When it comes to solving pressing business problems, the conventional wisdom is that two heads are better than one. With the advent of collective-intelligence tools, enterprises are realizing that thousands of heads are even better still.
Take Pfizer and AT&T, for example.
One of the largest pharmaceutical companies in the world, Pfizer Inc. knows that the best solutions to its business problems don't always come from the researchers on the front lines, says Rob Spencer, a senior research fellow at the New York-based company.
Often someone in another department or another country could hold the missing piece to a particular puzzle, he says. That's why Pfizer wanted to figure out how to tap into the collective intelligence of its 86,000 employees to address its business challenges, says Spencer.
To do that, Pfizer turned to Idea Central, a tool built on IBM Lotus Domino and developed by Boston-based Imaginatik PLC. Imaginatik customized its Idea Central for Pfizer, which then dubbed it the Pfizer Idea Farm.
The software-as-a-service platform gives Pfizer employees a vehicle for submitting ideas for new products or process improvements, according to Spencer, noting that it has saved the company $20 million while helping to solve hundreds of business problems.
For its part, AT&T is using an "innovation management" platform to provide a forum for its employees to share information and ideas on how to improve products and services, according to Patrick Asher, innovation leader at AT&T. The forum is open to all managers globally, and the company is starting to build prototypes based on some of the ideas that have been proposed, but "we're not ready to talk about those yet," Asher says.
The system features innovation software from Pleasanton, Calif.-based Spigit running on AT&T's own infrastructure, Asher says. Currently the program is open to the telecommunications company's 120,000 management personnel, but AT&T is on the verge of opening it up to non-management staffers as well.
So, what exactly is this thing called collective intelligence?
"The definition we like to use is 'people and computers connected in ways that seem intelligent,'" says Rob Laubacher, acting executive director of the Cambridge, Mass.-based MIT Center for Collective Intelligence, which brings together faculty from across MIT to conduct research on how new communications technologies are changing the way people work together.
The MIT researchers are trying to understand how to take advantage of collective intelligence to use it for things such as organizational effectiveness, organizational productivity, profitability and teamwork, he says.
"Our basic research question is: How can people and computers be connected so that --collectively -- they act more intelligently than any individuals, groups or computers have ever done before?" Laubacher adds.
One of the center's major findings is described in a paper that's scheduled to be published in the March 2010 issue of the MIT Sloan Management Review.
For the paper, researchers at the center gathered nearly 250 examples of Web-enabled collective intelligence, including Google, Wikipedia and Threadless, a unit of SkinnyCorp LLC that harnesses the brainpower and creativity of a community of over 500,000 people to design and select T-shirts that are then sold on the Web. The striking thing about the collection of ventures studied is its diversity, the paper says, noting that the examples exhibit a wildly varied array of purposes and methods.
But after studying the various initiatives in depth, the researchers identified a relatively small set of building blocks, or what they call "genes," that are combined and recombined in various ways in different collective-intelligence systems, according to the paper. For example, reliance on the "crowd gene" is a central feature of Web-enabled collective-intelligence systems, according to the paper. In fact, all of the examples the researchers studied included at least one instance of the crowd gene -- they all involve at least one task that anyone is welcome to participate in, according to the paper.
"There is still much work to be done to identify all the different genes for collective intelligence, the conditions under which these genes are useful, and the constraints governing how they can be combined. But we believe the genetic framework described here provides a useful start," the researchers say in the paper.
With this framework, managers can do more than just look at examples and hope for inspiration. Instead, for each key activity to be performed, managers can systematically consider many possible combinations of ways to generate new ideas.
Although this approach doesn't guarantee the development of brilliant new ideas, MIT researchers say it does increase the chances that organizations can begin to take advantage of the collaborative possibilities already demonstrated by systems like Google and Wikipedia.
Related to crowdsourcing
Collective intelligence is linked to crowdsourcing -- the idea that you can gain more wisdom from crowds of people than you can from one person or small groups of people, according to Chris Andrews, an analyst at Forrester Research Inc. in Cambridge, Mass.
