Data vs Insight for UX Design

UX Insight Elements

Funny how things can pop into your head when you’re not thinking about them. I can’t remember why this occurred to me last week … but it was one of those thoughts I realized I should write down so I could use it later. So I tweeted it. Lots of people kindly “re-tweeted” the thought, which immediately made me self-conscious that it may not explain itself very well. So now I’m blogging about it. Because that’s what we kids do nowadays.

My tweet: User Experience Design is not data-driven, it’s insight-driven. Data is just raw material for insight.

I whipped up a little model to illustrate the larger point: insight comes from a synthesis between talent, expertise, and the fresh understanding we gain through research. It’s a set of ingredients that, when added to our brains and allowed to stew, often over a meal or after a few good nights’ sleep, can bring a designer to those moments of clarity where a direction finally makes sense.

I’ve seen a lot of talk lately about how we shouldn’t be letting data drive our design decisions — that we’re designers, so we should be designing based on best practices, ideas, expertise, and even “taste.” (I have issues with the word “taste” as many people use it, but I don’t have a problem with the idea of “expert intuition” which is I think more what a lot of my colleagues mean. In fact, that Ira Glass video that made the rounds a few weeks ago on many tweets/blogs puts a better spin on the word “taste” as one’s aspiration that may be, for now, beyond one’s actual abilities, without work and practice.)

As for the word “data” — I’m referring to empirical data as well as the recorded results of something less numbers-based, like contextual research. Data is an input to our understanding, but nothing more. Data cannot tell us, directly, how to design anything.

But it’s also ludicrous to ask a client or employer to spend their money based solely on your expertise or … “taste.” Famous interior or clothing designers or architects can perhaps get away with this — because their names carry inherent value, whether their designs are actually useful or not. So far, User Experience design practitioners don’t have this (dubious) luxury. I would argue that we shouldn’t, otherwise we’re not paying much attention to “user experience” to begin with.

Data is valuable, useful, and often essential. Data can be an excellent input for design insight. I’d wager that you should have as much background data as you can get your hands on, unless you have a compelling reason to exclude it. In addition, our clients tend to speak the language of data, so we need to be able to translate our approach into that language.

It’s just that data doesn’t do the job alone. We still need to do the work of interpretation, which requires challenging our presuppositions, blind spots and various biases.

The propensity for the human brain to completely screw stuff up with cognitive bias is, alone, reason enough to put our design ideas through a bit of rigor. Reading through the oft-linked list of cognitive biases on Wikipedia is hopefully enough to caution any of us against the hubris of our own expertise. We need to do the work of seeing the design problem anew, with fresh understanding, putting our assumptions on the table and making sure they’re still viable. To me, at least, that’s a central tenet behind the cultural history of “user experience” design approaches.

But analysis paralysis can also be a serious problem; and data is only as good as its interpretation. Eventually, actual design has to happen. Otherwise you end up with a disjointed palimpsest, a Frankenstein’s Monster of point-of-pain fixes and market-tested features.

We have to be able to do both: use data to inform the fullest possible understanding of the behavior and context of potential users, as well as bring our own experience and talent to the challenge. And that’s hard to do, in the midst of managing client expectations, creating deliverables, and endless meetings and readouts. But who said it was easy?

Author: Andrew Hinton

I use information to architect better places for humans. More at

9 thoughts on “Data vs Insight for UX Design”

  1. I really enjoy this separation between data-driven and insight-driven, but I wonder if it applies to many disciplines beyond UX. Good marketing is insight driven, not just data-driven. The same could be said of good sales or good HR.

    I think the separation from other disciplines comes from the insights we seek: they’re usually about the lives of people and how our solutions — as experiences — fit in with their lives.

    But the most interesting relationship you’ve outlined is the relationship of data, insight, and taste. Do you think of ‘taste’ as being insight developed without data?

  2. @brandon

    For the sake of the rough provisional model I include here, I might say that the overlap of “Talent” and “Experience” might be called “Expertise.” That is, one might be considered an expert because of a combination of having a knack for a particular kind of work, and enough experience to make that knack perform at a high level. (In Glass’ video, he makes that distinction — you might have a knack, and be able to recognize great work when you see it, but lack the experience to do the same thing or surpass it yet.)

