On having an opinion
User testing serves as a valuable tool to ensure that the product or service meets the needs and expectations of the target audience, but it’s also not the be-all and end-all if it contradicts your internal beliefs and opinions. Sometimes you have to have a hunch and commit to it. This is particularly true now more than ever, where LLMs are rewriting a lot of the rules about software, and the best patterns or UIs are still being figured out.
Though, like many others, I strongly believe we’ve stopped having an opinion in design. Apple’s new design system, particularly Liquid Glass, is an OS having an opinion. Some may hate it. Others may love it. But whatever the case, it’s opinionated, just like iOS 7 was when it came out. M3 Expressive (terrible name) is also opinionated. Both, though, are uniquely opinionated.
New paradigms like LLM usage in products are influencing how we approach these opinions, pushing designers to make bolder decisions that align with emerging technological landscapes. But we’re also having to have these opinions at a pace unlike the past. This technology is moving so quickly that we’re designing with paradigms we already have just so that things make sense right now. But the final form of AI in product is not what we have now, not even close.
Take, for instance, Google’s recent updates to its search interface with AI Overviews, and now AI Mode. These decisions are highly opinionated and highly controversial. But I would bet hard that in user testing they performed great. People want answers quickly, and AI is novel. Great. But the reality has shown that people actually really dislike this and have seen a sharp decline in Google’s search result quality.
The problem is, user testing won’t tell you much in most cases here. The market, the approach, the UX, is all very unknown still. So we look to the past and resort to reusing old experiences. AI chatbots aren’t the answer, at least not all the time. And to quote one of my closest friends’ take from a year ago, “…the chatbot pattern comes with baggage.”
We’ve seen similar things ourselves at DuckDuckGo, where testing tells us the polar opposite of the market. Sure, one could argue it’s a bad set of tests that aren’t market representative. But I disagree. We work tirelessly as a unit to ensure our tests reflect our customer base, and to ensure that we have high confidence. Generally this proves out. But with AI, it always seems to be more polarising.
But beyond AI-alone, we seem to be pretty allergic these days about being opinionated. Having an opinion doesn’t mean you have to be rude or combative; it just means that you strongly believe in something. Design at its heart is highly opinionated. It’s a collection of opinions and it’s perfectly valid for everyone to have one, though it’s probable that the designer(s) on the team could have a more valid opinion (for the love of God can we please get over the “everyone is a designer” nonsense?), in that theirs is backed by expertise in the field. So why can’t we be opinionated and tasteful again?
Regardless, I think there’s a light at the end of the tunnel. People are getting more fatigued by the homogeny of modern digital design, and it takes a new generation of designers to push back on it. Design is not dead. And whether I dislike the rehash of my 90s internet youth by Gen Z and younger, I get it. They want something fresher, something challenging, something different.
They want to have an opinion.