An Inside View of IBM’s ‘Innovation Jam’

IBM Research is the world’s largest corporate research organization, with eight labs and 3,200 researchers in six countries. Every year Sam Palmisano, IBM Corp.’s chairman, visits its headquarters in Yorktown Heights, New York, to review progress.

When Palmisano toured the labs in early 2006, enthusiastic scientists showed him all manner of newly developed capabilities. One technology promised to forecast the weather so precisely that school districts could tell whether their town would get an inch or two more snow than their neighbors and therefore have to close school. Another project would enable the building of an Internet where shoppers could visit 3-D stores and see realistic 3-D demonstrations of products. Yet another new software program would perform realtime translation of speech so that the words on China Central Television or the Middle East’s Al Jazeera news network could appear in English underneath the speakers without human intervention.

After the demos, IBM’s Paul Horn, chief scientist, and Cathy Lasser, research chief information officer, met with Palmisano. “He was clearly very excited,” says Horn. But he was also already thinking about the next challenge — how to commercialize the breakthroughs successfully, a challenge IBM hadn’t always efficiently met. “He said, ‘Let’s come up with some novel way to get this stuff to the market more quickly. Let’s think out of the box.’ ” Palmisano felt that with 346,000 capable employees, there had to be faster ways to move products based on new technology to market.

The executives conceived the idea of a “Jam” to promote innovation. “Jam” was IBM’s term for a “massively parallel conference” online. IBM had developed its first in 2001 as a way to unite the organization. More and more employees were working at home or at client sites, rarely coming to IBM offices. The idea was that a Jam — a group of interlinked bulletin boards and related Web pages on IBM’s intranet, with systems for centrally managing everything and seeking substantive answers to important questions in three days or so — would give people a sense of participation and of being listened to, as well as generate valuable new ideas. From the beginning, the Jam process showed it could engage tens of thousands of people at a time. There were 52,000 posts in the 2001 Jam, addressing questions like “How do you work in an increasingly mobile organization?” and “How do we get IBM Consulting into the C-suite?” Subsequent Jams helped clarify IBM’s values and produced good ideas for improving IBM’s operations. A carefully designed system for reviewing huge numbers of posts enabled the company to initiate important courses of action.

But a Jam to accelerate the launch of new technologies was something different. IBM had only the vaguest of ideas how it would make money from most of the technologies that scientists had demonstrated. The challenges in creating new businesses substantial enough to benefit a company the size of IBM were radical. So in looking for ways to turn technology into profit, IBM was demanding more from the Jam system than it had before and seeking results more central to the company’s future. And if the proposed Jam involved as many people as Palmisano wanted, it would certainly be the largestever online effort to advance technological innovation.

The “Innovation Jam” took place in two three-day phases in 2006. We tracked the projects that received $100 million in funding based on the Jam’s results — the data shows that the Jam was successful to a considerable degree. It uncovered and solved problems in and mobilized support for substantial new ways of using IBM technology. It involved 150,000 IBM employees, family members, business partners, clients (from 67 companies) and university researchers. Participants Jammed from 104 countries, and conversations continued 24 hours a day.

But the Innovation Jam experience is important for the difficulties it demonstrated (and for how IBM struggled to overcome them) as much as for its successes. Ideas didn’t bubble up and get refined through continual, respectful dialogue. In fact, few contributors built constructively on each other’s postings. The Innovation Jam was organized to capture a huge number of ideas from IBM’s network, and it was purposely designed not to guide conversation artificially toward a quick focus on a few thoughts. But without organizers pushing toward an artificial consensus, conversations did not move toward consensus by themselves.

Rather than emerging during the online conversations, new visions emerged afterward. IBM had developed a carefully thought-out process that it used after each phase of the Jam to harvest ideas. Senior executives and others spent weeks sifting through tens of thousands of postings — gigabytes of often aimless Jam conversation. Working through the static enabled leaders to extract ideas they thought were key, put them together into coherent business concepts and link them with people who could make them work.

