SCENE: An upper-level manager at a large insurance firm is sitting at her desk. The organization’s underlying IT systems understands the following information: That she’s about to present some fresh numbers for a budget meeting. That, in order to move forward, she will need her colleague Larry in the room with her. That a PowerPoint presentation she used last year and which is saved on her personal drive would have some useful slides to assist with today’s meeting. That, her calm outward demeanor aside, she’s a little anxious about this one.
EXECUTIVE: (clears throat, taps at her keyboard) . . .
Okay, so it wouldn’t make a very exciting movie. But a script like this may be the best way to understand the unrealized potential around context-based computing, a concept that has been called other things for years but keeps being touted by analysts and consulting firms as a transformational concept in enterprise IT.
In January, for example, context-based services took up the first chapter of the Accenture Technology Vision 2012, a 52-page report that also examined the impact of converging data architectures and social-driven IT. Accenture took pages to explain that context-based computing is not just about cool mobile applications, the use of social media or the cloud. It’s all that and much more, the report said, adding that successful implementation of context-based services in large organizations could reposition senior IT executives by their employers.
“CIOs and other IT leaders who get it—who grasp the importance of context-based connections—will be able to establish themselves, and their organizations, as strategic players,” the report said. “They will immediately be able to offer new levels of insight that will differentiate their organizations from competitors.”
In journalism schools, novice reporters are routinely taught the “five Ws” – questions that must be answered in order to turn an event or issue into a news story. These Ws stand for “who,” “what,” “where,” “when” and “why,” and most journalism teachers would throw in “how” as well. In a nutshell, context-based (or as it’s sometimes called, context-aware) computing is about answering the five Ws as they pertain to a particular employee, customer or partner, and providing something useful to them.
First described back in 1991 at the Palo Alto Research Center (PARC), context-aware computing has been discussed by many experts in ways that sometimes make it seem like a utopian goal. Just two years ago Intel vice president and chief technology officer Justin Rattner said 2010 would be the year it hit the mainstream. It hasn’t, possibly because most CIOs and their organizations have been focused on cost-cutting and maintaining legacy systems than highly intelligent user experiences. As more companies slowly begin to develop strategies mobile applications and social media, however, it may be worth using context-based computing as a sort of organizing principle to create the kind of IT initiative that improves the ability to achieve business goals.
Context in action
If nothing else, context-based computing allows IT’s big thinkers to dream up some entertaining scenarios. These range from pure novelties – a TV remote that recognizes who’s holding it and adjusts the viewing experience according to their taste in volume and brightness – to more utilitarian.
Nicholas Bayley, Accenture’s global lead for IT transformation, used the example of a travel company that uses Twitter to see what users are talking about upcoming vacations or business trips. Once it pulls in that data, it might be able to then offer a discount or other incentive to attract new customers.
“It’s about aggregating data quickly and presenting it both to the business and customers. You have to be able to put together both unstructured and structured data and to bridge that effectively,” he said.
Internally, context-based IT might assist a CEO or marketing manager who has an appointment to see a customer target or client and needs all the vital information at their fingertips. Instead of going to your version of “My Documents,” the system would pick out certain files, such as the minutes of a sales meeting or notes from similar clients. David Jacobson, director of emerging technologies at PricewaterhouseCoopers, prefers to call such context-based services anticipatory discovery.
“It can take hours of work to prepare for such a meeting. With discovery, we’ll develop to the point where you enter the name of the client and system recognizes that you are going to be the one who will be visiting this client and will be in your calendar and will then automatically look for the information you require,” he predicted.
Does this sound a bit like customer relationship management (CRM)? According to Jacobson, most CRM systems require the user to do the majority of the organizing up front, based on a pre-defined scheme. It’s also usually predicated on keyword search, whereas context-based computing is a little closer to artificial intelligence.
At Tribal DDB, an advertising agency based in Toronto, technology director Joe Dee sees a different kind of internal opportunity for context-based computing. Think of all the internal projects that require various people in an organization to help out. Defining who those people are, tracking them down, coordinating their schedules and ensuring they can manage a project within their existing workload can be highly complex. Dee envisions an organization whose IT systems allow employees to “swarm in and out” of projects as needed, no matter how distributed their workforce becomes. There are already some products out there to assist with this, said Dee, including a task manager called Trello that sends notifications to developers when they need to work on something. HipChat, an instant messaging program for groups, can let teams create virtual “project rooms” where members can sit in on discussions as they’re needed, then move on.