    Add to that the element of understanding — doing the work necessary to get where your users are coming from, how they behave, what their expectations are, and how you might meet or exceed them. Then you gain new insights for how to design something. If you rely completely on your expertise, it seems to me you’re shutting yourself off from the possibility of much insight. Insight, to me, implies you’ve discovered or learned something new (or validated or improved upon an existing idea). Taking one’s expertise into the unknown and allowing for discovery and new understanding opens up the possibility for more insight.
    I think I avoided putting “data” or “taste” in the model specifically because they’re red herrings, of a sort. Data is just a set of inert facts. And “taste” is a vague, subjective term that folks bandy about the way alchemists used the word “aether” — a stand-in for something they hadn’t really nailed down yet. It’s fraught with cultural baggage and attitude, and in my experience only complicates any design conversation. (If one person thinks the design should be X because of ‘taste’ it implies the other person is lacking in ‘taste’ which is a pretty big insult, and probably wrong.)
    So, “taste” isn’t really definable. Expertise, while somewhat subjective, is a more useful term. And if a designer holds a particular position on a given design based on the designer’s expertise, it’s up to that designer to make the case for it, based on prior experience, logical argument, or charismatic flourish — whatever works 😉

  3. Thanks for the thoughts, Andrew. I like the way you characterize “expertise.” I’ve tried my own synthesis of the data/insight question, from an anthropologist’s perspective.

  4. Hey Andrew, I thought you’d like this quote:

    “I make all my decisions on intuition. I throw a spear into the darkness. That is intuition. Then I must send an army into the darkness to find the spear. That is intellect.” – Ingmar Bergman

    I like how he heroically articulates a common and perfectly normal (and in the right hands quite successful) creative process of coming up with the idea first, possibly even implementing it, and conjuring up the rationale after the fact.

  5. @chris That is quite a quotation. It just drips with a certain sort of heroism like you say. But also as you say, it’s successful “in the right hands.” One of my favorite movies is Ed Wood — there’s a guy who had all the vision & heroic belief in himself, but was really a mess when it came to the end result. (Though, hell, he has a huge cult following, so maybe he was successful too?)

  6. As a person who came from a culture of data-heavy and data-driven (usually poor) design, I couldn’t agree more. Data is not the stuff of design. It’s an emergent property. Data *informs* design…it can never drive it.

    Unfortunately I’ve seen many otherwise talented usability analysts say “we need more data.” Increasingly I say “feh.”

  7. @Andrew Great to bring up Ed Wood. For every “Genius Design” practitioner, Ed Wood makes for a great counter example. And the movie captures the blind optimism that usually greases the wheels of a design disaster. His cult fame is hardly, however, “success”. 🙂

  8. I’m tossing around the idea in my head that insight is generative, while data is restrictive, when it comes to design. That is to say, data can help you narrow down from a set of possibilities, but it cannot by itself create anything new. So the process is one of intuition or insight coming up with a range of possibilities, and then using data to cut away until you settle on one.

    I can’t decide if this is actually true in real life, though.

    What do you think? Is there a counterexample?

    1. @bri That’s a really great distinction — in general I think it holds up. That is, if we think of “data” in the typical sense of empirical stuff — numbers, charts, etc. generated by quants and business culture.
      Really, though, the word “data” is so abstract — it depends so much on context. For example, ethnographic observation is certainly data-gathering, and can be incredibly generative… it can be the raw material that sparks a lot of great ideas we hadn’t thought of before. But even empirical data can sometimes force us to see something in a new light, which can catalyze a new idea.
      So, I dunno … maybe it’s not the concrete “object” of data — the numbers themselves — that are restrictive, but the culture or approach? And maybe we just associate data with one culture over another because those are the folks who tend to use it the most, or by default … even though the insight/creative types really should embrace empiricism too, just in a different way?

Comments are closed.