This website is now archived. To find out what BERG did next, go to www.bergcloud.com.

Blog posts from September 2007

Drawing Olinda

Drawing hybrids and inbreds

We are around half way through the development of Olinda, the digital radio prototype we’re building for the BBC. Most of my efforts over the weeks since Matt’s post have been focused on how the object should behave and physically manifest. 

This post discusses some early drawing processes. We use drawing to surface and test many ideas easily and early. These drawing processes are also used to reach unexpected forms, and to examine why an object should look like it does.

About three weeks ago I met with Matt Ward from Goldsmiths. Ward has developed a drawing process which he works through to explore and interrogate ideas. Here we used it to develop ideas around products. His position for understanding how a product can manifest begins with a framework that includes how objects respond to anticipated contexts and tasks, in situations within a culture of consumption. He sketched this for me, and I’ve included it below. I like that it includes the designer, in a ‘context of production’. 

Ward diagram

Ward’s approach is this. Begin by taking an existing radio, and draw it at the centre of a page. From here, choose four contexts or situations for development, like ‘in the kitchen’ or ‘listening to the football’. Write these labels in the four corners of the page surrounding the original sketch of the radio. Then evolve the form in the centre towards imagined new forms in response to the four situations.

The point here is to get away from the original form as far as possible, and to make many drawings. Below there are radios that are – more and less literally – in the contexts of decorating, the bathroom, kitchens and shop shelves.

Early Olinda Sketches

Sometimes this leads to very strange things.

Hand blobs

Critically, the purpose for such an exercise is not to draw good products but to begin evolving forms outside of an expected mold. As soon as a form emerges which catches, it is redrawn on a separate page, and bred between other sketches to develop new hybrids.

Olinda radio hybrids

One is being selective in this process, but it is surprising how little control there is over what you expect to emerge, forcing issues with the sketches rarely yields anything satisfying. But this is not a storming, random process. It is very methodical, as a process of deconstruction. It is using drawing as thinking, which is its power.

What emerges is the discovery of what it is about that original radio that persists, in spite of the violent evolution. The drawings are really about ways of housing these commonalities, so you start thinking in terms of materials very quickly. The other thing that happens is you see particular twists. For example a kitchen radio should have legs, in order to sweep crumbs out from underneath it.

Making and drawing

In parallel to these processes, we have been in the workshop making objects from which to derive further drawings. This process started by thinking out a critical aspect of the form, in this case the connection between the two separate Olinda modules.

Early connector experiments

Once things start to get made, materials start to influence drawings and further made experiments. As the pieces of wood were cut, the shapes started to yield new directions and the wooden blocks emerged as a combinatorial way of interrogating traditional and less likely forms.

Early form tests

These are then fed back into the drawing and imagined interfaces are penned onto surfaces.

Early interface drawings

Some of the drawings begin to imply unlikely material qualities. The social module here looks like it’s been knitted from wool. The drawing is from a little over a week ago, and is based on a model used to investigate certain materials and assembly.

Olinda wool module

When Olinda is an object, it will be a product of unusual influences. It is unlikely that in this project such radical deviations from expected form will be appropriate. But these processes have made it possible to interrogate the assumptions embedded in the form of products. Objects like Olinda respond to forces from many territories, but the reasoning around that is a separate discussion.

The Experience Stack revisited

Since the central point of my experience stack presentation was somewhat obscured because of my playing with the structure, I thought I’d have another bash at it here. Setting an idea up like this feels unnecessarily dogmatic, but frameworks are only meant to be rocket boosters to the actual design so it doesn’t feel too constraining. Anyhow. You can read the original presentation in addition to this article.

Five layers

The experience stack is a way of thinking about the different levels at which experience design operates. Experience design can be thought of as…

  • branding;
  • service design;
  • product design;
  • interaction design;
  • human factors.

Just as with the OSI seven layer model of computer networking, these layers can be thought of in a tiered stack… but not reducible to one another. It is not possible to express the experience of discovery, in the service layer, to the cognitive components in the human factors layer, for example.

Let’s go into these, from the bottom up, and look at the contribution to experience from each.

Human factors

Human factors covers physicality and cognition. Cognition I’ve covered in the Mind Hacks-derived presentation, Assumptions, Attention and Affordances: it means we have to know how people pay attention and the limits on it; the impact of the workings of visual perception, and how things like arrows, shading and visual change are more important than we realise; and how to take advantage of all of this. I would include peripheral vision and tactile feedback in this.

