Good intentions are usually assumed. Generally, everyone understands that most people act on good intentions, and out of good faith. Many social behaviours are “unwritten contracts” rather than established in law because it’s understood that most people will abide by these contracts. But when good intentions aren’t sufficient, mechanisms are necessary.

As with many other productivity concepts like kanban boards or just-in-time manufacturing, the Andon cord was pioneered by Toyota in an attempt to reduce defects and improve customer satisfaction.

Traditionally, before the Andon cord was introduced, vehicle manufacturing companies used to have rigorous quality checks and strict systems in place to ensure products are up to a certain standard. However, these systems were deployed in a hierarchical manner where only managers and supervisors have the capacity to stop production. The problem in this situation is that, given a defect, accident or assembly failure is detected by a worker, having to go through the usual route of reporting to a manager would take too much time, and by the time the supervisor has to take the decision to stop assembly, there could already be several defective parts produced.

For this, the concept of “Andon cord” was introduced. It is a physical cord that runs throughout the entire factory floor, within reach of any employee regardless of rank or seniority. Employees are encouraged to look out for defects and problems and pull the cord, which in turn will significantly reduce the time required to stop an assembly line, therefore avoiding potential economic losses due to malfunctions or manufacturing defects.

This concept is what I call a “mechanism”. Mechanisms are differentiated from good intentions in that they are tools, available to anyone, that can be actioned whenever good intentions or promises don’t suffice.

This can also be applied to daily life: traffic lights or speed limits are “good intentions”. Everyone knows what they mean, and that you’re supposed to respect and abide by them. But when they are not followed, there exist mechanisms in the form of speed cameras, or police cars equipped with speed measuring radars, which will incur in fines or even jail time.

Every certification, every rubberstamp, every word, is just that: A promise, a good intention.

And when mechanisms don’t really exist, problems can spiral out of control. In most cases, the good intention of law is backed by a mechanism in the form of fines, arrests, lawsuits, etc. But a good example of good intentions without mechanisms to back them is national constitutions. It doesn’t really matter whether it’s the Russian, American, or Chinese constitution, but national constitutions simply establish a baseline for laws to be developed on, and aren’t really easily enforceable. In many cases, constitution article require further interpretation by bodies like a constitutional court.

Some examples of questionable constitutional articles:

• Amendment VIII to the Constitution of the United States: “Excessive bail shall not be required, nor excessive fines imposed, nor cruel and unusual punishments inflicted.”: Unclear what “excessive” or “cruel” may mean here.
• Article 58 of the Constitution of the Fifth Republic of France (4 October 1985): “The Constitutional Council shall ensure the proper conduct of the election of the President of the Republic. It shall examine complaints and shall proclaim the results of the vote.”: This is a good baseline, but it does not limit or contextualise the powers of the Constitutional Council in any particular way. It does not specify what “proper” means, and therefore it gives the Constitutional Council complete control over this matter.
• Section 19 of the constitution of the Kingdom of Spain (29 December 1978): “Spaniards have the right to freely choose their place of residence, and to freely move about within the national territory. Likewise, they have the right to freely enter and leave Spain subject to the conditions to be laid down by the law. This right may not be restricted for political or ideological reasons.”: This section is not useful because it defers the capacity to enter and leave Spain to the law. Therefore, if it’s illegal for a specific person, or for a specific period to enter or leave Spain, this Section would still be valid and unaffected, as it is the law that’s forbidding you, and not the Constitution.
• Article 98 of the Constitution of the Kingdom of Norway (17 May 1814): “All people are equal under the law. No human being must be subject to unfair or disproportionate differential treatment.”: This is a common article in many national constitutions. It’s not clear what “all people” means because some laws affect certain people, while other laws don’t affect them. For example: some laws can specifically target people in certain living arrangements, genders, or nationalities, and therefore are treated differently, contradicting the Constitution in principle. Additionally, “unfair” is not well defined, and is entirely up to interpretation.
• Article 35 of the Constitution of the People’s Republic of China: “Citizens of the People’s Republic of China shall enjoy freedom of speech, the press, assembly, association, procession and demonstration.”: Another common, recurring theme among national constitutions. The fact that Chinese citizens don’t enjoy a significant degree of freedom of speech or press is well documented and hardly a surprise. However, while cracking down on free speech in China is literally unconstitutional, it doesn’t seem particularly enforceable.

Thus, while national Constitutions are often regarded as the be all, end all of government counterbalances putting the interests of the people before the interests of the national government, the reality is more complex and depends on specific power structures and hierarchies related to the application of the Constitution and laws.

Designing mechanisms

Of course, mechanisms or “Andon cords” can be ineffective, or inflict unwanted collateral damage. Mechanisms must be carefully designed and engineered using best practices, in order to ensure their effectiveness while limiting the blast radius. Indeed, you might be wondering why not every manufacturing company has adopted the Andon cord principle. Or perhaps you have already realised this cord specifically can be used to sabotage production.

An example of bad mechanisms: Alert fatigue

Examples of badly implemented mechanisms, or mechanisms that, due to inaction, have been allowed to degrade over time exist in every discipline. One common case in Systems Engineering is Alert fatigue: This is the most pure example of Andon cords in action outside of a manufacturing plant. A practical example of this problem is as follows:

• An alarm is set up to watch a certain metric or indicator. A decision is made at some point that this indicator must remain below a certain threshold for a specific amount of time. Breaching this threshold will fire an alert, requiring human intervention.
• Over time, as the system grows, this alert fires every so often, with increasing frequency over a period of weeks or months.
• After months of receiving the same alert, the alert is subconsciously treated as noise, and its effectiveness is almost completely eliminated.

