Do knowledge management incentives pay off?

When employees make up their minds that a knowledge management (KM) system is more trouble than it's worth, they simply stop using it. This decision affects the employer's bottom line and is a crucial factor in whether the big aspirations for a KM system pan out. Indeed, such investments frequently under-perform, which leaves companies wondering how they can get more value out of their knowledgebases. Govind Iyer and Sury Ravindran, associate professors of information systems at the W. P. Carey School, decided to look at the problem from a new angle — and found some surprising results.

Your printer stops working — what do you do? You could go to the manufacturer's web site where, in many cases you can access a vast storehouse of information that might help you fix your problem. At companies like HP and Microsoft, what you see when you type in your issue is often the very same thing that the employee on the technical support hotline sees.

If you try a few search strings and don't find the knowledgebase (KB) useful, you cast it aside, resolving not to bother with it again. "This is exactly what people who are working also feel when they are faced with a client issue or an internal issue where they have to solve a problem," says Govind Iyer, an associate professor of information systems at the W. P. Carey School.

When employees make up their minds that a knowledge management (KM) system is more trouble than it's worth, they stop using it, just as an individual consumer would. This decision affects the employer's bottom line and is a crucial factor in whether the big aspirations for a KM system pan out. Indeed, such investments frequently under-perform, which leaves companies wondering how they truly can make good use of the know-how that's locked inside their employees.

How to get more value out of knowledgebases then? The question intrigued Iyer and his colleague, Sury Ravindran, also an assistant professor of information systems. The two decided to look at the problem from a new angle and found some surprising results. They have presented papers on the topic at the Workshop on Information Technology and Systems, the Indian School of Business and America's Conference on Information Systems, and are currently in the process of preparing for journal publication.

Much of the research dedicated to the value of KM systems focuses on getting employees to contribute to these knowledgebases. Employees may be strong-armed into writing up their experiences after projects end or they may receive some sort of incentive to put their two cents into a database. Here the thinking is that the more "knowledge objects" — such as a step-by-step guide to fixing a paper jam — in the system, the better it will be and the more employees will use it.

This rationale is not wrong. In most cases, a KB with 1000 or 10,000 objects would be more of an asset than one with only a few dozen articles or tutorials. But rather than focus solely on increasing the girth of the knowledge storehouse, Ravindran and Iyer looked at what truly makes for a successful knowledgebase.

"One of the striking things when we read research in knowledge management is most of the focus is on contribution: We need to give incentives for people to share their knowledge. That's fine and good and companies are investing billions of dollars. But there's one critical element that seems to be missing in the argument and it is: Is there anybody who's using it?" asks Iyer.

A nuanced look at incentives

To better understand the issues, the pair surveyed 191 MBA students. In the surveys, the participants where given a scenario where they were assigned a job to be completed in a certain amount of time. In theory, the job could take any form, from something very lengthy and complex to something relatively short and straightforward like troubleshooting a malfunctioning printer.

The participants were told that they would receive their regular base salary as well as a bonus if they finished the job on schedule or early. In a second scenario, they were told that they'd receive the original salary and bonus as well as an additional bonus if they documented their experience and submitted it to the knowledgebase.

The third scenario saw the same base salary and ahead-of-schedule bonus but in this case employees would receive a bonus if they used the knowledgebase to do their job. In this final scenario, they were told that the employee who submitted the write-up they used would also receive a bonus. This last feature functioned like a fictitious click-through payment scheme whereby knowledgebase contributors could make money simply by documenting what they knew in a way that would make others value and use it.

Ravindran and Iyer also looked at people's tolerance for ambiguity and how that affects the way they use knowledgebases. Iyer, who is a chartered accountant and a CPA, says that research in the accounting field often looks at how individuals handle inexact situations, but this is not usually addressed in information systems research.

The introduction of this personality variable was relevant because one's job (which determines where and how a KB is used) often correlates with personality type; people with a lower tolerance for ambiguity gravitate towards more structured jobs while those with a higher tolerance for ambiguity tend towards less structure.