The Innovation Jam’s important lessons are thus highly paradoxical: On one hand, it showed how many people throughout an organizational network may have important strategic ideas. It demonstrated that online conversations and sophisticated technology can bring those ideas to bear on important societal problems and make them worth millions to a company. But the Innovation Jam also revealed limitations in how most people recognize and build on others’ ideas online. The result of these limitations is that analysts and managers near the top of the corporation were essential — together with sophisticated software for combing through vast amounts of verbiage — to making the rank and file’s ideas useful. Leaders found themselves taking a different role than in the past. Their new role was about identifying and nurturing a good idea as it was built on by the organization. But they were still the drivers of progress.

This article explores this innovation effort, unique in size and unusual in the amount of management resources invested in it. The article is based on participant observation in the Jam itself, review of Jam Web pages and postings after its completion, online use of some of the emerging technologies and more than 20 interviews with Jam organizers, participants, idea sponsors, senior scientists, senior executives and others.

Worldwide Brainstorming

IBM’s intellectual capital portfolio is immense. IBM holds more patents than any other company in the world and each year extends its lead. Yet only a few radically new IBM technologies have moved smoothly from lab to customer and produced substantial new revenues. Like many companies, IBM often lacks clear mechanisms for fitting innovations into its already complex product portfolio. One executive expressed a common frustration within the company by saying, “At IBM, new products aren’t launched, they escape.” Innovations in services and software have kept IBM highly profitable. But its sales have grown more slowly than that of other technology businesses, and in 2007 IBM lost its position as the world’s largest technology company to Hewlett-Packard (which had acquired Compaq Computer and had been more successful than IBM in big — though not necessarily highly profitable — consumer markets).

Speeding products to market is a priority, and to make it happen, Jam organizers sought to lay out key emerging IBM technologies for participants. Web pages described 25 clusters of technologies in six broad groupings. The Web sites included digitally recorded minilectures from IBM experts on some technologies and demonstrations in Adobe Flash of others. Experts from the Research Division and other IBM units participated in online discussions as moderators to help participants understand the technologies and address customer needs.

Each IBM Jam has included new experimental elements. In the Innovation Jam, several of the experiments sought to help emerging ideas evolve into something more complete. The Innovation Jam was the first Jam organized in two phases: one in July, when the company posted information on key technologies and participants brainstormed new ways to use them, and a second in September, in which participants refined ideas from the first phase. In Phase Two, participants were able to click to a separate site where they could work on business plans for key ideas using wikis.

Not everything worked ideally. Many of the participants logged on just to look around. But participants posted more than 46,000 ideas. They enthusiastically offered many potential money-making suggestions. This, from a participant in India, is fairly typical:

We go places and capture lots of pictures with digital cameras and camcorders, but there is a limitation in storage when it comes to video. There should be some way to upload videos to a Web site and there should be a device embedded in the camcorder which transfers and stores data to a remote server.

An IBM press release described the Innovation Jam as “the largest online brainstorming session ever,” and even with 150,000 participants the Jam managed to take advantage of many of the strengths of brainstorming sessions. People could raise their ideas freely, and the management of the Jam was based on the concept that “every idea counts.” However, the Jam also shared many of the difficulties common to large brainstorming sessions, as well as having serious difficulties peculiar to its online format.

Naturally, the brainstorming approach produced many ideas that were completely impractical or irrelevant to IBM’s businesses: a solar-powered toilet and vending machines selling flaxseed, for example. Monitors, however, found that guiding the conversations was even more difficult than in traditional brainstorming sessions. A senior scientist spent considerable time explaining why it would be impractical to import water from comets in outer space for consumption on Earth, and still failed to engage the advocates of the idea in discussions of ideas that could advance IBM’s business.