Crowd-sourcing at Sun & Ski
For the past year, Sun & Ski Sports has been using the Product Recommendations engine from San Jose-based Baynote Inc. The system makes recommendations by tracking the behavior of all the consumers who visit the outdoor sporting goods retailer's Web site.
Unlike Amazon's recommendation engine, which mostly relies on the past behaviors of users to make recommendations, Baynote observes users' behaviors on a site in real time, says Baynote CEO Jack Jia. Then it looks for patterns based on such things as how much time shoppers spend on particular product pages, whether they scroll up or down or through the whole page, or whether they highlight any text, click on any links or add any products to their shopping carts. Baynote then uses collective intelligence and an affinity engine to analyze the data, says Jia. The system then suggests products to the user.
"Our challenge was that we're not Amazon, so we don't get the repeat shopping over and over so it's hard to build a profile on someone to recommend products. It takes too long to build a profile that way," says Scott Blair, director of e-commerce at Sun & Ski. "So collective intelligence was the better option for being able to offer up quality recommendations to our visitors."
Since the Baynote system went live, Sun & Ski's revenue has increased by 13%, says Blair, who declined to give exact dollar figures. And 50% of the company's online sales now come from customers who click on items recommended by Baynote, he says.
But there's also another definition of collective intelligence.
"Collective intelligence is collecting information about what lots of people are doing and using that information to help produce better decisions for your interactions with customers," says Susan Aldrich, an analyst at the Boston-based Patricia Seybold Group.
Houston-based Sun & Ski Sports is harnessing the collective intelligence of its customers to provide good product recommendations for first-time visitors to its Web site, says Scott Blair, director of e-commerce at the outdoor sporting goods retailer. (See sidebar, right.)
The concept of collective intelligence was foreign to people even just a decade ago, says Forrester's Andrews. But the Internet has made the concept much more accessible and much easier to apply, he adds.
"The market for innovation management tools is still developing and is therefore amorphous, but we expect this market to steadily evolve and mature -- driven by strong corporate demand," says another Forrester analyst, Chris Townsend, in a report. In his paper, Townsend quotes a 2008 IBM report that says 93% of 1,130 senior business executives from around the world cited innovation as a top strategic priority.
Building on the ideas of others
Pfizer's Spencer says Imaginatik's Web-based collective-intelligence application lets employees build on the ideas of others and then allows the company to review, structure and track those ideas. The Idea Farm helps Pfizer find additional applications for drugs that are already in use, Spencer says. It lets the company pull together the right resources to rapidly and efficiently solve problems and uncover opportunities so it can have an edge over its competitors, he says.
Spencer says Pfizer has been using the Idea Farm for about four years to solve business problems, which typically come to him from heads of various business units.
After a 20-minute conversation with a business unit leader, Spencer says he can determine whether to pursue a problem based on its potential value to the company, how passionate the person is about finding an answer, whether that person has the resources to manage the problem and whether or not the problem could benefit from the large-scale focus the Idea Farm supports.
When he decides to tackle a particular problem, Spencer says the next step is to come up with the problem statement, which he calls the challenge. Then, through the Idea Farm, the challenge is e-mailed to a group of people invited to participate. Spencer says he has conducted challenges for groups of 200 to 20,000 people.
Participants then enter suggestions into the system via a preconfigured electronic form that's quick and easy to complete, Spencer says. Technically, the system needed to conduct a campaign can be completely set up and ready at scale in about 30 minutes, he says. However, the discussion about the business need that people are being asked to address is really the most important part of the process, and that could take as little as an hour -- if the business leader has a clear and urgent need -- or as long as a week, with several meetings, he says.
All participants can access the ideas that are submitted and comment on them. Imaginatik says that its software has over 50 features to help capture, manage, develop and evaluate the ideas. Idea Central is much more than a wiki, Imaginatik says, because the software is specialized to support the innovation and idea-management process, with security and configuration-management, for example, geared to this particular market.