Thus, the alert becomes ineffective, and may even become a distraction that takes away engineering time from other, more pressing problems.

There are entire books with chapters dedicated to alert management. A good example is Google’s Site Reliability Engineering book, which is available online for free.

When engineering complex systems, the balance between proactive detection and operational flexibility is crucial. For instance, mechanisms designed to monitor or halt production based on predefined conditions can become bottlenecks themselves if they aren’t refined for both accuracy and impact. An aggressive, but ineffective mechanism might lead to frequent interruptions, while a lax mechanism may fail to prevent critical issues. Therefore, effective mechanisms are characterized by:

1. Precision and Relevance: Mechanisms should address specific, meaningful metrics that indicate real problems rather than triggering alerts for minor issues or outliers.
2. Threshold Tuning: Mechanisms must be tuned to strike a balance between catching genuine issues and avoiding false positives. This requires an iterative approach where thresholds and conditions are adjusted based on historical data and changing operational contexts.
3. Escalation Paths: Beyond simply detecting a problem, effective mechanisms should incorporate clear escalation paths. This might mean the system gradually escalates alerts or actions based on severity, or allows for automated responses under predefined conditions.

Mechanisms of Social Exclusion and Inclusion

We humans, also, without thinking, use and apply mechanisms in daily life. Societies aim to promote a certain type of behaviour and discourage other types. Depending on where you live, littering can lead to aggressiveness from others or social exclusion: this is apparent in Japan. Whereas in other societies, for example, there is a common understanding that, if you are able to steal from someone else by deceiving them, that means you have earned that money. Similarly, Tipping culture in the USA is widespread and part of daily life, however, in Japan, tipping is seen as offensive and insulting.

Through social networks (not the Internet kind, but groups of people connected in some way), cultures and societies arrive to a common, implicit understanding of what the ideal society looks like. In groups or communities, failure to conform to shared standards can lead to exclusion, ridicule, or distrust. Meanwhile, actions aligning with group values often lead to increased inclusion and trust, granting individuals access to social resources and opportunities they might otherwise lack.

Social media exemplifies how modern platforms amplify these mechanisms, with the power to reinforce or dismantle norms rapidly. A single misstep can lead to “cancel culture,” where an individual’s reputation is called into question.

Mechanisms in Law and Social Structures

Mechanisms within legal structures also illustrate how society balances freedom with order. Laws often reflect deeper societal values and function as both deterrents and signals of what is considered intolerable or unacceptable behavior. For instance, severe penalties for violent crimes signal society’s commitment to personal safety, while more lenient punishments for lesser crimes may indicate an understanding of human imperfection and the need for proportionality.

But laws are only as effective as the mechanisms supporting them. A rule or law that goes unenforced, or is ignored without consequence, risks becoming ineffective or even encouraging defiance. This is apparent in cases where outdated or unfeasible laws are still technically on the books but widely disregarded. Without consistent reinforcement—through mechanisms like policing, judicial oversight, and public sentiment—laws lose their regulatory function and become symbolic, a reminder of an ideal rather than an enforceable standard.

Mechanisms vs interests

Usually, mechanisms are set up to reinforce or complement good intentions. This is certainly the point of the post. Unfortunately, however, mechanisms can be turned against people, converting the situation into a zero-sum game.

There are many examples of this:

• Prices ending in .99: If a product is priced 39.99, the product might appear to be priced at 30, but in reality it’s essentially 40.
• Social media addiction: By exploiting the primitive human brain, social media companies can weaponise their platform to ensure you stay longer on their site, which in turn leads to higher revenue through advertisement exposure.
• Casinos and gambling: Through comprehensive psychological and sociological studies, casinos design their premises to be highly attractive and addictive, which drains gamblers of their hard-earned money.
• Payday loans: These services are advertised as a solution for people in urgent financial need, but the extremely high interest and fees trap individuals in a cycle of debt, making it difficult to escape once they’ve borrowed.

Generalising mechanisms

Not all mechanisms are mere engineering challenges and measures. As shown earlier, the law is in itself a series of mechanisms. Some people don’t commit crimes simply because they don’t believe murder is a good (read: effective, measured, reasonable, etc) solution to a problem. But there exist people who don’t commit crimes simply because they are afraid of going to jail. Some crimes have very few repercussions for the criminal. For example: while murder can have personal consequences for the criminal, shoplifting has no direct consequences unless you are caught: That is what laws are for.

A generalised approach to mechanism design is a recursive process and could look like this:

1. What is the problem to address?
2. How can this problem be resolved? (Propose several options)
3. Reason through, arguing pros and cons of each proposed solution.
4. Do any of the solutions cause collateral damage? Can this damage be contained? If it can’t be contained, a decision must be taken to tolerate the risk, or reject it, invalidating the entire solution.

These mechanisms then can be grouped up by field. For example, it is helpful to write down all mechanisms being enforced in a manufacturing plant. At some point, it is inevitable patterns will become apparent of interdependent mechanisms, or mechanisms that have to be applied differently depending on a set of circumstances. It is here where all those mechanisms are now considered problems, due to their incompatibility or interdependence, and the mechanism design process must be restarted from step 1.

Conclusion

Mechanisms are an effective tool, and must be carefully designed for a specific purpose. But more important than experience or practice, mechanisms and mechanism design require a very specific mindset and intent.

Good intentions not backed by mechanisms are doomed to fail, and when the stakes are sufficiently high, only good intentions won’t suffice.