Guiding Principles for Product & Design

The Role of Guiding Principles

The role of guiding principles is to help teams make decisions. You cannot come up with guiding principles unless you have thoroughly understood the problem and have clear rationale on why you can’t or WON’T do x, y, z. You don’t know what you are unless you can define what you’re not. It’s the team’s distilled point of view around the human value and the why behind the soul of the product.

From my experience, I can’t craft a principle until I’ve done my investigations, explorations and qualitative research; I’ve attempted to falsify my position, tease out divergent approaches to understand users’ motivation and trade offs. It’s not something I arbitrarily pull out of my ass. If you’ve conducted good research, the principles should feel obvious. The dots will appear for you to connect.

A principle must be clear, actionable, and align with the mission. “User friendly” or “Intuitive” are useless principles. They are not specific enough and won’t help you make decisions.

Some Examples

Ten years ago, when I was a designer at GoPro, we explored a machine-assisted storytelling application where users select moments in their video to instantly create an emotional video synched to music. They could do this within 5 minutes, instead of hours on something like Final Cut Pro. These are some of our principles:

• Lightweight: This was an important principle, because it makes the team mindful of complexity. There’s also a correlation between content creation time to the scale of content created.

• Moments over Manipulating Time: This has a system wide repercussion on how we thought of engaging with emotional moments in a video. It’s a different mental model than creating clips. It’s about speed over effort, emotionality over technicalities.

Here’s another example. When I worked on Search, we’d get arbitrary requests from certain stakeholders to make certain assets rank higher. So our principle was:

• Search is agnostic: We won’t artificially inject weights. It should be obvious, but you’d be surprised how people justify poor behavior. Principles can reflect your values, what lines you won’t cross. I’ve often used principles to create boundaries to manage poor behavior.

I’m always looking for principles behind a position, especially in discovery of emergent tech products, because when we invent new ways to interact with tech, we set the expectation for consumer behavior at scale. Every iteration we explore, I test and refine the governing principles. Good principles create coherency, focus and provide teams with a compass.

Now I’m at Zillow, without divulging what I’m working on due to NDA, the guiding principle I define informs the focus for UX automation, the physics of the interaction in immersive experiences.

Principles to Mitigate Harm

The principles above provide a POV about the product itself, but there can also be principles around mitigating harm.

Years ago, I worked at a startup on IoT for smart homes, and we had a principle called, “Local First.” This principle prioritizes storing data locally when possible, protecting privacy and leveraging cloud only out of necessity or for enhanced functionalities.

When we think of massive platforms like Amazon or even new tech like GenAI, it’s easy to see where there’s a lack of principles to mitigate harm. If there is one unspoken principle that we see in tech over the past decade, that has to be, “Market dominance over society well-being.” Growth at all cost.

I started thinking about what types of principles I’d like to see in AI products. Here’s a few.

• Accuracy through accountability and transparency: I’ve been working in ML and AI products for almost a decade now. The longer I’ve been at this, the more I believe in the importance of humans-in-the-loop. Just the idea that we have to define when and where human input is needed to augment AI output creates a kind of reflection muscle at a system level to address the long tail of variability in suboptimal results.

• Fairness by default: This is about proactively address sources of bias in data and models, as well as questioning inequities in outcomes across diverse groups.

Guiding principles are a way to crystallize your intentions, conviction, and what lines you won’t cross. Principles can change. When teams debate if a feature fits a principle, they have to address why they want to break that principle. Sometimes breaking a principle force them to ask, “Is this who we are?”

I’ve always included guiding principles as part of my strategy and the story of the product. Applied research, investigation, divergent prototypes and proofs yield guiding principles. The reason why they are powerful is because over time, the team will know it by heart. It’s like a trigger for reflection. It’s not about platitudes. Principles can be the DNA that inform every layer of your product.