Physicality is about understanding two things: First what I’ll call body thinking (where physical movement affects our emotions), and second, the physical and physical context of the interaction: how does one stand to approach an object, like a radio; how do its movements and shape indicate the possibilities and constraints of interaction. See for example our material explorations in wood and have a look at the identical and complementary shapes photo: that construct would be perfect as a replacement for the Bluetooth pairing process, because it shows clearly and two (and only two) objects are supposed to join together. Or also read about how to make objects ‘disappear’.

Interaction

The interaction layer includes many of the ideas in The Hills Are Alive with the Sound of Interaction Design and From Pixels to Plastic: it takes common interaction patterns like play and sociality (see social software), and positive and negative emotions and drives, and uses them appropriately.

In the Experience Stack presentation, I discussed the different ways in which games start… learning from successful interaction patterns and applying them to the product at hand is definitely part of this layer. (For example, seeing that we enjoy observation can lead to product feature ideas.)

Knowing how people will adapt their product – applying customisation – and showing how that is possible is also, to an extent, part of the interaction layer.

It’s about making the interaction not just something to be learned from the manual, but part of a pleasant, intuitive, engaging experience–and the best way to do that is to learn from other experiences.

Product

In the middle of the stack is how people experience their products–not while directly interacting with them, but nor in the more distant, less controllable way of brand. Products should be firmly identified: a product which has a sentence to identify it and a target audience can also have goals and metrics to tell when it’s performing well, and will be better understood by its team, and its host organisation. Customers will be able to tell other people about it, and develop a personal understanding of the product and how it will behave.

Metaphorically, a product should be ‘shelf-demonstrable’–able to be understood simply from a first look, even if that understanding growth in breadth and richness in the future.

The point of this layer of the stack is two-fold. It says that the best way to have a good product experience is simply to have a good product. But it also says that predictability and a common understanding of the product as a discrete thing is important (it is for this reason that I’ve said previously that products are people too).

Note that this doesn’t mean that a product must only do one thing, because people are capable of seeing very abstract bundles of behaviours as ‘things’… but if a product does two things, the conception of what it’s truly about may need to change. And adaptive design points out that products have fuzzy edges that change at different speeds, but that’s just something to incorporate into the overall understanding of what a product is, and design for accordingly.

Also in this layer are firm identifications of the other actors and situations. Who will be using this product? In what situations? Considering archetypal people and situations can help target product features.

Service

Experience hooks (more) are those moments you remember in your story of your life with a product: how you meet, how you show your product off to friends, how you clean it. When I’ve discussed this before, I’ve referred to unboxing as a key moment in the product life-cycle.

Designing for these moments in the product-as-service is the service design layer of experience, and lets us associate brand ideals with product features. For example, certain experience hooks can be made a feature of–especially social, or particularly easy, for example.

Brand

The experience of the product is cloaked in messages, direct and indirect, which set up mental understandings of the product. A person then uses these models for implicature–the understanding of a product beyond its explicit communication.

It’s rare the brand – this high-level experience – can be seen as separable from the product itself, but Volkwagen’s Night Driving commercial and the site They’re Beautiful by Jackson Fish Market both create experiences beyond the product. These brand experiences affect the use of the product itself because the different layers of the experience stack are indistinguishable in people’s mind, and even weakly associated still bleed into one another.

Consideration of the brand influences what design and feature decisions are made at every other layer in the stack.

Rationale and approaches

Central here is the Generation C trend (people form communities; expect to be connected socially and electronically; Gen C are comfortable with complexity and want control; they are creation and expect co-creation), and their demands. And then ideas like social software and adaptive design point in the direction of a more holistic kind of design. Experience design is what bundles this all together, as it implies that all aspects of the product experience (from every part of the stack) are considered as one by the user, customer, or viewer… and if the experience is awful, they feel awful.

Approaches come out of using the stack–looking for experience hooks and considering the context or situation can help generate ideas. Using the brand as a guiding principle can help select features. Running through the interactions with the product and making sure they’re aligned with emotions and behaviours that are enjoyed or avoided also helps. These are all very direct uses of the experience stack.

More generally, it has to be realised that experience is very badly understood by observation: the designer has to take part. I can sum this up: Nothing is easier than believing we understand experiences we’ve never had (source).

The Experience Stack at d.construct 2007

At d.construct at the end of last week, I presented on the experience stack. This is an idea I’ve been mulling, thinking about the various layers of experience in a way analogous to the seven layer OSI model of computer networking.

As usual, I’ve put the slides and transcript online. Read the Experience Stack.

Some portions of the presentation went well, but overall I wasn’t pleased with my articulation of the ideas. The comments I’ve received and read online are also mixed–some people got a lot out of various slides, while others felt presentation was confusing. That’s a shame. I used to feel okay about confusing pieces, because nonlinearity can be fun and confusing is a price you pay. But I’d prefer now to have a baseline accessible to everyone, so I want to review what happened.

