Data informed design, not data-driven design

5 minutes read

You might have heard that data is now a big thing in business, especially BIG data (as opposed to small data – that’s just for wimps). Business folk simply can’t get enough of it. Data telling them what their customers are up to. Data telling them how their products and services are performing. Data telling them how ‘productive’ their employees are. Governments are now even using big data to help subtly influence the behaviour of their unsuspecting public. For example the British Government has a well published ‘Nudge unit’, or behavioural insights team to give them their official name, which in their words, “help people to make better choices for themselves” (and no doubt the Government).

Data-driven design

And of course data is also a big thing for UX, especially in the context of data-driven design. On the face of it data-driven design sounds great. Rather than using hunches, some mythical UX design sixth sense, or simply letting the Hippo in the room (highest paid person’s opinion) dictate things, you can make informed design decisions based on cold, hard, unflinching data. You might test out multiple variations of a design to find out which performs better. You might benchmark a new design against an existing design to see which has the best KPIs (key performance indicators). You might see which features get the most usage and focus your design efforts on those.

Data can be used as a handy shield with which to fight off bad design decisions, and a mighty sword with which to slay your design enemies (metaphorically of course – it’s considered poor form to bring an actual sword and shield into a design review). Data has no allegiance. It has no master. Data is the Switzerland of world, taking neither side of the argument whilst quietly yodelling to itself in the corner. But as with all powerful weapons, data must be handled with caution, and it’s certainly not the fabled silver bullet that some seem to think it is.

Let the data Gods decide?

The more data we have the better decisions we can make. And when it comes to digital products and services the data is coming in so fast we don’t know what to do with it all. Analytics data, A/B testing data, customer satisfaction data, marketing insights data, the list goes on. And data is great, because we no longer have to make any hard decisions, right? Just let the data Gods decide. What should the label for this button be? I don’t know – let’s test a number of variations and let the data Gods decide. Which of these two designs is better? I don’t know, let’s A/B test them and let the data Gods decide. And with the rise of Lean UX, we are now told that we have to prove our hypotheses with data. “Mr product owner sir, I think that this change will improve registration on the site”. “Well son, you give me the data to show me that, and we can make that change”.

Of course I’m exaggerating a little, but it’s certainly true that some organisations have perhaps taken this data-driven decision ethos a little too far. And this extends to their design process. Their teams have unwittingly become slaves to the data, letting the data drive design decisions, rather than the designers and the design team. Just have a read of the infamous Goodbye Google blog post by ex-head designer Douglas Bowman to see that Google certainly used to be guilty of this.

Data – A false God?

Now don’t get me wrong, I think that data is great. Data is a designer’s best friend (sorry to all the dogs out there), but used in the wrong way it can be a force for evil, just as much as it can be force for good (very much like the Force *spurious Star Wars reference alert*). If there is an inability to make any design decision within a team or wider organisation, no matter how small, without oodles of data to backup and support that decision, then you’re potentially in for one seriously slow (not to mention frustrating) design process. Similarly an environment where designers are continually challenged to prove every microscopic design decision with BIG data is not an environment that most, if any designers are likely to want to work in. An environment where creative and innovative ideas can all too quickly get squashed because the initial data doesn’t seem to support them. I’ve written before as to why the user is not always right, and I’m reminded of the quote from Henry Ford that I used in that article:

“If I had asked people what they wanted, they would have said faster horses.”

– Henry Ford

In my experience data-driven design can sometimes feel a bit like the tail wagging the dog, rather than the dog wagging the tail. For those unfamiliar with this rather idiosyncratic but wonderful expression, this means that the tail (i.e. the data) is dictating things, rather than the dog (i.e. the designer). The tail wagging the dog might make for fantastic YouTube videos, but it seldom leads to fantastic designs. Data has no notion of design cohesion. Data doesn’t look at the bigger picture or consider the wider product vision. Data simply runs around with its tongue out chasing a big fat KPI stick.

Data-informed design

Ultimately it should be the dog, I mean designer driving a design, not the tail, I mean data. The risk with letting data run the show is that not only can great design be stifled, all too soon you can end up with a fragmented mess of a user experience. Data can certainly help to make design decisions along the way, and inform the wider UX strategy, but designers should be using the data to help inform those decisions, not blindly doing what the data tells them to do. This is why I’d advocate data-informed design over data-driven design (incidentally I can’t take credit for the term, I’ve borrowed it from a rather fine article called How to choose the right UX metrics for your product from funnily enough Google Ventures).

Data should be informing the design, but not driving it. It’s the job of the design team to drive the design, not the data. Data should be your side kick, the Robin to your Batman, the Dr Watson to your Sherlock Holmes, the Chewbacca to your Han Solo. After all, Batman doesn’t always battle the Joker or Mr Freeze (Arnold, what were you thinking?) with Robin by his side, but he knows that Robin will always be there when he needs him most.

If data-driven design is a bit like the tail wagging the dog, then data-informed design is more like the dog’s trusty nose, telling it where to find a big fat juicy bone. Sometimes the dog will follow the scent, but other times it will simply follow its own path.

P.S. Given that you’ve gotten this far you’re clearly interested in using the power of data to help inform UX design. In which case, you might be interested in my next post titled 12 tips for better data-informed design. Same Bat time, same Bat channel.

See also

• Data-driven vs. data-informed design in enterprise products (Medium)
• Know the difference between data-informed and versus data-driven (Andrew Chen)
• Data-Informed Design: Three Data Stories (UX matters)
• Be Data-Informed, Not Data-Driven: Balancing Design, Qualitative Feedback, and Quantitative A/B Testing (Presentation – userconf)