5 Tips to Rock at UX Service Design
People very very occasionally ask me what are the methods and tools that I use most frequently and recommend. So over the years I have been compiling a small but value set of tools of the trade for doing DHCD. DHCD? That’s Digital Human-centred Design and as well as skirting around the UX, Service Design, HCI lacuna with that term.
I am also putting forward these tools and methods as starting points for a more ambitious goal of Design Science. So it’s a bit of mixed bag this toolkit but they are things that I use in my day-to-day work and have applied in various organisations, roles and teams. There’s lots of other stuff to do and other methods and tools to use but if you do these you should be well on the way to doing principled UCD.
1. Contextual Inquiry
It’s the best way of eliciting requirements for me and something that stakeholders can understand ahead of the more jargon laden ethnography. It’s also particularly good for getting designers involved in research too.
2. Reuse first
Recycling is cool and good for the environment. Reuse ui patterns and stuff that works. If there is a situation where something completely new is required then treat that with caution. My favourite story on this comes from Bill Buxton who tells the story of a multinational that spent years of R and D developing a gesture ui only to discover it was almost exactly the same as an ancient civilisation’s script alphabet. So don’t reinvent the wheel.
3. Pen and Paper
The most powerful, design and research tool. Use this before anything else. Sketch, create models, talk through and test; then iterate and then use digital tools to polish.
4. Prototype Early Test Early
Following on from the lo-fi approach make stuff early and test it. Make stuff, in whatever form is quickest and cheapest to do and try it out with peers, representative and even unrepresentative users, whatever get feedback and improve when you have good reasons to do so.
5. Structured Usability Testing
I am not a fan of unmoderated or unstructured testing. I prefer to go to perhaps the opposite extreme by using a task approach and layering that with questionnaires, interviews and standard metrics. I use this methodical approach because it generates more data and I think triggers more insights than more open approaches.
6. Always Collect Data
Workshops, meetings, whatever it is, extract data from it. Structure these activities so they generate outputs themselves .
I always use SUS in tests and if I can in general research too. It has validity and beyond that it is good to have your own benchmarked portfolio to compare projects to. I haven’t use so many of the standard usability questionnaires but highly recommend this one. It’s referenced by my UX Data Hero Jeff Sauro which only adds to its value for me.
Staying old school — whatever you can reuse from the golden age of HCI is going to be good as it has rigour behind it and guess what: most practitioners won’t be aware of these tools so you can be one step ahead. GOMS is great for mobile and there are updated timings for touch devices now so it’s bang up to date.
9. Grounded Analysis
I am not sure about the exact term for this but it’s similar to methods found in coding, qualitative analysis and grounded research. It’s very simple but seemingly hard for newcomers to grasp. When you have data, code it; that means layer on a descriptive tag for each item of data.
Collect all your codes and tags, count them up, group them on post-it notes like all of those trendies do, whatever you want to get deeply immersed in it. Then stand back, think, reflect and then create an explanatory model that shows how the codes or categories relates to the data and research question. Iterate that model til it explains everything and is ‘strong’. You’ll know when you have got there because suddenly everything makes sense.
What is usability? Hierarchy of user needs? Usabilty Honeycombs? Layers? Whatever, since 2003 I have used this acronym based on solid research to define the user experience as Accessible, Usable, Engaging and Rewarding. One benefit of AUER is that it closely matches other qualities including business benefits and ethics, you can extrapolate cases from the framework, so it works.
Google’s HEART framework for measuring UX is the best thing since sliced bread. It has industry relevance and uses a holistic set of datapoint that map to AUER. Not too many worked examples, but even as an aspirational approach its good at cohering organisations around a common, human-centred, data driven metric.
12. Design Games
You can find my examples of these on SlideShare. They take group work, structure it around decision points and hopefully package them into an activity that is engaging — maybe even fun. The two I use the most are Ideation Games at the beginning of projects and the Scoping Game to well agree what that project is.
13. The Wall
Don’t keep ideas, design, research in devices, liberate it by putting up, creating talking points, co-creating and all of that stuff. Done at the right time this simple trick can change organisations into design led ones. The Wall is also good to change how people work by making it an easy thing to talk about work and collaborate without over formalising it.
14. Vision over Process
No single process works in all situations so you need a few that map to existing delivery processes and contexts. The three I use the most for different flavours of DHCD are ISO 9241–210–2010 if you want to go for the full waterfall UCD version.
15. Post-agile Freestyle
Then there’s the various Agile/Lean UX process, I don’t have a preference other than stipulating a research driven Sprint Zero. Lastly, there is the accelerator approach, Google have a good one but it’s a condensed UCD process that works over 5 days or whatever short timescale you have. They all work in different situations so hope they work for you too.
The most important part of any process is getting the vision right. That means collaboratively getting to the right place through visioning workshops and user research. Then focusing on the key interactions, perfecting them through iteration. If you get the basics right everything else will follow and you’ll embrace the new Freestyle design approach.
© John Knight, 2016