Someone working at a technical support call center who handles well defined, step-by-step troubleshooting would likely be less tolerant of ambiguity than a higher-level technician whose solutions must often be, by definition, less well defined. Similarly, there is a wide gulf between the exactitude of determining what's wrong with a printer versus the work undertaken by a consulting firm that might try to codify more nuanced projects, such as reforming a client's corporate culture.

Companies seem to understand this, and Ravindran and Iyer say that the differences are expressed in different types of knowledgebases. The high-level technician working on a printer might be better served by more freeform articles about electronics and circuitry that have less step-by-step to them so, in theory, this knowledgebase would look and feel very different from the one used by the front-line technicians.

The write-ups that result from a debriefing about how a consultancy changed its client's business would look altogether different from the other two. In many cases the same company may be best served by multiple knowledgebases targeted at different types of employee; HP's enterprise consulting team is not well served by the 1-2-3 tutorials used by the phone support reps even if both deal with printing and printers.

No one-size-fits-all

As there are different people with different personalities who use different types of databases for vastly different purposes, it is not surprising that Ravindran and Iyer found that these different personalities affect how a knowledgebase is born, grows and improves. Acknowledging this factor may help explain why previous research that did not consider this possibility produced mixed results.

When dealing with people with a low tolerance for ambiguity — the more concrete thinkers — Ravindran and Iyer noted that giving them incentives to contribute their knowledge does work. However, incentives to reuse the articles or tutorials that others wrote did not make them more inclined to contribute their own knowledge; when they took from the pot, they didn't feel they had to give back.

With this group, if the goal is to build a useful, comprehensive KB with lots of articles, it makes sense to offer incentives to get people to contribute their knowledge and to do so for a long period of time. With the other group — those with a high tolerance for ambiguity — giving them incentives to contribute their knowledge also proved effective in building the knowledgebase. Where this group differed was in how reusing the knowledgebase affected their actions.

Drawing on others' tips and tricks to solve problems didn't inspire the low tolerance group to contribute their own knowledge. The high tolerance group did, however, feel compelled to contribute more to the KB the more they used it themselves and benefited from others' contributions. "They are likely to contribute more when they perceive that their contribution is going to be used or is going to be useful," says Ravindran.

This, the researchers insist is an important distinction. Further, if this group feels their time is well spent they'll use the KB, but if they feel it's a waste, they'll ignore it even if there are incentives to use it. Therefore, when dealing with less concrete thinkers, a company should institute incentives for contributions to the KB in the initial stages of building the knowledge storehouse.

But keeping these incentives for the long-term doesn't help and will only cost the company unnecessarily because once these more free-thinking employees see that the KB is useful, the reuse itself will drive future contributions. All of this points to the professors' conclusion that the design and application of a KB will vary based on the use and users — as will effective start-up tactics.

Still to be determined are the specifics of what incentives work best and for how long, says Ravindran. He and Iyer are now planning to refine their research by interviewing actual users in real KM implementations where they are rewarded for reuse — a difficult task because they say most companies only offer incentives for contribution.

They also plan to simulate the use of KM incentives in a controlled setting. What they find going forward will affect specific recommendations in this part of knowledge management but they believe in their overall conclusion: "Come up with different plans for different types of people or different types of departments, different types of work, because you can't have a one-size-fits-all. It's not going to happen. It's not going to help anybody," says Ravindran.

Bottom Line:

• Employees judge knowledgebases based on their usability. If they come to the conclusion that a knowledgebase frequently will not be helpful, they will disregard it and try to invent their own solutions to problems. This reduces the anticipated value of a KB and costs a company money.
• Frequently, a KB perceived as useful and practical is one that has many contributions, much like a full set of encyclopedias in lieu of a single volume.
• Different personality types, job types and companies will require different types of knowledgebases. A one-size-fits-all approach will not work.
• Incentives work to get both concrete and non-concrete thinkers to contribute their know-how to a KB. People who are more comfortable with unstructured tasks are more likely to contribute of their own volition — to "give back" — if they find the KB useful to them.
• Companies may be wasting their money if they continue to offer incentives for contributing to a KB when, in some cases, users will develop a habit of contributing when they see the utility of mutual contribution.