What made the talk confusing

I threw the structure baby out with the bathwater. My presentations so far this year have shared a very definite rhetorical structure. In an attempt to get some distance from this, I organised the material as a fool’s alphabet, starting with A is for adaptive design and finishing with Z is for zooming user interface.

The scattered approach was intended to mirror the experience of people with products, where every pattern is assumed meaningful and so we have to design for all levels of the experience stack simultaneously.

It also meant, however, that the fundamentals of a good presentation were skipped: I did not say, at the beginning ‘this is what you’ll get'; I did not, at the end, give a simple, graspable take-home message (violating my own point about treating everything as products with simple statements of intent); I did not use pacing to help people know when I was illustrating versus when I was building up to a more substantive point.

In short, I took away the contextual cues people use to feel situated and comfortable within a presentation, and removed the explicit ‘resets’ I usually use to help people who have drifted off get back into the flow. In short, people felt lost!

I had too much material. There are simply too many examples in this presentation, and they aren’t grouped well because of the constraints I set myself. I usually leave room to dive into examples if the audience are looking lost, but because of the quantity of material I didn’t give myself time to digress on the fly.

I tried a live demo. No matter that it worked in AV check on the previous day… I was using a Wii remote control to pan and zoom an image embedded in a Quartz Composer movie embedded in a Keynote slide, controlled by an application broadcasting Wiimote data to a custom-written Quartz Composer patch. That’s precarious at the best of times. Ironically it functioned perfectly well technically–I just forgot that I gesticulate a lot more than I notice when I talk, that my hands shake, and that – when I’m on stage – I get so focused that I forget what buttons do what (seriously: I can’t even take photos on stage unless my camera is already turned on, because I can’t remember what button does what).

Other technical problems were the metadata portions of my slides being cut off (my computer recognised the projector differently on the day), a couple of movies freezing (I think because the presentation slides had been set up a lot longer than I’d tested), and the loss of sound.

I didn’t explain the core premise sufficiently. This is partially due to the absence of structure, but also because I could have used the experience stack as a common thread but choose to leave it till later to introduce it. Even then, I didn’t spend enough time explaining myself. Consequently some people didn’t get what they came for (and for those folks, I’ve written up the stack concept itself in a separate post).

Why I made these mistakes

There are two reasons. First, I tried to change too much at once from my usual style. There are problems with using the old rhetorical structure, so I needed to find an alternative… but it would have been better to make that shift with old material. I don’t understand this new material well enough yet to really, really condense it, and my discomfort with this style of presenting almost made me freeze on stage about halfway through. It’s been a long time since that happened.

Second, I took a risk and it didn’t pay off. I wrote the full talk in long-hand before delivering it, and it looked unusual but I genuinely thought it would work. In retrospect I can see that it works better as something written: you need to be able to refer back to make sense of the structural reveals and the examples.

What I was pleased with

All of the above makes it sound like my talk went terribly. It didn’t – I’ve had good comments too – but I’ve had a few talks recently go as well as I could ever hope and that’s pushed my standards pretty high.

I’m still pleased with the experience stack concept itself. I think I have, now, a decent way of understanding and explaining experience design, in a way that draws together my various other explorations.

I enjoyed putting together the slides. I colour controlled the slides again (as discussed in this first slide of Products are People Too), and it really binds them together visually. And the first slide uses gentle video in the background, with the ocean washing the beach behind the title. I’m going to do more with that.

This was also my first outing of the new version of last year’s 2D presentation code, this time integrated with Keynote and using a Wiimote. It didn’t go smoothly, but it had to be used to be understood and I know a lot now about what needs to be built.

Finally I was pleased – and thankful for – the friendliness of those who came to d.construct 2007. I talked with a ton of people afterwards, and my talk got a good reception with at least some of the audience. It was fun to see the notes they’d taken, and to hear about their work.

Books

I’ve been asked for a list of books (and other things to read) mentioned. They are, in order:

And the presentation slides and summary of experience stack idea are also online.

Say hello

A couple of months ago I put the feelers out for interns. Well, we have two starting today. Say hello to:

Alex Chadwick is an electrical engineer. In-between touring the canyons of the US and university in October, he’s coding and building the guts of Olinda, getting us stocked up with microcontrollers, and has me really, really wanting a decent oscilloscope (apparently it’s super useful, but I’ll be happy if it can sing and dance). Many thanks to David Smith for putting us in touch.

Jeff Easter is an interaction designer currently studying Design Interactions at the RCA. He’s been with us for a week before, on user flows and screen design, and for September it’s more of the same, plus photography, carpentry, and iterating, iterating, iterating on form design.

Both super chaps. It’s going to be an excellent month.

Recent Posts

Popular Tags