Losing Our Heads — Separating Information from Interface
The way people use information services has undergone a major shift in the past few years, with destination portals like Yahoo! diminishing in popularity, overtaken by single-purpose mobile or web apps, such as those for weather or news. This shift represents the ‘unbundling’ of services from pages, and is being pushed even further in modern mobile devices, where we can see information being unbundled from the user interface itself.
New devices and interface models aim to give us access to the information and services we want, when we need them the most, and without having to open an app or website. If your train is delayed, it’s better you’re told before you leave the house. The most obvious way that this is manifested is in the mobile app notifications that most of us use every day, but there are other interaction models that show the path to a future where information is divorced from its source.
This new mode of interaction is sometimes referred to using an old computing term: headless. Traditionally this means an application or service has no interface exposed to the user — that’s not quite true at the moment, but there is a clear trend in that direction.
Fast Web is destination-based. Slow Web is interaction-based. Fast Web is built around homepages, inboxes, and dashboards. Slow Web is built around timely notifications.
Notifications can be sent for many use cases, from the trivial to the critical. At a recent ‘notifications summit’ hosted by the startup studio betaworks, attendees defined a rough spectrum of effective notifications: from content-based and contextually relevant (e.g. news alerts) at one end, to service-based and actionable (e.g. location check-in) at the other.
In the past few years the two major mobile operating systems, Android and iOS, have introduced interactive notifications. With these a user can not only receive notifications but can also act on them — reply to a text message, for example, or like a post on Facebook — without opening an app at all.
In the Browser
Notifications have largely been available only for native mobile applications until now, but are now emerging in web browsers. Known as web notifications, they act like their native counterparts in alerting users to new events. But web notifications have a drawback: they’re only active as long as the web application is active in a browser tab. Once the tab closes, the application can no longer send notifications.
This is changing thanks to a recently developed technology called service workers. These essentially work as a proxy, acting on behalf of a web application outside of the browser: that is, a website can register this proxy to remain active on a user’s device, and operate even when the browser tab is closed. Service workers enable a whole new class of interactions for web browsers, but the most obvious and useful to the typical user will be push notifications. As the name suggests, these are pushed to a user at the time that’s most appropriate, even when the application tab or browser is closed.
So far these push notifications are only available to a fairly small number of users, but already some big publishers are on board. The most notable is Facebook; their users on Android devices can opt to receive notifications of relevant new posts, and open the site in their phone browser to read them. No app install is required; this matters because it’s becoming increasingly difficult to get users to install and continuously engage with apps, so the web remains an attractive option for broad reach with low-friction interaction.
Problems with Notifications
While notifications provide genuine value, there is a danger that they can be over-used. One study suggests that in some product categories — notably, social and media — some 60% of notifications are disabled by the user. This contrasts sharply with ride-sharing apps, where the opt-in rate is almost double, with almost 80% of users choosing to receive alerts. Users prefer notifications to be relevant, time-sensitive, and valuable. Steven Levy says that, in order to make notifications to work for everyone, we’ll need better automatic filtering:
We need a great artificial intelligence effort to comb through our information, assess the urgency and relevance, and use a deep knowledge of who we are and what we think is important to deliver the right notifications at the right time.
Android already provide some help in this direction with notification priority settings for both developers and users, and algorithmic control of notifications from the Google Now service.
The broad purpose of the first wave of smartwatches, from the Apple Watch to the Pebble to any of the Android Wear devices, is to provide headless interaction with apps. Android Wear in particular has two core functions: access to Google search, and outboarding the notification tray (indeed, Android 4.1 saw a redesign of the notification tray specifically with the later release of Android Wear in mind). Apple Watch has Glances, which provides summaries of information from your most-used apps, along with appropriate controls to interact with them.
The stated mission of smartwatches is to reduce the time a person spends looking at their phone. So you can see useful information, read and reply to messages, and perform key interactions with apps — the very definition of headless interaction.