Mark Dean, a vice president at IBM’s Almaden Research Center, served as a moderator. He noted that the freedom of the Jam made it difficult to exert influence. In Phase Two, Dean was assigned to moderate a group of discussions gathered under the heading “Going Places” — ideas related to travel and to virtual communities. He said:

I was assigned to moderate “Going Places” because I was good at moving ideas forward in groups, had experience with my own blog, I was willing and they knew I liked cars. … In a face-to-face meeting you’d have an easel where you could write down ideas both to disassociate people from the ideas and as reminders for the specific topic. That wasn’t available here and it was difficult keeping people on track.

Many of the skills that made Dean successful face-to-face weren’t applicable: “When typing, you need to be able to write crisp responses. The body language skills you need in front of groups don’t apply.” Moreover, he inevitably had to leave the conversation for long periods. “You came back after eight hours of sleep and couldn’t tell where ideas had come from,” he said.

These problems were particularly notable in Phase Two, devoted to “refining” ideas from the first phase. Groups of managers and professionals had carefully sifted through the posts from Phase One and come up with 31 “Big Ideas.” Phase Two participants were asked to indicate which ideas they thought were best and to propose and discuss refinements. Yet even with wikis provided for work on rough-draft business plans, it was rare to find suggestions that built on previously posted ideas.

On the other hand, executives found that none of the major ideas from the Jam were completely original. People who had really important ideas had already spoken of them to some IBM managers.

The Jam’s value, however, was in bringing many ideas together. It helped IBM listen both to big ideas that had already been mentioned (but that no one quite knew what to do with) and to smaller ideas that in many cases complemented the big ones or helped executives think about how innovations already under development could succeed.

Capturing the Ideas

The brainstorming created a vast pool of thoughts from which related concepts could later be plucked and combined. The methodology that IBM utilized to find and make use of the most valuable comments started with sophisticated text mining software that the company first developed in the late 1990s. It also demanded a great deal of management time.

Text mining software examines the words in ordinary sentences and then classifies items into categories. Though it requires powerful computers, the process is fundamentally simple. Classifier software looks for items that contain the same or synonymous words and puts them together. When the classifier reviewed Jam postings, it noted that many contained the word “health” or synonyms such as “wellness” together with “billing” or “invoice” or “payment.” It could then create a category with all “health payment”-related postings. Executives interested in health care finance could later review all these postings at once.

However, the computer creates many spurious categories. It might find a large number of postings with the word “meeting” in them, for instance, and create a category. But “meeting” could be incidental to the authors’ meanings. Because of the errors, humans have to review the clusters and discard many. (IBM has explored using artificial intelligence techniques to reduce these errors, but they did not improve accuracy.) Human and machine together can effectively cluster many related postings, and they did that for each phase of the Innovation Jam. In addition, volunteers read each posting after each phase of the Jam and highlighted some that seemed of special interest. Only after both computer and human processing could senior executives review the Jam’s output to seek its important ideas.

Some 50 senior executives and professionals spent a week reviewing the output of Phase One to synthesize the 31 “Big Ideas,” and then an overlapping group of 50 reviewed the output of Phase Two to consider which ideas could become worthwhile businesses and how they should move ahead. For each phase, the reviewers worked in nine subgroups of five to eight people, with each subgroup focusing on a related group of ideas (health care, environmental, Internet related, and so forth). Participants came to Yorktown Heights, New York, from as far away as India for the reviews.

After Phase Two, the senior management groups were ready to propose substantial new businesses IBM could enter. IBM professionals had already been advocating some of these — a business to support the creation of a “3-D Internet,” for example. Other new businesses were aggregations of related ideas discussed during the Jam — businesses that could never have emerged so quickly without it. One example was a unit called “Big Green,” aimed at creating environmentally oriented businesses and bringing technology to bear on management of water resources. By November, Palmisano could announce appropriation of $100 million for launch of 10 new businesses.

Two Kinds of Progress

The Jam had achieved at least two positive results. First, it had enabled people with big ideas to articulate them to a wider audience, including skeptics, to hear others’ complementary ideas and to win funding. Second, and probably more important, it had allowed people whose ideas weren’t quite so big and who hadn’t been able to find the place for their ideas within IBM to present them in ways that senior people could understand. Related ideas could then be combined in major new initiatives. The 3-D Internet and Big Green businesses represent these types of achievement.

