A Social Web of Things
Back in 2008 we started working with some projects involving interaction design for large networks of connected products and services. Not Internet of Things-ish stuff, but in that direction. The challenge was to come up with a graphical user interface which was both scalable and very easy to understand. Or that was what we thought the challenge was to begin with.
The scaling-part was difficult. A relatively small number of things, say ten fifteen or so connected multi media devices, was easy to design for, but hundreds of totally different kinds of devices, vehicles and appliances, environments plus a layer of various ubiquitous services and data sources was quickly becoming a bit messy. Apart from the obvious issues with limited screen real estate and graphical complexity, we discovered an implicit and more of a pedagogical problem: it wasn’t just the GUI that was hard to scale; in a way people’s minds didn’t scale either. And it is not that people are stupid. This has nothing to do with tech-savvy or IQ. It has to do with mental models and the ways we conceptually understand something. From the studies we conducted we found that people’s mental model of the Network as such didn’t scale.
Mental models are internal images or representations of something which we use all the time to make sense. People in general, myself included, have little clue about how networks actually work, but most of us try to guess anyway, consciously or not. So if something (say a system or technology) is opaque to us and it seems to just work by magic, we try to make sense of it anyway with a mental model. We can often hear echoes of mental models in our language when we use metaphors, metonymies and figures of speech. When we for example talk about things like Wi-Fi, 3G or Bluetooth we call them wireless network technologies, relying on the image of a wire to give an idea of how it works and what it is.
And it was through talking to people we discovered something. It seemed like the general mental model of the network as such was something like “very many point-to-point connections”. In one way that is absolutely correct, the physical network infrastructure is a lot of cables connecting everything. But then most people don’t differentiate between what a network is (physically) and the mental model for how the networked objects works, which means that we tacitly understand the ”network-ness” of products and services in terms of serial point-to-point.
This is quite different from the “continuous dynamic many-to-many interrelations” which was the characteristics of what we tried to design into our GUIs, and the implication was that nobody understood the value or capabilities of our concept prototypes. The mental model of point-to-point caused a mental blind spot making it difficult to imagine dynamic holistic and heterogeneous “total network-ness”. (I’ve previously written about how this may be a problem for the whole industry, here)
Said in a different way – we designed for driving a car while people, who had never seen a car, tried to make sense using their knowledge of horseback riding (loosely paraphrasing something Henry Ford allegedly once said, only upside down). Initially the car and the horse provide somewhat similar functionality, but one level down the concepts differ so much that people got confused and didn’t understand.
We have of course been thinking about a solution. The thing is that this particular discrepancy between mental model (inherited from the analog ancestor, the Cable) and concept (such as an Internet of things) is valid only when looking at technological networks. The Nature or social relations are examples of other heterogeneous eco-system-like networks which we understand differently. Probably better. The concept of ‘friendship’ and ‘social relations’ are for example understood so well by virtually everyone that they are intuitive to us. So with that in mind we came to think that a solution to both the practical scalability issues and the mental model/pedagogical issue could be to simply “dress” a network of things as if it was a social network. We imagined a social network service for your connected products and services where they could talk and say what they do or need, follow each other, discuss with each other, collaborate, create events and do things together (as in mash-ups) and perhaps even ‘like’ or ‘+1’ or ‘retweet’ each other. That would capture the core characteristics of the “network-ness” in a familiar way.
We called this concept the “Social Web of Things” (Social Web + Internet of Things). We built a few prototypes which we showed to selected people in different parts of the world and literally everyone we showed it to got an intuitive understanding the dynamic “network-ness” between everything, no matter what kinds of things or how many. Social-ness replaced tech-ness, and bingo! Problem solved!
Or rather, it solved our two initial problems: the scaleability and the understandability. Of course, just making things talk and borrowing interaction design principles from Facebook or Google+ doesn’t solve everything, but it helped people understand by providing context and meaning to the rather abstract concept of an “Internet of Things”. It simplifies something complex without reducing the complexity, which means that ‘average people’ became aware of some of the more important and more interesting challenges and opportunities provided by having literally everything connected. Some of the more important Internt-of-Things-challenges we can think of are related to for example trust, integrity, security, information overload, technology-dependency, loneliness, anthropomorphism, playfulness, morality, ethics, consumerism, ownership, business models, open APIs, industry-horizontalization, governance etc.
The two films below were produced for Ericsson’s exhibition booth at the GSMA Mobile World Congress in 2011 and 2012. The attention and reactions it received from the press and public may indicate that people do relate strongly to the just mentioned challenges.
We will continue exploring. The Internet of Things is already here. It’s user experiences are just not very well understood yet.