Peek & Pop
In iOS9 Apple introduced a new mode of interaction. Called Peek & Pop, it allows users of devices with Force Touch screens (such as the iPhone 6S) to press lightly on an app icon to see a quick preview of the content contained within (increased pressure on the touch would open the app fully). An extension of Peek & Pop is Quick Actions, shortcuts to commonly performed actions in enabled apps. For example, the Maps app allows you to send your location to a friend, and Camera lets you jump straight to video mode.
Peek & Pop is a way to get the information you want, without the overhead of finding an app, opening it, and browsing for the information within. It’s not a fully headless interaction, but is guided by the same motivation: fast and low-friction access to information and actions.
From GUI to CUI
Taking the idea of headless services to its logical extension, one route to the future of interaction may be through no interface at all. The digital assistant is now an important part of every major platform, from Siri, Cortana and Google Voice Search on our devices (and in our cars) to Amazon’s Echo and the forthcoming Jibo in our homes. These assistants perform queries and actions using commands stated in natural conversational language, and become more powerful and capable with every software update.
The biggest opportunity for services and brand interaction will be in messaging apps. Messaging is the fastest growing social sector; there are an estimated 2.5 billion people active users of messaging apps, with an extra 1.1 billion predicted to join in the next three years. Across Asia, most notably in China, the top messaging apps (largely QQChat and its follow-up, WeChat) are platforms in their own rights, giving users access to vast arrays of services including retail, medical and transport.
Interaction with these services uses the same natural, familiar interface as interacting with friends: written messages. Natural language text and voice interaction is known as the conversational user interface (CUI), and is set to dominate interactions in future. There are no graphical user interface (GUI) conventions to learn, you can just chat informally to an assistant that will carry out tasks on your behalf, and complete any business exchange using money from an associated digital wallet.
The fairly functional conversational apps and services of the West are somewhat behind the curve compared to their Asian counterparts, but that’s set to change with a new wave of services this year. From shopping to restaurant bookings, meeting organisation and travel planning, these exist only as contacts in our email or messaging address books, there’s very little dedicated graphical interface at all — to the point that some people refer to this as no UI (or, less often, contextual runtime).
In the West the big actors in this space include SnapChat, KIK and Telegram, but the biggest by far is Facebook. As well as owning WhatsApp, their Messenger app has 800m monthly active users and is currently being developed into a fully-fledged platform to match the big Asian rivals, with tools to help businesses meet (and sell to) customers. As Facebook observe in their trends for 2016: “threads are the new apps”.
Many messaging and conversational services currently rely on human operators to answer the queries, but the possibility of scale is offered by recent breakthroughs in artificial intelligence. As humans answer the user queries, the results will be delivered to learning machines which will begin to understand the demands of customers and, with systems now much better at understanding natural language, AI ‘bots’ will begin to take more responsibility over time.
These conversational bots will be the workers of the next generation of services. Facebook is beginning to trial them with select partners, and has its own human-augmented bot concierge service called M, which lives in Messenger. Meanwhile the rapidly-growing business-focused communications tool Slack has recently announced their own platform that makes it easy for any company to add or build their own services and bots, and companies like Howdy are already offering bots as a service.
Of course, the rise of the bots will have consequences for the way we work:
Our AI-powered assistants will manage more and more of our digital activities, eventually diminishing the importance of individual, siloed apps, and the app stores that sell them. Many websites could come to feel as outdated as GeoCities pages — and some companies might ditch them entirely. Nearly all of the information they provide can be fed into a bot and delivered via messaging apps.
Challenges for the Near Future
The rise of headless services and interactions won’t replace the GUI entirely; whether in the browser or a mobile application, some things are simply more easy to use and interact with using on-screen controls and displays. But it’s certain that fewer interactions will take place on a destination site or app; if your business is information or services, it will become more feasible to provide that to customers in the most convenient way: through a notification, a briefly-glanced screen on a smartwatch, or (more likely) in their messaging app of choice.
How do you design information for a large screen, tiny screen or no screen at all? What is the critical interaction of a service? If a service is accessed through a bot assistant, what personality should it have? It’s clear that new modes of interaction will require new ways of thinking, and new design disciplines.