Opinionated design

A few years ago, I wrote a blog post about curating taste. I was grappling with the inherent duality between a designer’s mandate to solve real user problems and their professional ambition to put out game-changing, world class designs requiring some level of taste. I wasn’t entirely sure if I had this so-called taste, and oftentimes felt like it was unnecessary flourish that impeded or distracted from the problems that needed to be solved. After attempting to describe my taste through many meandering paragraphs which took a wild detour into inspirational influences from other forms of media in that blog post, I ended it by hastily concluding that my taste was actually minimalism and that’s what was reflected in my work (which wasn’t even true, I would know). 

I’ve now been a designer in tech working full-time for a decade and felt like it’s a good time to re-examine this idea in the context of the sea change coming to tech with AI. I want to explore the idea of opinionated design, a concept that several design managers I’ve had over the years have tried to drill into me. It’s also closely tied to the idea of taste since opinions about how products should be used come with the prerequisite of knowing what people like about using products. I also had a related post about me not having strong opinions on things early in my design career, a perspective which I’ve now done a complete reverse flip on because I’ve convinced myself that in my career and in design as a whole, you need to have strong opinions in order to clearly and definitively communicate what your design stands for. Otherwise, you run the risk of obliterating any seed of elegance and thoughtfulness in the design by others who reshape it in some vision with their feedback — which can often be misguided — and turn it into a strange conglomeration of mixed opinions and design-by-committee style iterations where your original vision completely loses its hook. If you’re hosting a dinner with the promise of a showstopper main entrée, you don’t want multiple cooks in the kitchen adding their own flavor of spice and seasoning in a dish that was meant to showcase your vision for how it comes together as an exquisite meal. 

Opinionated design, as I understand it today, is the approach where you present a point of view for what the user experience of interacting with a product should be like. This may sound close to the definition of just UX Design, but there’s a critical difference. Opinionated design takes into account all the complexity of systems and processes that make the product work and then presents a simplified worldview of engaging with them. Let’s work with a simple example — Apple Notes. Despite its age, Notes has remained seemingly unchanged in its UI since the beginning. You open a new note, start typing, and it saves it for you. At their core, this is the most important value proposition of every note taking app, but tools like Evernote and Notion have ballooned in scope and have failed to deliver on this vision due to the added complexity that have shown up in their products over time. Notion continually shows options for AI prediction when I’m typing a backslash assuming I want to use a slash command. Evernote keeps asking me to organize and categorize my notes and whether I want it to sync to the cloud. Apple Notes never gets in the way of actually transferring your thoughts onto the tool.

Opinionated design tells us that rethinking how a tool like Notion or Evernote should work requires a fundamental rethinking of what we’re presenting to the user and whether it serves their core needs from the tool. When these companies get too big, internal processes severely hamper this mode of thinking. Entire organizations and teams within the company will have created complex dashboards that they display in the office proudly tracking adoption or retention metrics due to a single feature that prominently displayed in their product at the top of the homescreen. Multiple teams are doing the same thing due to a business mandate to increase some core metric that’s important to leadership. It becomes nearly impossible, then, to come at the product as a designer with an opinionated point of view of how it should be used when everything within the company is constantly moving in the opposite direction. 

For a long time, my modus operandi at these larch tech companies was to accept that reality. I knew what a good design solution to the problem was and I also knew everything that stood in the way of delivering that vision — resourcing constraints, deadlines & timelines, metric degradation, organizational misalignment on strategy, and so much more. After nearly burning out at one point trying to push against these forces, I had come to the conclusion that it’s not possible for a single IC designer to completely transform the culture of a company. I obviously wanted to do great design and would try my best to do so, but if the teams implementing my solutions are hearing from their leadership teams that speed and metrics are more important than quality and craft, then there’s only one real outcome that would lead to. 

Lately, I’ve been adopting a different mindset here. It’s now been one year at DoorDash for me, and it’s been a fun, exciting, challenging, ride so far. The company very much operated like a startup for the past ten or so years, shipping features scrappily and quickly to capture market share and rapidly acquire customers. They’re now in the phase where their product has a hundred different metrics that are apparently important to some individual Product Manager’s career prospects, and is stuck in a cycle of experimenting with small tweaks to existing features to see what gives them the biggest bang for the buck. Every PM’s dream is to ship the smallest possible change that gives the largest possible return in cost savings or user retention. And if every team is operating like this, there’s no team — aside from the design team — keeping track of the overall user experience of what it’s actually like to use the product on a daily basis. 

