The afternoon of Interaction16 workshop day offered the event named “The network as a design material: distributed systems UX for the internet of things”, which I had a pleasure of attending. According to the workshop hosts, Claire Rowland and Helen Le Voi, it is predicted that 33 billion devices will be connected to the Internet by 2020, three times as many as in 2014. The range of services that will be built on top of this enormous mass of devices will range from smart homes, over wellbeing assistants to emergency services and beyond. Some of these will be extremely critical and will be saving lives. Others will be less critical and simply making lives more convenient. One thing is common for all of them: they will require network connectivity to bring all the devices together into a seamless experience.
Can we assume that the network connectivity will always work as expected? Certainly not. Network connections in most cases involve a heterogeneous mix of complex protocols and technologies.
Outages, glitches and delays are part of network’s everyday. When they happen the digital service user might be experiencing unresponsive connected physical objects that were otherwise responding immediately in an unconnected world. This might be difficult for a user to accept, once it was taken for granted. Take an example of a light bulb, which is turned on and off by a physical switch in an unconnected scenario. In IoT world, that same light bulb might become unresponsive to the switch implemented as a mobile application. Reliability and latency of the network connection and limited power of the connected devices will translate into unreliable user perception of the real environment being sensed and controlled through these devices. The dimensions that should be understood and explored in the IoT case range from most visible like are UI/visual design, interaction design, industrial design, over medium-visible interusability and conceptual modelling to least visible like service design, platform design, and productization.
To get our brains into mode of thinking what all must be considered when designing a service over a volume of connected objects we took several cases to think of and ideated in groups. The cases of feeding a fish in a fish tank through a remotely controlled automated system and sending a beacon automatically off the smart mountain-hiking jacket each contain their own set of challenges.
Notifying the user about whether the fish is hungry or not and whether it has been fed or not can go wrong in several ways. The information may be unreliable, or reliable but many hours old. Communicating that the fish was not hungry and not fed 2 hours ago does not require immediate user attention. However, information that the fish was hungry but not fed 2 hours ago may require immediate user attention.
Similar is the case of beacon-equipped mountain-hiker jacket. In addition to the previos case the power limitations of the beacon’s battery Beacon power limitations may not even permit more timely information. Having 2 hour old information that person was all right does not require immediate attention, if any at all. However, having 2 hours old information that the person was in trouble in an unreachable mountain region requires different sort of emergency that needs to be communicated to the beacon user (i.e. rescue crew member).
In summary, the real status of environment that the IoT system operates in might not be the same as the one that is communicated to the user, or it may be the same but somewhat delayed. User experience designer needs to identify what are the possible environment states, and what all can go wrong in conveying correct information about that state over the network. Different mismatches (incorrect, correct but outdated) between the real and conveyed environment state call for different actions from the user and therefore shall be communicated to the user accordingly. Approaches based on assumptions that everything works fine might work in majority of situations and may positively surprise the user. On the other hand, transparent approaches might be much more appropriate in some situations, and might be more effective in building desperately needed user trust.
Rowland C., Le Voi H., The network as a design material: Distributed systems UX for the internet of things, http://interaction16.ixda.org/workshops/the-network-as-a-design-material-distributed-systems-ux-for-the-internet-of-things/, Accessed on 3.Mar.2016