Typically Pfizer's Challenges last for six weeks and end on a particular date with a decision about which idea will best solve the business problem, Spencer says.
"It is the business sponsor who sets the criteria for good ideas and then the path to implementation," Spenser says. "Serious business needs quickly get detailed and complex, so this is far more effective than a generic approach to innovation."
No cash rewards for employees
Unlike other companies, including AT&T, Pfizer doesn't offer cash rewards to employees who submit winning ideas because, Spencer says, tangible rewards, including money and prizes, are actually quite a bad idea. "They devalue an altruistic contract with the business and beget their own bureaucratic costs," he says. "Recognition, however, is always an excellent follow-up to a successful campaign."
Spencer says he has conducted 240 challenges over the past four years, and he does it as a no-cost internal service that benefits both R&D and the entire company. Since 2006, Spencer says the Pfizer Idea Farm has saved the $50.0 billion company about $20 million.
But it's not always just about the money, he adds. One Idea Farm project helped Pfizer shave three months off a 12-month process for setting up human clinical trials for a new drug, he says.
"So the ROI becomes fairly simple -- did that process help you deal with the problem faster, better, cheaper? The answer is nearly always yes," he says. "We cut three months off a 12-month process. A finance person could put a dollar figure on that."
Pfizer's implementation is an internal software-as-a-service setup, but it isn't cloud-based, Spencer says, explaining that the company has added extra security features.
"We own our own specific server, it is electrically isolated from any other computer with a full static installation of the necessary software on it," Spencer says. "We share no disk space with anyone else. In addition, Pfizer and Imaginatik jointly installed a VPN to this server and also jointly implemented a third-party single-sign on solution which both gives a better end-user experience and still another layer of authentication."
Innovation management at AT&T
"About a year ago, we wanted to re-implement how we innovate inside AT&T, so we went out and got Spigit software and created a big process called the Innovation Pipeline," Asher says. "On the front end, we're using crowdsourcing to get all the people across our company, not just the researchers, involved in innovating in a Web 2.0 environment. People are voting on the ideas and commenting on them, and we're taking the ideas that come to the top, funding them and turning them into products."
Asher says that each quarter AT&T asks employees to submit new ideas, and those ideas are then refined and ranked using Spigit's analytics engine. The top 10 ideas each quarter that can generate revenue or enhance the experiences of AT&T customers become candidates for internal funding, he says. Asher says the top 10 ideas are completely generated from the crowdsourcing model, and he has no control over the selection of those ideas.
And unlike at Pfizer, Asher says there are rewards for employees who submit the winning ideas.
"We have several ways to motivate people to be involved in the site," he explains. There are leader boards and rankings, so "that's a competitive motivator and those are placed prominently on the Web site," he says. "We give trophies to the people who have the highest-ranking ideas each quarter, and they're presented by our senior leadership at town hall meetings. If the idea actually gets funding, we give them cash rewards. And if the idea gets another round of funding to turn it into a product, they get another cash reward."
The Innovation Pipeline has been in operation for six months, and "we have 22,000 people [contributing] on our Web site, and that number's growing every month," Asher says.
The collaboration software is already yielding results. A researcher in AT&T's lab in Austin who was looking for information about a particular idea posted it to the Innovation Pipeline Web site. Within two hours, an employee who works in the company's network operations center in Dallas responded, saying he had written a report with all the supporting data the researcher needed, Asher explains.
"Normally it might have taken weeks for the researcher to find out about that report, if he found out about it at all," Asher says. "We're not ready to put a dollar figure on that, although we will, but we're really speeding velocity here. We're getting things done faster."
A former Computerworld reporter, Linda Rosencrance has written about technology for 10 years and has been a reporter for 20. Currently a freelance writer in Massachusetts, her fourth true crime book for Kensington Publishing Corp. will be out in October. She can be reached at email@example.com.