"Quality" is a tricky concept, so I wanted to write down what it is here in a way that can be referenced. This should save some time repeating myself in future posts. We can but dream.
In the most simple form, "quality" is the measure of how well something does at being that [thing]1. Scientifically minded people, including many testers, tend to view this through a fairly binary lens; the scientific method is that we establish a hypothesis, a process for capturing insight into the [thing], and compare the insight with the hypothesis. Testing takes this broad approach, but the hypothesis is supplemented with an expectation. The components of a test are to determine a correct outcome, identify a way to observe whether that outcome has been met, and then to compare the actual outcome with the correct one. If they are the same, congratulations, the test has passed. If not, further testing is usually required.
While this is good science and good testing, it's not necessarily capturing the quality of a [thing]. We will know from personal experience that a [thing] can be functionally correct, and yet considered poor quality. Perhaps we've driven a car that's underwhelming, or eaten food that was functional but otherwise unremarkable. This is complicated further by a colloquial understanding of "quality" being a synonym for "good", or even of higher than average value. To be clear, that is not accurate, but perhaps a story for another post...
In technology, the additional elements that help define the quality of a product are often referred to as non-functional attributes. Where any product can be split into a series of features that define what exactly the product does, the non-functional elements of a product represent how the product does it. A common example is the performance of a product. Sure, this software application allows people to access a secure user interface. However, the developers need to consider how long it takes a user to load the page, and how responsive the page is when there are multiple users loading the page simultaneously. An application that is fast and can handle multiple people has a greater performance than one that is slow and crashes with only a few users. For our scientifically minded friends, this is the metric through which we can finally determine quality.
However, this is still not the proper way in which we determine quality. It might seem obvious, that the non-functional performance of a product clearly provides a degree of separation from a functionally similar, but worse performing alternative. Yet there is more to consider. Firstly, there are tradeoffs. Many non-functional attributes are incompatible or can compromise each other. A food vendor may have to balance speed of service with freshness. Someone designing a digital interface needs to balance design features with accessibility concerns. The security measures of a bank directly impact the user experience. Secondly, nothing is free. Whatever the [thing] in question is, there's an inherent cost to non-functional requirements. It might be quite literal, paying for compute in an IT system. It might be reputational. As such, while we can determine what the thing should do, and how the thing should do it, we must also account for why it needs to be done that way.
Perhaps the easiest examples to use are commercial technology. How can you determine the "quality" of a car? We can talk about functionality, or the non-functionality of a car. What features does it have? How does it perform? Judging a car by these criteria alone can give you an indication of quality for sure, although the trade-offs still exist. A fast car is unlikely to be an efficient one. In fact, there are several publications online and in print that dedicate a portion of their output annually to judging and ranking the best cars. However, what's interesting is that the cars that appear on these lists aren't always the best selling cars, and the best selling cars are rarely the ones that are universally agreed to be of great quality.
Some of this is necessity - people are often looking for utility over pure quality, and are constrained by finances. However, even great selling cars that are generally considered good quality can be very different. In the UK, the VW Golf is frequently in the list of best selling cars and ranks well among car enthusiast publications. Similarly, the Tesla Model 3 was, for a while, the best selling car in the UK and was a Car of the Year in those same publications. If these two cars are both popular, both highly ranked in publications, both do the same basic job, yet both have very different functional and non-functional attributes, how can quality even begin to be assessed?
The answer is often unsatisfying and subjective. Unlike the colloquial idea of quality meaning "good", the truth is that quality sits at the intersection of desirability, functional correctness and accessibility. It's not mathematically predictable either; any one of these criteria can overcompensate for the others. Historically we've seen superior technologies struggle against functionally inferior alternatives. We see people motivated to visit difficult to reach countries. We see cheap Chinese cars hit bestseller lists. While these [things] may not meet the traditional definitions of quality, the key thing is that they achieve the goals of the participants.
If you are designing for quality then, you need to understand what behaviour you would like to see, and how to achieve that. If you are building a commercial product, you need to understand the desire of your consumers. If you are designing a process or concept, you need to understand the desire of your participants. You need to understand what the [thing] needs to do, and how to do it. You need to make it accessible enough, but also understand that there's sometimes an inverse relationship between accessibility and desirability2. You also need to accept that there is no such thing as engineering the "best" [thing], nor is there a clear roadmap to achieving this. Sometimes it might require continuous improvement or adjustment.
I may come back and add to this publication. But this is the crux of it all, and also why so many people get quality wrong.