By: Roger Hislop
Anyone who was at Tech4Africa in 2010 will remember ClearLeft’s Andy Budd and his awesome presentation on user experience design. He drew the parallels between a great retail store and a great website design, and how this often fails by designers having the wrong departure point. His colleagues Cennydd (pronounced Kenneth if you were wondering) Bowles and James Box took to the Tech4Africa stage again this year, talking to a packed room.
These two tag-teamed each other, talking about UX design, but focusing less on the gee-whiz concepts and more on process and techniques.
They started with the old UX ra-ra – UX is now all-important, the most successful products of today have been designed with superior user experiences: the age of competing on features is over. It’s user experience, user experience, user experience.
They then traced the path of UX design from the very first computers – from ENIAC and the old punch-card machines. In this world, operators had to learn the language of the machine. As computers have got more powerful and the interface can be abstracted, controls are no longer mapped to the guts of the device.
Now, it’s about huggable technology. In books like The Brand Gap we can see that these are concepts marketers have understood for a long time: commercial success comes from bridging the distance between expectations and experience, between assumption and real life. This is the role of the modern UX pro.
In UX the thinking and discipline has become clearer and more formalised, with a hierarchy of needs that moves from the task-led beginnings, starting with something being functional, then useful, then reliable, then usable, then convenient, then pleasurable, and finally meaningful. The end-goal being to address the experience, not what the user may have wanted to achieve.
Bowles talked about experience design as a profession – one which is described by this umbrella term that stretches over a range of disciplines from info architecture, to research, to usability studies and more.
The key to good user experience design, said the Clearleft twosome, is that it’s not a checkbox to be ticked, it needs to be integrated into every part of the process right from initial product strategy to final release. It’s becoming a clear discipline, with career paths and levels of seniority as companies start to build UX into their business structures.
The presentation looked at a four-step process:
Step 1: Design research
Begin with a pause… understand the people who will use it. This part can easily lead to an immediate taking of the wrong tack. Design research is not about listening to a wish list of features from user focus groups and feedback forms. They most vital part is understanding user behaviours, their problems, how they conceive of and understand these problems – and how can we (as UX professionals) solve them.
Their key tips were:
– Not look at market research as core, rather design research around quite specific, personal, and one-on-one interactions with possible users.
– Conduct ‘expert reviews’ – using experienced pros that evaluate against tested design principles (made for humans, forgiving, accessible, self evident, predictable, efficient, trustworthy), and doing competitive analysis (the risk here is significant, to get hung up on what competitors are doing and everyone chase their tails, not innovating)
– Look at analytics closely – although again they warn against getting hung up on analytics just because it’s ‘hard data’. Analytics tells you WHAT people are doing, not WHY. It’s important when studying analytics to talk to people what they were doing (maybe they took a path to a page because they couldn’t find another way, not because they preferred it!)
Research methods recommended
– Interviews (one-on-one)
– Focus groups (as usual with focus groups beware of danger of groupthink, or a dominant individual driving the discussion)
– Questionnaires
– Diary studies (where people record their interactions with a system over a period)
– The quick and dirty corridor tests (stick your head out the office, grab someone walking down the corridor and stick them in front of your new design, and watch them use it)
Once you’ve done your research, you need to produce some outputs – normally a report. The speakers were quite anti the normal practice of producing a couple of slides with some nice charts of the numbers, as this is often not useful when resolving design disagreements.
They suggest rather using a persona (a fictitious user that represents a user group). It’s a lot easier to design for a person than for a heap of data, they say, and each persona acts as a common reference point that is clear and intuitive.
Reports can include the personas, and they maybe even extend into comics and storyboards that show how the personas would integrate your system with their life – the idea is to build insights into a broad view of the user behaviour.
Step 2: Generate Ideas
Once this research has been done, it’s time to generate ideas.
You’re all excited, you have lots of information, and even understand your users. The temptation is to jump in and start to design right away.
No, say Bowles and Box. As in the classic tome on chess strategy, “when you see a good move, look for a better one”. You need to explore the possibility space. At this point you can throw many ideas at the wall to see what sticks – risks are low, nothing to change, nothing to hold on to.
If you jump in and start designing immediately, there is a temptation to fixate on obvious solutions.
To generate ideas, there are some conceptual tools that can help. Two good classic books in this area are the advertising industry standby, “A Technique For Producing Ideas” by James Young, and Edward de Bono’s “Six Thinking Hats”.
They are big fans of sketching – taking a koki pen and piece of paper and scribbling down ideas and layouts. It’s quick and easy, says Box, and importantly you don’t get attached to them.
Sketches can also be posted up onto a sketch board – basically a big board with rough maps, sketches and mock-ups pinned up and arranged in some kind of rough order. Sketch boards encourages stand-up communication – trying to resolve design disputes via email is often a giant, slow fail. The other advantages of sketch boards in the ideas generation phase is that they are clearly not finished, so it invites input.
Step 3: Detailed design
Historical UX designers and visual designers worked separately. In modern development this is increasingly seeing this as undesirable – with famous architect Frank Lloyd Wright’s adage, “You can fix it on the drawing board with an eraser, or onsite with a sledgehammer”.
This is where all the work is turned into something concrete, the dreaded ‘deliverables’, the risk that management pressure and corporate culture can fetishize these deliverables into an orgy of reports and planning.
Some of the types of deliverables include logic maps, site maps, storyboards (which are better for a more interactive site), and wire frames – although these make it difficult to capture something really interactive, so instead you may need a prototype.
For very top-down design management, there would also need to be a functional spec – great to keep middle managers and procurement departments happy, but in trying to describe how every function works they often end up being milestones that document the chasing of a moving target.
Step 4: Testing and iteration
Finally: get feedback, and test the ideas on a real site. Say the UXtabulous twosome, “You need to accept that first attempt will not be right.” There is certain emotional baggage around this that you need to keep an eye out for: Egoless vs Genius design. A bad designer will cling to concepts too long, especially if they think they’re genius.
An example of truly successful Genius design that’s often held up is Apple – although, says Box, the genius is probably more that they’re just really good at making sure we didn’t hear about all the design iteration.
Usability testing should include all stakeholders – and a structured critique is often missing from design process. This means user feedback, analytics, and even A/B testing. A risk in formalising testing in a clumsy way is codifying incorrect assumptions simply to be able to check the “tested” box.
You can compare Summative vs. Formative testing. In Summative testing you answer the question “does this work”. If the answer is no, there may be pressure to ignore or misinterpret this testing, as the cost of changes at the end is quite high. Compare this to the Formative testing, where you “test early, test often” to inform the design process.
The presentation touched briefly on some testing software, such as Clearleft’s own cheap’n’chearful Silverback, or something like Morae.
Their final bit of advice is to look to improve, not perfect, in the testing stage. Perfection is not possible, and this is the real world.
Where is UX design heading?
To wrap up, the two talked about where UX design is heading. There is a (somewhat debatable) quote that “Our medium is not technology, it’s behaviour”. More interesting was the concept that “behaviour is a function of a person AND his/her environment”, that how someone acts changes with context.
There is interesting work being done in choice architecture – it has become clear that decisions can be deeply influenced by the choices that we are presented with, and how they are presented. A lot of the work UX people do is around changing behaviour, which requires understanding the psychology of users.
A final nugget of wisdom is that good design is less about UX tactics, and more about what we can do to make their lives better. In the modern world this means there is a growing focus on cross-channel UX, where there may be a native mobile app, a mobi site and a full site, all performing slightly different functions for the user. The modern UX designer has to make sure these are all consistent, but make the best use of each device – and also synch with real world stuff like the physical presence (retail stores, etc).