So my default mode of “doing the best that we can with the resources we have” is changing. I’m questioning Product Directors on whether or not we’re working on the right things. I’m asking Engineering Managers why basic frontend changes are so difficult to make. I’m presenting designs that showcase an extremely simplified and opinionated view of using the product that removes all the internal complexity we’ve built up over the years and shows a future that we can work towards and is actually achievable. For a long time, I assumed that bringing up all of this would make me seem green or clueless. I was worried that pushing too hard on my cross-functional peers would reflect poorly on my peer reviews where my peers may say that I didn’t understand the realities of our business and was “difficult” to work with. I thought that these “tech adults” who spoke a seemingly foreign language of acronyms and basis points knew what’s best and that my job was to simply manifest the way our system worked in the interface to ensure the best possible user experience that worked within our constraints. 

But the longer I work in tech, the more that illusion fades away. It becomes clearer that everyone is optimizing for their own gains due to the organizational incentives that are often incorrectly set up to propel the different disciplines in their own careers. It becomes extra clear that if I don’t keep the end-to-end user experience in mind and continue to question every new initiative or project that spins up, then the experience will continue to degrade. It’s unfortunate that this monumental task which has such big implications on the overall usability of a product is falling squarely on designers today, but that’s the world we’re currently in.

I think a part of this recent change in mindset has to do with me actively observing, as a consumer, how products are visibly getting worse to use if they don’t have an opinionated point of view on how they should be used. Apple’s Control Center update in iOS 18 is a good example. A quick-access view that once used to be simple, intuitive, and graceful has now turned into a fully customizable, confusing, and clunky UI to interact with. You can choose which controls you want and reposition them in different spots with different sizes. While this is great for users who have specific HomeKit controls or quick-actions they want to put in there, the experience for the majority of users is now one that is more cumbersome and unintuitive to use. There are multiple pages and actions that were previously accessible with a single tap are now buried in widget menus that require two or three taps. Sure, this can be customized with some work, but it involves a significant amount of tweaking and setup to get just right. I’d argue that the friction here is far too high for an average grandparent who already struggles with technology that has surpassed well beyond their mental models. 

Compare this to the very first iteration of Control Center that debuted with iOS7, where a simple bottom sheet contained frequently used actions in a specific layout that worked best for each function. It wasn’t customizable, you couldn’t choose what you wanted in there, and it did not have every feature that Apple had on iOS in there. But it was opinionated. A team at Apple decided that WiFi, sound controls, brightness, AirPlay, Screen Mirroring, Flashlight, and some other key features were the most important ones to have easy and fast access to. Everything else is secondary. When you don’t have this layer of opinionated decision-making from the creators of the product, the interface becomes a free-for-all, with users choosing what they think they want quick-access to but ultimately not using those things much or never bothering to redo their Control Center layout after the first time they set up some combination of actions. It completely misses the mark on the “Don’t make me think” guidance that good design aspires to, because only a very small subset of users will be continually optimizing their Control Center with their evolving needs.

A big component of opinionated design is simplicity. Imagine a three stage life cycle of a product, where it starts off as a simple solution to a problem and is minimal in its execution. The creators likely weren’t striving for simplicity, but because it’s likely a very focused product to address a specific user need, it’s simple by design. This is uninformed simplicity. Over time, the product scales to address more user needs and needs new features. New teams, initiatives, and operations are setup to account for all those things and the product grows in scope and scale. This is informed complexity. Accounting for all these features in an interface that stays usable, intuitive, and elegant isn’t easy, but it can be done. This is where most big tech companies are at right now. They have so much baggage from years of legacy features and value props that made sense at one point in the past that they carry it forward without questioning it since someone somewhere internally is tracking it as part of some metric. The final stage is an aggressive stripping of all the systems that make it difficult to provide a simple and straightforward user experience. This is informed simplicity. It’s a disciplined approach to be aware of everything that we’re told we need to communicate on the interface and yet choosing not to do so in favor of keeping the user experience opinionated.