3-D Internet The 3-D Internet business unit was based on the ideas of a self-organized group of revolutionaries. Many IBM staff had long believed that emerging technologies could transform the Internet, allowing people to experience shared, simulated worlds in three dimensions. They thought these new simulated worlds would create enormous opportunities for commerce. Many virtual-world supporters had come together on a virtual “island” a few weeks before the Jam in the online virtual world Second Life (a simulated universe with many of the features that innovators want to bring to the mainstream, but a universe whose features are difficult to use). Cartoon-like avatars of IBMers from all over the world, meeting in Second Life, created the IBM Virtual Universe Community, and even Palmisano joined the conversations. (You could recognize his avatar right away: While most avatars are funky or outrageous, Palmisano’s was a cartoon man wearing a conservative blue suit, the kind for which IBM salesmen were once famous.)

Yet it was still not clear how IBM could make money through the 3-D Internet. During the Jam, virtual reality fans at IBM held conversations among their avatars within Second Life in what they called “the world’s first 3-D Jam.” There they discussed the “Big Idea” — “Real Markets for Virtual Worlds.” They also contributed their ideas to the conversation on the regular Jam site.

These Jammers contributed many ideas about how IBM could make money. After the Jam, IBM launched a business unit to provide tools for companies seeking to build the 3-D Internet. Executives and 3-D Internet evangelists agreed on the goal: to sell technology that would provide “faster response time for realworld-like interactions.”

Big Green While supporters of the 3-D Internet were well organized, most people with ideas were not. Most did not know how to fit their ideas into any IBM business, and none could promise their ideas would create the volume of business necessary to justify an entirely new IBM business unit.

But intelligently combined, many of these ideas could promise great things. A considerable number of ideas concerned human beings’ interactions with the natural environment. Senior managers reviewing the postings noted this after the Jam’s first phase and grouped seven “Big Ideas” for Phase Two under the heading “A Better Planet.” One of these “Big Ideas” was “Predictive Water Management.” It combined smaller ideas, including accurate weather forecasting for public utilities, sensing devices to help utilities know exactly how much water they have (and where), long-term climate forecasting and simulations of the impact of specific weather events on water supplies. No one in IBM had previously discussed water management as a potentially large IBM business, yet the Jam was revealing that the company could make a major contribution.

Jam participants pointed out that the water industry, though facing crises of supply, was failing to use important technology. “It is very archaic,” says Gary Rancourt, a member of the executive team that reviewed energy- and water-related postings. “It is perhaps the least IT-enabled asset of importance in the world.” After Phase Two, senior executives organized the business unit called Big Green Innovations to address predictive water management and other environmental opportunities. The team also found enough good ideas to support creation of a new business unit that would promote “green” data centers.

In addition to providing ideas for these new businesses, the Jam provided many of the people. The team for the Big Green business unit came largely from those who articulated environmentally related ideas during the Jam.

A Global Exchange on Ideas Being Born

The Jam created whole new businesses and allowed people from all over the world to give input on ideas that management was already working on.

Joseph Jasinski, program director for Healthcare and Life Sciences Research at the IBM Thomas J. Watson Research Center, served in subgroups that reviewed health care-related postings. He found the power of the ideas was in how they could complement each other and complement existing schemes. He “didn’t see anything that just blew [his] mind” among the postings, but said that many were “good augmentation” to ideas already being considered. For example, IBM had long wanted to help improve health care management. Many postings offered suggestions.

IBM leaders had hoped to create standardized personal health records that would comprehensively describe a patient’s health, to be accessed securely by each health care provider that the individual visited. The personal health records would give each provider complete background on each patient from the first encounter. IBM’s biometric authentication and data standards technologies could be central to such a system. However, technology was not enough. A health record business would need to appeal to an enormous range of stakeholders.

