Mike Atherton is an information architect and a content strategist, having started his digital quest for the gestalt of design in the early 90s as a web designer.
Straddling bravely the fields of content strategy, information architecture, user experience and interface design, Mike does curious (and hard!) things. He designs interactions, builds spaces with words, visuals and other digital elements. Helping organisations deliver better-structured, more-engaging digital products and services, Mike has worked at BBC and Huddle and currently is a content strategist at Facebook London.
[Update] Shortly after this Dialogue was posted Mike’s book, co-authored with Carrie Hane was [published. It is called Designing Connected Content [the link is to Amazon’s page] and is about the end-to-end process for building a structured content framework.
He is one of those guys that you would see posting in the Information Architecture group on Facebook:
I was musing as I passed the separate signs for ‘Valet Parking’ and ‘EV Valet Parking’ how we often set something apart because of its novelty. Electric Vehicles are a new thing (and maybe they really do need valeting differently) but one can imagine at some point soon their popularity will mean they’re just ‘vehicles’ in most contexts.
Is there a name (and better explanation) for this phenomenon, where something starts out a taxonomically distinct, but loses its distinction over time as it tends toward the norm?
I met Mike via a video on Vimeo (see UX Oxford: Mike Atherton on “Designing with linked data”) and instantly fell in love with the way he saw the world, the content pieces it consists of and, as you will see him talking about this the Dialogue –
the design behind the design.
It is a huge pleasure and thrill for me to introduce you to Mike Atherton.
Enjoy the encounter.
Building Spaces Out of Words and Navigation Menus
How do you build spaces out of words and worlds made of navigation, sub-navigation, menus, drop-downs, buttons, links, windows, rounded corners, shadowing, error messages, alerts, updates, checkboxes, password fields, etc.?
As my friend Andrew Hinton says, language is infrastructure. Our understanding of any space or subject begins with an understanding of the things within it and how those things interconnect. A user interface serves to represent those connected concepts, so any design process must first begin by modeling the concepts and relationships inherent to the subject domain. That gives you some terminology and business rules. With a strong understanding of those things, interface design becomes much more straightforward. You’re essentially designing the controls whereby an < actor > can perform an < action > upon an < object >.
Take the time to understand your actors, actions, and objects and the rounded corners and shadowing will follow.
Straddling the fields of content strategy, information architecture and UX what do you manage to take from all these worlds and what do you think they take from you?
IA and content strategy are different sides of the same coin. CS focuses on the merit of the content itself and the processes of making it happen. IA provides the structural framework for that content; a structure which may be reliant on top-down classification or may tease out, bottom-up, structure inherent to the content itself. Together these drive the ‘experience’ of information retrieval and consumption, though from your question I presume you think of ‘UX’ as as the presentational design part. Interface design choices affect our ability to consume content wherever we are, and our appetite to do so.
In the work I’ve done I’ve tried to bring these fields closer together and figure out that the stuff you don’t know isn’t always so scary.
Knowledge doesn’t want to be constrained, it’s messy. But how do we tame knowledge for the sake of a smooth ride through a given domain, space of interest, place of interaction?
We can’t. The best we can do is try to model its complexity with the best abstraction we can manage. Domain modeling is the best technique I’ve found to capture more of that complexity in a way that can be translated to usable products.
What is it like to build User experience into the fabric of the Web?
Thinking at web-scale means thinking about more than just your product. Thinking about how you’re contributing to a bigger conversation. Editorially that can mean focusing effort on original content, rather than duplicating information already available. Technically it can require you to publish semantically marked-up content in formats consumable by people and machines, to get the best possible reach. And strategically it means recognising that no-one does all their fact-finding in one place. The best you can do is to put relevant, quality content where the people are.
Thinking about the current and future state of voice and chat interfaces, with all the “algorithmic helpers” we get, what do you think is the most important to consider when designing UX?