Informed simplicity is often mistaken for uninformed simplicity, where if you ever try to show your design to some key stakeholder without any of the critical features that the company has worked so hard to build and track, you will likely get grilled by some executive about how your design does not account for everything the company offers. The truth is, of course, you’ve considered all those things already and have formed an opinionated point of view that they’re not relevant to the majority of users who are mainly looking for a simple and useful product that solves their problem. By simplifying and stripping the fluff and excess, you’re bringing more focus to the core offering that allows for building a stronger emotional resonance with your users. The freed-up mindspace for your users that was previously occupied with processing new feature offerings and figuring out where things live in a complex navigational IA can now be allocated to appreciating the elegance and quality of how a technology product has been carefully curated by a team of experts to give you a unique, hand-crafted solution to your problem. 

The biggest challenge, of course, is convincing people internally in your company that this is worth pursuing. That this is worth doing away with over a decade of measuring some metric that has been communicated to investors and shareholders as a key measurement of business growth and that we should focus instead of winning users by a change in form. It usually does not fly at most big tech companies, despite the best efforts of design leadership. The market forces of quarterly growth and increasing pressures to keep costs low blocks even the initial investment to explore the idea at all, with executives preferring instead to pour the resources into more optimization experiments that will show a marginal improvement in that holy key metric that measures the “health” of the business. 

The death of a product doesn’t happen suddenly. Your metrics can continue looking great as apathy and indifference slowly builds in users over time. You can add stop-gaps in the form of incentives and new features, and they work temporarily as users check out your new outfit for the day, but it quickly turns into the same old almost immediately. You can try putting out fires and outages and incidents that happened only to have a senior engineer tell you that the platform was “working as intended,” simply because an edge case in an interaction between a feature from three years ago and another from seven years ago wasn’t accounted for in a new feature. You can continue tracking your dashboards to have the retention graphs go upward and cost graphs go downward while celebrating with champagne, all while a new incumbent is building a competitor that is so singularly focused on solving the core problem that they launch a week later with a highly opinionated design that your users latch on to and give App Store verbatims hailing it as the simplest and most intuitive app in your category while your executive leadership scrambles to figure out how to chart a course forward. 

And now with AI making its way into product development, building products will get even cheaper. The baseline expectation for every tech company will be that they can build and deliver a product that solves a user need. The differentiator, much like it was when Apple originally brought this to widespread awareness, will be with design. Can you work to mold your product into a highly curated, specifically targeted, and very opinionated experience that clearly states its intent to users in an intuitive way? Do you really know what your core user needs are and can you deliver on them without getting distracted by a shiny new adoption metric? Can you build an emotional connection with your users through a product that reflects your users’ values, needs, and identity? This is what will be the differentiator in the next decade of my career in tech, and that’s what I intend to focus on. 

Taste is then the act of conscious filtering of product experiences you’ve had over the years into a refined sense of aesthetic expression that you then manifest into an opinionated design. You do this based on taking the inputs of the business and operation you’re given and determining, through careful synthesis, how it should translate into an experience that will make a product feel like it was made to communicate a specific point of view. The stronger your taste is, the stronger your conviction in this opinionated design will be. You’ll have a wealth of inspiration and reference to draw from to make the case of why this will work and why you believe it is the best possible direction for the organization to pursue. Nearly everybody will disagree with your proposed approach and you’ll likely hear an infinite number of excuses on why it’s impossible to achieve or why it will never see the light of day. This is why building your taste by observing what happens when products fail to build a connection with the users is important. This is why having conviction in your opinions is important. This is why, despite all the people internally that tell you that your design needs to account for a specific feature or scale to handle seven upcoming new cases, you need to have an opinionated point of view on what the best experience for your users actually is. At the end of the day, having bad taste is better than having no taste. Because bad taste will eventually give way to good taste and confirms what isn’t working, but having no taste is where you’re stuck in an endless limbo of stagnation from which you can’t move forward because you can’t get your product unstuck from the quagmire of decades-old baggage that it’s trying to make relevant into a world that has moved on.

In many ways, realizing this ten years into my product design career has renewed me with a sense of optimism about the future. Designers who have opinionated taste will become more valuable in the future. It’s still up to us to convince companies to hire us and that changing their products is something they need to do, but the good news is that AI can’t curate taste. AI is, in its current form, derivative. It knows what existed in the past, but it doesn’t know why. It knows what people liked and didn’t like, but it can’t create something original with that information. It repeats and recycles and regurgitates. The mark of great products in the future will be something that feels like it was carefully crafted and handmade in a crowded marketplace of bland, similar, and samey products that feel like there’s nothing unique about any one of them. And opinionated designers are the ones best positioned to deliver those great products.