Like almost everyone in the world, IBM employees and customers are major stakeholders in health care systems. The Jam gave them a chance to articulate ideas. IBM management had been wondering whether a personal health record should be controlled by medical professionals who would have training in what should be entered into it, or if it should be open for viewing and change by the person whose medical history was chronicled. From countries all over the world, Jammers indicated that they and their friends would want to see and be able to change their own records.

On the other hand, some ideas about how a personal health care record should work varied dramatically by country. Americans wanted a health record system to reduce medical inefficiency. The Chinese and Indians wanted it to fix flaws in their own medical systems that kept people from getting adequate care. But the Chinese and Indians also differed between themselves: The Chinese cared about tracking traditional Chinese medicine, while Indians cared about tracking nutrition.

“In the Innovation Jam, we took some ideas that we were sort of messing around with, used the Jam to get others’ thoughts and then used the post-Jam process to launch pilots,” Jasinski says. IBM launched both a personal health records business and a business to provide smart health care payment systems. Other health care ideas were reviewed and rejected because Jam postings did not indicate they could add up to real businesses. For instance, participants suggested using technology to keep track of elderly people in their homes using such devices as heart monitors and TV cameras. But the suggestions did not indicate how the project would build enough revenue to justify it. “It was hard to see how we could build a business case,” says Jasinski.

All 10 businesses funded as a result of the Jam began functioning — assembling their product offerings and test-marketing them — in early 2007. The evidence indicates that at least some will be substantial successes. The Big Green unit has developed a comprehensive approach to water management for utilities and is piloting it. Success of other green initiatives led to expansion of Big Green into the “broadest initiative ever undertaken by IBM,” according to its Web site announcing the initiative, reallocating a billion dollars a year to help IBM and its customers address environmental challenges.

A unit aimed at building “intelligence” into electric utility systems made enough progress that its products “graduated” from test-marketing and became a major offering of IBM’s public-utility information systems salespeople. The Integrated Mass Transit Information System unit won contracts in London, Singapore, Dubai and Queensland, Australia.

There have been setbacks, too. After a year or so of effort, IBM decided not to market its electronic health records yet, concluding that the market was not ready to accept them.

Extracting and using the good ideas of tens of thousands of people is not simple, but it is potentially powerful. On the whole, the units produced by the Jam seem to be achieving enough success to have made the effort worthwhile.

The Jam represents a way for leaders to fulfill their role in managing innovation. “Idea generation is in some ways the ‘easy’ part — and darling star child — of innovation, whereas advancing, refining and building support for those ideas is the really tough part,” says Edward Bevan, who played a key role in running the Jam as vice president for communications at IBM Research and is now IBM vice president for technology and innovation programs. “The online portion of the Jam is the rather large tip of an iceberg.”

When questioned about whether any of the difficulties from the first Jam had provided lessons that would apply to the next, Bevan jokingly responded, “Lessons? The first one was perfect!” And it clearly did generate huge numbers of ideas, enable leaders to choose from and combine them, and bring selected ideas to market with deliberate speed. Bevan’s unit is organizing a new Innovation Jam for October 2008, and the approaches being used for the 2008 Jam show that IBM sees it as a product that will always require further innovation. Like each previous one, the 2008 Jam will involve new, experimental approaches. The limited degree to which Phase Two participants in the 2006 Jam focused on refining existing ideas has led to the return to a single 72-hour online conversation. Instead of building from IBM’s technology this time, IBM will start with customer needs. The Web pages from which people build ideas will be created based on IBM’s latest Global CEO Study, a report based on a survey of more than 1,000 chief executives and other leaders worldwide.

It is important, particularly for an organization trying to adopt lessons from the Jam, to think of these changes as part of the purposely built iterative process of the Jam and of innovation itself. Exactly how far these changes and the Innovation Jam can drive IBM (and others) and how thoroughly it can put underexploited technologies to use is just beginning to be seen. But given the energy and resources it takes to put on an event of this magnitude, it is safe to say that IBM finds significant value in the endeavor.