump, the environment can simply make the undesired behavior physically difficult or impossible. When this is available and appropriate, it’s often the most reliable option.

A Note on Ethics

The power of Rational Control raises a legitimate concern: isn’t this just manipulation? If you’re changing someone’s behavior without engaging their reasoning, aren’t you bypassing their autonomy?

This concern deserves a serious answer. The short version is: it depends entirely on who is doing the designing, in whose interest, and with what transparency.

Sunstein and Thaler’s framework explicitly defends what they call “libertarian paternalism” nudges that steer people toward outcomes that are good for them or for society, while preserving their freedom to choose otherwise. The cafeteria arrangement doesn’t force anyone to eat vegetables. The opt-out pension enrollment doesn’t prevent anyone from opting out. These designs respect freedom while acknowledging that the structure of choice always influences choice.

The ethical rule of thumb should be something like: design simple environments which support the people. The speed bump serves pedestrians and drivers alike. The well-designed cockpit serves the pilot. The thoughtful product flow serves the user. When Rational Control is deployed in service of the people whose behavior it shapes it is not manipulation. It is good design.

The manipulation version would be Rational Control deployed against users’ interests dark patterns in software that trick people into subscriptions, slot machine mechanics engineered to exploit addiction, confusing forms designed to suppress participation. Certainly these things all exist but the distinction is not about the tool; it’s about the intention and the alignment of interests. Should we avoid using Rational Control altogether because it could be used for ill? I don’t think so?

Rational control works

Here’s the bottom line for anyone who designs anything: you are already practicing Rational Control, whether you intend to or not. Every environment you create has a structure. That structure influences behavior. The only question is whether you’re influencing behavior deliberately and thoughtfully, or accidentally and poorly.

The developer who puts the “Save” button in an unintuitive location has made a design decision. The manager who arranges the office so that collaboration requires a cross-building walk has made a design decision. The policymaker who sets the default contribution rate on a retirement plan has made a design decision. In each case, someone’s behavior is being shaped. The question is whether the designer was thinking about it.

Rational Control asks you to think about it. To ask not just “what information should I convey?” but “what environment should I create?” Not just “how do I convince people to do the right thing?” but “how do I make the right thing the natural thing?”

This is a shift in mindset as much as a set of techniques. It asks designers to take responsibility for behavioral outcomes, not just informational outputs. It asks architects to think about movement, product managers to think about defaults, policy designers to think about friction. It asks all of us who shape the environments others inhabit to accept that we are always, inevitably, influencing behavior, and to do so with intention and care.

The speed bump simply works. That’s the whole point. The best design doesn’t need to convince anyone of anything. It just makes the right path the easy one.

That’s Rational Control. And it’s one of the most powerful tools in any designer’s arsenal.

You finished the article!

It was free!

Please: