Kiran K. Karthikeyan
Hyper context is the next evolution of contextual applications, when multiple bits of information are weaved together at the right moment in such a way that it is no longer seen as invasive, but useful. Hyper context enables serendipitous moments for the user, allowing them to perform tasks at their moment of need or immediacy.
Consider Google Now notifying you to leave early to make it on time for flight because of traffic conditions right now. To achieve this, Google uses the information about a flight booking you’ve made and if you’ve allowed it, your current location. From users using Google Maps on the route from your current location to the airport, Google knows how fast traffic is moving on average on that route. This allows estimation of the time it would take to get to the airport. So when Google realizes that you need to leave in 15 minutes, it sends you that serendipitous notification. If the notification also carries with it information about how long it would take to get a cab to pick up the user, you will most likely book a cab then and there to avoid missing your flight.
Now imagine two friends chatting about catching a movie premiering that weekend. A typical conversation would include mentions of when and where they are planning to go. A notification at this time to see movie listings for this weekend from the BookMyShow app would spur the user to view listings and perhaps book a movie that interests both of them.
There are few differences between the two scenarios, even though both are hyper contextual. In the first scenario, context is formed over time and the immediacy is manufactured based on multiple data points. The notification then at the right time creates intent in the user to book a cab. However, it also requires that any change in the information in context needs to be constantly monitored. If you cancelled the flight by calling the airline’s customer service number or through your company’s travel agent, Google Now would not know about it and send you the notification anyway. In the second scenario, intent is being expressed by the user and a way to fulfill it is made available within the context of the conversation they are having. To the user, the latter is more reliable and valuable resulting in constant use.
So why don’t we have more technologies focusing on expressed intent within context? Primarily because it is very difficult to capture intent at the moment it is expressed. Another reason is that intent is usually expressed vaguely or weakly. In our second scenario, the conversation could have also been about catching a specific movie on the opening day or just about watching a movie sometime next week.
So if an expression of intent is captured, it needs to be narrowed down. For most users, they narrow it by looking at options through search. However, searching for what you want requires you to express the intent to a search engine, a different task in itself which is usually disjointed from what the user was in the process of doing. Digital personal assistants like Siri, Cortana and Google Now referenced earlier are a step in this direction and make it easy for the user to express intent by just speaking into the air and use context to narrow options down. However, the behavior has not caught on.
For most of these tasks, users use some text input app that is either part of the operating system or specialized, more feature rich applications more tuned to their use downloaded from app stores. Features such as remembering commonly used words and phrases, auto correction, or even moving across multiple characters at once without lifting your finger off the touchscreen are now offered by most leading app developers in this space. Given that whenever a user is having a textual conversation about half of their screen is occupied by a keyboard means it presents the most immediate canvas to identify and serve user intent hyper contextually.
(The author is the associate vice president – Product Management and Product Marketing, KeyPoint Technologies Pvt. Limited)