Context-aware computing could become more urgent as organizations discover they need to keep IT in pace with what internal and external customers demand, Dee said.
“When you look at IT and rolling out platforms or rolling out new approaches to the way things work, it’s already heavy. It’s not fluid and it’s not quick. We may be able to get some tools that will help us work faster,” he said. “It’s going to be that evolve-or-die point. You’ll have to be able to service the needs of a consumer that just wants more and faster, or they’re going to go somewhere else.”
Big challenges
There are third big barriers to context-based computing, and they will come as no surprise to any CIO who has tried to even approach it on a superficial level. One part has to do with the massive complexity in many organizations. Another is about making sure responding to context doesn’t come off as creepy – or criminal.
Accenture’s Bayley said few firms have properly grasped the scale of information management required to identify where context would be useful and how to organize it properly. “It’s not just about having access to big data, but access to structured and unstructured information. You’d better get the architecture right,” he said.
Bruce Stewart, a Toronto-based consultant who has worked for Gartner and the Meta Group, among other organizations, said there may not be enough balance in the way enterprises tackle the varied formats of formats of data in order to understand its context.
“Everyone trained on information in university is being trained for unstructured (data), because they’re looking at it as the electronic follow-on to paper. The reality of the business world is that most records are structured. They’re part of operational systems. That isn’t going to change,” he said. “There is a real disconnect. People who learn how to administer databases come from a very structured, traditional IT background. The unstructured types, meanwhile, don’t understand the database world.” Context-based computing may require more holistic skill sets than are common in companies today.
Even if they get this part right, Dee said organizations have to avoid any activity that could contravene privacy laws or make users feel they are being overly monitored. “Companies are going rogue in a land grab for personal data to inform the contextual services they’re trying to build,” he said. “Consumers are starting to become really aware of what that means and how it’s used. They know their rights.”
Jacobson, however, thinks such concerns might be overblown. “(It) does not necessarily compromise security, privacy and so forth. It could if it was badly used, but that was true of almost everything,” he said. “If you’re operating in your business you’re typically behind a firewall and all your colleagues are working based on codes of conduct. Incidents of compromising one’s privacy and security and are certainly not non-existent, but they occur far less because of those guidelines.”
The third big hurdle may be the most common. “One of the biggest challenges is recognizing context as a phenomenon and how it correlates to the business,” Gartner analyst William Clark said in an online video posted last year. In several research reports on the subject, Clark urges CIOs to start talking about it with senior management now, so that initiatives related to improving context will be on the radar during corporate planning sessions.
Stewart said such initiatives will require major changes in organizations, ones that may prove highly disruptive to technology professionals.
“The typical IT organization remains very process-centric, almost process-myopic. If I can’t fit it into a standardized business process, I don’t know what to do with it,” he said. “A process-centred mindset won’t get the results to business value.”
Not that IT organizations can’t improve in this area, of course. Hopefully that comment hasn’t been taken out of context.
Next steps
The conventional wisdom around context-based computing is to figure out your data sources, begin experimenting with services and refine as you go. But that’s fairly general. Try these steps instead.
Keep a “bottleneck diary.” Context-based services should ideally help users get more quickly to an outcome than they would without contextual information. Jot down any instances you experience or are told about where users can’t complete a task until they get more information. And document where those sources of information are – their desktop, a database or a person. These are the sources you’ll be using to draw context out.
Be your own pilot project. The best way to understand how context can be useful is to look at your own gaps. Talk to staff or colleagues about the many simple, free online tools that help organize their lives and see how their functions might be further developed and integrated into your organization’s own systems. You’ll be more convincing talking to business leaders when you speak from personal experience.
Map the “five Ws” to a decision. To help your team (and yourself) figure out how to bring more context to enterprise processes, look at a recent historical decision or action, like signing a new deal with a customer or organizing an internal project. Identify the “who,” “what,” “why,” “where” and “when” and figure out how that information might have been more seamlessly presented to decision-makers (and don’t forget “how.”).
Talk ‘customer service,’ not context. Unlike the cloud, big data and other tech terms, context-based computing hasn’t caught on as a buzzword. That’s okay. Better to look at how the concept could be used to improve areas that are readily understood by business stakeholders like customer service, whether they’re internal customers like employees or customers who generate revenue.
Pick metrics that matter. Context could lead to many things – faster time to value, greater value offered to customers or internally, less duplication of effort. Some will be more concrete than others, but try to pinpoint these as early as possible as your experiments begin.