Be honest and transparent. If you set expectations of human-like cognition or understanding, you better deliver on them. Most chatbots right now are less like Eliza and more like an audio command line interface. And that’s fine if you set expectations about what your voice UI can and can’t do. And consider how your design choices affect the relationship people have with their device. “OK Google” as a wakeword was a cold decision from a corporate engineering culture. By contrast, “Alexa’ gets marriage proposals daily.
Designing and Thinking with Linked Data in Mind
When did you first hear about Linked Data and how did it hit you that this is something to design awesome stuff with?
When I started working at the BBC, Tim Berners-Lee had just given his TED talk about Linked Data. It just made so much sense to connect not just the documents, but the things – people, places, objects – that the documents were about. If the whole web could be one big relational database, then it could be one big content management system. Which means anyone could create great products from almost nothing, Lanyrd founders Simon Willison and Natalie Downe once spent a weekend hacking together a product called ‘Wildlife Near Me’ using nothing more than APIs from Flickr and the BBC Wildlife Finder!
Don’t you think we (as content creators, who doesn’t write and create these days :)) are still in the grip of the old “codex metaphor” that is we think content in the frame of a page or maybe an app, we still have a lot to walk till we get to the Linked Data talk. As a working (accidental) taxonomist what are your observations?
Our industry still defaults to designing products from the interface in, rather than the content out. That makes the process sound less informed than it really is, since those interface decisions are often driven by research into a domain and the needs of users. But in my experience, the people primarily responsible for the hands-on design of digital products are people with visual communication skills. That means the process puts sketches and wireframes of interfaces ahead of the planning and structure of content. To get to the point where we’re comfortable designing products from Linked Data we must first get comfortable with designing content-first.
In a world where experiences are increasingly driven by algorithms (many times with with semantic technologies involved), what are feelings (and emotional responses to design) driven by? [you talk a lot about enchantment, magic]
When it comes to content, algorithms right now are mostly there to assist personalisation. To curate the corpus into selections based on past behavior, stated preferences, or current need state. Done right, we have the Spotify Discover Weekly playlist that feels like it was made just for us. Which gives us a positive emotive reaction to the brand. Or Google Now, pushing flight or ticket information to us right when we need it. We’re at a point where these advances still seem unexpectedly clever, and deliver us something of value. That combination of intelligence, unexpectedness, and value can drive love and loyalty for the service provider.
Given the beauties of the Semantic Web, why do you think “semantic web methods” are so hard to “sell”?
Because it’s boring. Because in all this time it hasn’t fulfilled its promise. And because aside from Agile and Diantetics, no one ever bought a method; what people buy are outcomes. But tell people that properly-structured content can boost their search rankings, or put their movie showtimes, recipes, or ratings right into Google results, and they’re interested.
How do you manage to explore the lands of the boundless language and then get back to tasks as simple as creating a button here, adding a tiny info box there? There must be some magic pill :)
Eventually everything connects :) That button is probably labelled with a verb-noun pair (like ‘Invite Member’) which is a surface manifestation of a controlled vocabulary of agreed terms for all system actions and objects. That tiny info box might include content about China, which you want to assert is the country and nothing to do with porcelain. Keeping all these entities under control is just as much a creative process as anything at the interface level. It’s the design behind the design.
At its core, to me UX is about designing a relationship – which is kind of brave and very optimistic (to try to predict what the other side will need, want, have time for). What is your biggest challenge in doing so and what is the reward for taking the time to be considerate at creating smooth experience?
I find the biggest challenge to any kind of holistic and consistent systems thinking is the ‘fail fast’ shipping culture ushered in by Lean and Agile. Products expand and course-correct in small iterative increments, and design execution can feel less like architecture and more like an endless series of home improvements. In any kind of digital-physical relationship design, you need to orchestrate all the touchpoints; every email, notification, product description, package design and customer service experience. They all have to work in concert. A coherent voice and tone system throughout. Consistent design patterns throughout. Coherence brings the illusion of simplicity, since we can rely on what we learn. That’s difficult to design for when you’re fixing tiny leaks one Jira ticket at a time.
