|Photo: Aiden Kenny|
- The dominance of messenger apps, which are become better attuned to how we operate as social animals.
- The improvements in contextual predictive systems, such as Android Now and Siri, which are bootstrapping themselves to ever-more sophisticated levels. The availability of ever-larger datasets is accelerating the pace and capability of global-scale machine learning projects.
- The rise of push-button service ecosystems, such as Uber. In which the app is merely one surface aspect of the greater ecosystem — which is the true product.
Photo: Aiden Kenny
Adams’ advises that it is critical for designers today to think in terms of overarching systems, rather than of individual destinations such as apps. For the reason that we should not expect our primary computing paradigm today – that of a screen of apps on a sheet of glass – to persist over the long term. Products and services are already morphing into ‘digital fluids’ that flow into different forms conforming to the affordances of their destinations. That process is only going to accelerate.
Change is one of the defining attributes of the Internet. Today this implies that UX and UI designers have much to learn from the skills and mindsets of adjacent disciplines: such as animators because now they need to design for time, and architects because now they need to design for space. There are many hidden traps in continuing to use tools originally conceived for static design to now create experiences that have change, movement, transition, and flow as innate properties. We need to choose our tools carefully, or indeed build all-new tools fit for purpose.
Designers need to understand that we are in the midst of a transition from an ‘Internet Made of Documents’ into an ‘Internet Made of People’. We humans are social animals. So it should be unsurprising that the most successful Internet companies, services, and products are those that leverage our innate social nature. Adams sees the successful systems of the future as those that can best leverage our information to provide ever-more tailored feeds of personalised content. (Just don’t be creepy.)
Intercom innovate at the level of Product Strategy and Product Concept, and not at the level of User Interface. Adams works on the basis that we are still only in the earliest days of the Internet Era. So, as we collectively shuffle forwards out of the dark, we are going to need to rely less on existing design paradigms. More often we are going to have to rethink from first principles. That means there are now amazing opportunities to investigate, explore, define, and craft novel interactions and design patterns.
One design challenge is that to move large numbers of users along any technology innovation curve, it is often necessary to only introduce iterative changes, so that you do not break user’s existing mental models and habitual workflows. Designers and developers are often early-adopter personality types — eager to move fast and break things. We need to remember that most people are not. Even on a (relatively slow) one-year cycle time, changes to Apple’s iOS are still treated as suspect by many types of users.
Adams explores the idea of replacing a screen of apps with a design pattern based on cards. Cards are one of the atomic units of the web, as explained by Brad Frost. Cards can be considers as containers for content from any app. You can observe this today in the interactive notifications in iOS and Android. So that more often we do not need to open an app now to act. The notification itself can be the complete experience for many activities.
Interface of Facebook’s ‘M’ product.
Increasingly the content of cards and notifications will be brief conversations and interactions initiated by bots. Bots can communicate to users via messaging interfaces (think of Slackbot and Facebook’s prototype personal assistant ‘M’). Bots can already take on many basic interaction tasks, and so leave humans free to address more nuanced interactions. In future a considerable amount of our time spent interacting with companies will be communicating with such AI-enabled systems. Just as we all once had to acclimatise to ATMs instead of bank tellers, so people will adapt to this mode of interaction as well.
Facebook’s new ‘M’ product is one facet of their current initiative to overcome their lack of a native operating system. ‘M’ places a novel AI assistant inside their Messenger app. Users interact with the service using natural language texts. Currently this is only a beta service limited within a strict geographic area. At present the nascent AI still has to be supported by teams of humans to ensure it parses user requests correctly.
However, I wonder whether one master-stream of notifications could ever make sense to users? Even today, my full Twitter-stream is such an overwhelming juxtaposition of friends, bloggers, news organisations, apps, parody accounts, and much more, that it is essentially unusable. So I depend upon my Twitter Lists to mentally shift contexts and to filter Twitter into a usable and beneficial service. I work with a different mindset, and range of potential actions, when I am reading my ‘Friends’ list than when I am reading my ‘Innovation & Ideas’ list. Classification tools such as Twitter Lists and Google+ Circles still seem to be more of a power-user preference. That said, lists are also a problematic design pattern in their own right. They make us define arbitrary categorisations. Unfortunately they need ongoing curation and regular maintenance to stay useful. Despite years of innovation and iteration no-one has truly cracked categorisation of social feeds yet.
Content within our apps is cumulative. Even the most diligent data-editors amongst us are always adding more contacts into our address books every year, following more people on social media, and accreting more connections on LinkedIn. These, and all the other data silos on our phones, represent ever-more opportunities for unwanted notifications. Any successful card-based interaction model is going to succeed or fail depending on careful tuning of the signal-to-noise ratio. Will that be better served by the user spending precious time curating and adjusting? Or will an algorithmic solution prove good enough? While we know that spending hours triaging email is neither an enjoyable activity nor a productive workflow, many people also spend blocks of time ‘going through Facebook’ or ‘getting caught up on Twitter’. These are the kinds of siloed activities that the proposed new interaction models seek to minimise.
How can design thinking help us to find the balance between people either achieving optimum flow and living in the moment with all timely and relevant information immediately at their fingertips, or descending into the near-adjacent possibility of existing in a state of permanent distraction?
One issue arising from the card-based systems discussed is – as they become optimised to present an ongoing stream of micro-events needing response or action – when will we get the time to sit, to think, and to process the inputs requiring more than immediate reactions? In one sense this is of course a personal question, relating to our own productivity habits and how mindfully we manage our time and our attention. That said it is also an issue that the technology industries need to address. If an ever-increasing number of notifications are not to overwhelm us, then automated filtering and prioritisation become critical problems to solve. Then we also need to consider whose algorithms do the filtering, and what do those entities gain from providing that service?
Adams poses these three questions:
- The first is will this new model operate at the level of the app, or notification or OS?
- Secondly, will these experiences occur in one consolidated stream, or in multiple streams?
- Thirdly, will that stream be owned at a company level, or as an open interoperable standard?
His point is that apps are changing; they are not siloed destinations anymore. Increasingly we engage with them across multiple surfaces: notifications, cards, and whatever comes next, etc. Designers ought to think of apps as publishing systems and not as a destinations.