Seven Plus or Minus Two

08 April 2021

At work, I inherited this tool to help people use a poorly designed product. While the long-term solution was to address core issues in the product, this supporting tool had low-hanging fruit, such as glaring usability holes. Rather than preach about these gaps, I leveraged the ever-relevant heuristic from writing: Show, don’t tell.

I asked several users to screen-share their use of the tool with my colleagues. One module had a list of troubleshooting scenarios—three columns of 20 options each. The team was shocked to discover that users didn’t use CTRL+F to find what they needed. One team member went on mute to laugh at the poor user. Thing is, that user wasn’t dumb. Despite using the damn tool each day, my eyes glazed over when looking at the wall of options. Too much for a human to process, especially when they’re already frustrated with the broken product. As a contradictory aside, I have a theory that Americans create very usable software because we design for the lowest common denominator of intellect: Americans!

This isn’t a story of innovation or novel insight, but it’s a prime example of Miller’s Law—the rule of seven plus or minus two. Miller’s Law states that humans can only hold limited things in their working memory. For instance, when presented with a random list of words or letters, a person’s recall ability deteriorates after five to nine characters. User experience design (UX) appropriated this into a principle for creating interfaces. Also called “chunking.”

We see this everywhere, from number strings—phone, credit card, and social security numbers—to menus for restaurants and mobile apps. I even find this in my personal life. At a given time, I’m tracking 7+/-2 habits, spending energy on 7+/-2 relationships, etc. Small chunks are easier to process.

Like everything, this is a contentious topic. Some believe Miller’s Law is an essential design principle. Others think it’s an unnecessary constraint. A useful heuristic on one side, a perennially broken rule with no scientific footing on the other. Given the “law” was coined to describe the outcome of a single experiment about memory, there isn’t science to support its use in design. People argue about it because it’s easy to understand. It’s why many UX conversations devolve into bike-shedding…

Anyway, I can appreciate the skepticism of 7+/-2. Dogma in nothing, excluding a dogmatic disdain for dogma! 7+/-2 isn’t a silver bullet. Frankly, it’s hardly even a bullet. It’s more of a foam dart. But it’s a pretty reliable foam dart that hits close to the target. Not every decision requires thoughtful analysis—especially menu design. Yes, the idea of 7+/-2 arose from unrelated research, but ideas are appropriated all the time, it’s in their nature. That’s why I don’t believe 7+/-2 is a law, but a decent guideline. A constraint for young designers, like training wheels.

To make progress, we need “mostly accurate” heuristics like this. I’ve struggled with the use of heuristics for a long time, especially in fiction writing. Because I’m a special snowflake, I start projects believing I need to invent everything from scratch. An entirely new plot structure! An entirely new prose style! An entirely new genre! It’s an exciting prospect, but it’s exhausting and not very productive. I have this weird belief that if I invent every component, I’ve somehow graduated from training wheels and achieved mastery. It’s a limiting belief, especially when I’ve hardly practiced the basics. Instead, how could I use existing conventions and constraints to free my focus for the few things I truly want to invent? I don’t need a new genre for my work in progress; I can use thriller conventions. I don’t need to define new formatting structures; I can use 2-3K-word chapters. Using guidelines isn’t a lack of creativity. If anything, guidelines like 7+/-2 enable creativity.

Every guideline, even every law, has a place to be broken. But why break them when we don’t need to? Often, the basics do just fine.