UX and UI in Plain English
If you were to use simple words, words that are outside our context and our immediate understanding of things as they are in all things digital. What word would you use instead of content strategy and instead of UX? For web writing, for example, I use exchange, or “tell someone something :)”
I often describe content strategy as “UX for Content”, since it advocates applying the same user-centered principles to content production as we do to product design. For the less UX-friendly, I’ve described it as “communications management” since apparently that’s already a real job. As for UX itself, who knows? You’re right that it’s about relationship design so maybe ‘digital relationship design’?
Also convincing businesses that UX is paramount, as is content strategy, is hard. Why do you think this is so hard?
Nobody buys process. They buy results. They don’t care how the sausage gets made, so practitioners shouldn’t expect stakeholders to get excited about sticky notes and Sharpies and Blackwing pencils in quite the same way. Frame your work in terms of business results, and then go ahead and deliver on those business results. That’s how you get investment.
Content Strategy, Retro and LOD
What brought you to the beautiful world of content strategy?
I started out as a web designer in the 90s, when the web was all content strategy. But I was always looking under the hood, trying to get to the gestalt of design. That led me to information architecture, which is my love but not without fault. IA can be a little clinical and based on content classification. So the next step was to get close to the content itself. At the heart of every good product or service is good content. Also people said there would be cake.
Where is your love for Retro coming from? :) [ref. Most of your slides on Slideshare :)]
It isn’t retro – I’m just old. I’ve always been fascinated with yesterday’s vision of tomorrow. It reminds us that no matter how we imagine the future it will always turn out differently. And that wherever we are now, we still have a long way to go.
If you were to use Linked Open Data for one cause, what it would be?
To make medical science more open and connected.
Quick Favourites
Favourite process in the product design
Usability testing
Favourite design paradox
How UI is UX and UX is UI, but UX design is not UI design
Favourite content strategy challenge :)
Here I’ll defer to the great John Saito who called out engineers for asking their UX writer for ‘one word’ to describe . Oh, you said favorite?
Who’s gonna interview the interviewer: Their Majesty the Narrative and a Nautilus
Mike, I have a section in the Dialogues called Who’s gonna interview the interviewer and I would love to answer a question of yours.
Mike: Teodora, you describe the Semantic Web as a way of ‘weaving narratives’. How will this non-linear web of connected stories affect the way we perceive narrative form?
Teodora: Wow, thank you for a thoughtful, three-dimensional question!
The moment I finished reading sentence, an image of a nautilius popped in my mind. If before I thought the narrative form (and content strategy as a tool for providing the right space and wording for this form to emerge and live) is very much like Ariadne’s thread (see Content Strategy perceived as Ariadne’s Thread), these days I feel more connected to a vision where a narrative is a multidimensional nautilus with Fibonacci numbers woven into its structure.
I searched Google to connect my narrative to someone else’s and the algorithms did their job, surfacing this:

This metamorphosis of the narrative I see is two-fold.
First, we will have more ways of knowing (isn’t it beautiful that narrative comes from the latin narro, which comes from the greek gnarus – “knowing”).
Second, it will expand the planes a piece of story touches. It will bring us the desire and the opportunity to allow for multiple perspective on an infinite number of levels, while still grounding us in one starting (for someone it could be the ending point) point of knowing. One point, in a different context will serve as a middle, a start or and end point (to use Aristotle’s story skeleton) simultaneously.
And in that swirl of worlds and words we will be seeing things on many levels. The question about how will this be happening technically still remains wide open and hard for me to grasp from the perspective of code and design. And this is why it was such a pleasure talking to you about things content, design and information architecture.
That said, maybe it’s time to thank you for a beautiful, enriching Dialogue and for the many doors you opened towards more awareness when it comes to content!
And let’s leave the narrative open …