Experimentation beats expertise

I found the design process utterly transformed once I decided to stop trying to be the expert and start trying to encourage a culture of experimentation.

Battles that would rage, angrily, for months – dying down when the provocateur was busy with other work but rising up as soon as they had a little time on their hands – these battles began to go away. Long frustrating and unproductive sessions of trying to explain, defend, rationalise why the design that I suggested had more merit than the many and varied suggestions (or requirements) coming from stakeholders all but disappeared. People who would usually sneer derisively at the design team became participating members of the design process.

It’s not perfect, it’s no silver bullet, but for me, it’s been a transformation.

And it’s pretty simple. To embrace experimentation you just need to stop talking about design in a Socratic way (other related but less civilised methods are also very common) and start formalising hypotheses and tests.

Stop having meetings to argue about which design approach is better  - endless meetings with stakeholders full of defensiveness and crazy arguments where the people who tend to win are those who are loudest, most persistent or highest paid. Start making decisions based on lightweight research that provides evidence (sometimes stories, sometimes numbers) to support the design that most strongly supports the agreed goals.

Goals. That’s one pre-requisite you need for this experimental approach to work. You need to have agreed what your goals are for the design. What success looks like. Without this agreement, no change to methodology will save you.

The experimental mindset is an egalitarian approach to design. It allows that anyone can suggest a design solutions and, rather than argue endlessly about whether it is better or worse than other approaches, you design a test. Find out how to find out which design works best.

Hypothesis, prototype, test.

There are loads of tools you can use to test ideas quickly – from some quick in person user research, to some A/B testing (if you’re not set up to do A/B testing, meet your friend Google Content Experiments and get onto this immediately), to an online card sort, to one of the range of tests that places like VerifyApp offer. The methods for testing are limited only by your creativity and are mostly inexpensive.

Sure, you can’t design from the ground up this way – you will still need a good designer that you trust get you to a good starting point from which you can experiment up, but once you’ve got the framework in place, don’t waste time and goodwill bickering about the details but encourage experimentation throughout the entire organisation. You’ll raise the overall ‘design IQ’ and happiness quotient of your company, your design team and, most probably, even yourself.

2 Responses to “Experimentation beats expertise”

  1. John Goode May 2, 2013 at 9:04 am #

    This is a great approach and has always been your way since we started working on the surrey.ac.uk project. In Higher Education there’s always many stakeholders and gathering opinion in an attempt to reach consensus simply doesn’t work. We had a website that was the product of that approach, so we “flattened it” to start over. The first iteration was the product of ground-up design and it’s a result we’re proud of. It gives us a baseline. Moving forward I hope we’re able to adopt this egalitarian approach:

    “Iterative, goal-driven, test-driven decision making must trump personal preference.”

  2. James May 2, 2013 at 9:39 am #

    Wholeheartedly agree. It’s always best to just get something out there and test, verify, iterate.

    I find sometimes however there’s still the old design argument when you just know something won’t work — but again, at least if you go with it and it fails you have proof it doesn’t work. And if it does work, then sometimes being wrong is a learning experience.

block trafo