It was the memo heard around the world: In late February, when Yahoo CEO Marissa Mayer ordered the company’s staffers to stop working from home, she set off a ferocious debate over workplace productivity. “Speed and quality are often sacrificed when we work from home,” wrote the company’s human resources chief in a leaked memo implying that telecommuting employees were less productive than those in the office.

Nonsense, shot back advocates for flexible arrangements: Without interruptions from co-workers or time wasted in traffic, telecommuters are often more productive than their in-office compatriots, not less.

The debate highlights a broader issue that goes well beyond the struggling Silicon Valley giant — the widespread disagreement and lack of clarity over what constitutes productivity in the modern workforce.

More than 50 years after management guru Peter Drucker first wrote about the difficulty of defining and measuring the productivity of knowledge workers, management experts say many companies still do a poor job of it. “In general, organizations have not truly come to grips with how to think about productivity in a knowledge economy, let alone how best to manage it,” says Jordan Cohen, a productivity expert with PA Consulting Group.

At its most basic, employee productivity is simply a measure of how efficiently he or she works: For each hour spent on the job, how much does the person produce?

That may be easy to calculate for assembly line workers in traditional manufacturing jobs once at the core of the economy. But it’s far from the case for the service jobs and knowledge work that dominate today’s economy. “When it comes to knowledge work, productivity is really hard to measure; it’s nowhere near as simple as the number of bushels a worker picked in an hour or the widgets produced,” says Lynn Wu, a Wharton operations and information management professor. “It simply doesn’t work to measure productivity by output.”

Quality is often as important, if not more so, than quantity: The sheer amount of time spent on a project or the output produced may have little to do with how productive an employee truly is.

Wu points out that it would make no sense to judge the productivity of a software developer by counting the lines of code produced; a shorter string of code might yield a far better solution than a longer one.

Despite the difficulties, companies often end up relying on crude numeric tools, notes Matthew Bidwell, a Wharton management professor. Some rank engineers based on the number of patents awarded, for example. But who is more productive, the engineer who files many low-value patents, or one with fewer patents which prove to be more useful?

Moreover, Bidwell adds, such measures rarely take into account what he calls “citizenship behaviors” — behind-the-scenes work such as recruiting, mentoring or communicating across teams. Such behaviors are critical in knowledge organizations, yet they are frequently overlooked in favor of hard data. “The risk is that we focus more on quantity over quality, just because we can measure it,” says Bidwell.

A Bonus System That Backfired

The answer, according to management experts, lies in creating broader and more nuanced measures of productivity that account for the quality, effectiveness and impact of an employee’s work, as well as the quantity. “You can’t just say more code is better,” notes Cohen. “Better code is better.”

Cohen suggests that managers must be rigorous in defining what behaviors and outcomes are most critical to enhancing productivity. And they must be equally diligent in determining how to measure those qualities over the medium to long term, not just immediately.

“How a company defines productivity will determine what infrastructure they build to measure and manage it. If they don’t really question the traditional assumptions around productivity, they end up with an industrial-era notion — simply that ‘more output with less input’ is better,” he states.

Cohen points to the experience of a recent client, a fast-growing fashion website. The manager sought help in improving the productivity of sales representatives who handle customer orders over the phone. “He thought that if he could shrink the time that reps spent with each customer a little bit, they would be more productive,” says Cohen.

Many customers wanted advice on the items they saw on the site, however, particularly the first few times they called. That interaction was key to building a strong relationship between the site and its customers. It also provided valuable insights as to what shoppers wanted.

Rather than cutting back the time each rep spent with a customer, Cohen suggested that the employees stay on the phone longer. “If you think about productivity in terms of effectiveness, that extra time was all about trust-building for the customer, at least the first few times they called,” he says. “Maybe after the fifth or sixth call, you would want the reps to shorten time on the phone, but not before that.”

For lawyers or engineers, that might mean including more subjective criteria when determining the impact of their work, for example. Measuring how often others cite an engineer’s patents or refer to a lawyer’s briefs, or ask to work with them on difficult projects, might give a fuller picture of effective productivity than the number of patents filed or hours billed alone.

Francesca Gino, a professor of business administration at the Harvard Business School and author of Sidetracked: Why Our Decisions Get Derailed and How We Can Stick to the Plan, warns of another danger: Managers often judge productivity by whether performance targets are achieved, without looking closely at how they are met.

One company Gino works with sought to boost the productivity of its sales people by introducing a system of bonuses for those who met monthly targets. Sales went up, bonuses were paid and everyone was happy — until the managers realized that returns were also soaring. The poorly designed system allowed the sales people to get their friends to buy merchandise, then return it the following month as soon as the employees pocketed their bonuses. “If you aren’t measuring the right thing and you have incentives in place, you might trigger behavior you don’t want,” says Gino. “Too much attention is paid to outcomes, and not enough to process.”

The Facebook Problem

Beyond measurement, another critical challenge is managing a workforce with varying levels of productivity. On any team, some employees will inevitably be more productive than others.

That was the issue Yahoo’s Mayer faced in addressing the firm’s stay-at-home workers. Internal data showed that many appeared to be less productive than those working in Yahoo’s offices, in part because many didn’t log in to its virtual private network (VPN) for much of the day. In the memo sent to employees, Yahoo’s human resources director also argued that increasing face-to-face collaboration would improve the company’s performance. “Some of the best decisions and insights come from hallway and cafeteria discussions, meeting new people and impromptu team meetings,” she wrote.

While few management experts doubt the value of face-to-face communications in fostering ideas and collaboration, many believe Yahoo erred in addressing its productivity problem with a one-size-fits-all solution. By penalizing both high and low productivity employees equally, the company risks losing the strong performers among its telecommuters, without getting to the bottom of why others’ productivity sagged.

“Blanket policies are a really bad idea to address performance issues” that involve only a few people, says Peter Cappelli, a Wharton management professor and director of the school’s Center for Human Resources. “The right answer is to deal with those individuals.”

Cappelli argues that managers should work directly with their low-productivity employees to determine the source of the difficulty and whether it can be resolved. “The biggest problem is a lack of supervision — setting out objectives and following up on their progress,” he notes.

In Yahoo’s case, that might have meant only ending telecommuting for those who didn’t log in to the VPN frequently. The laggards could have been more closely supervised from the office, while leaving alone those able to work productively from home.

To address Mayer’s other concern — her desire to improve Yahoo’s performance through greater employee collaboration — other, less draconian solutions might also have been found. Yahoo might have limited telecommuting to one or two days a week, Wu says. Or it might have urged managers to schedule regular on-campus meetings or brainstorming sessions to foster more of the creative interactions Mayer sought.

Wu points out that other companies have come up with more innovative solutions to develop that sort of collaboration and knowledge sharing. IBM has created an in-house database of expertise, for example, that allows consultants to tap into the knowledge of others throughout the company. The technology services giant has been able to leverage the expertise of its most experienced and productive employees far beyond their teams or personal networks. As a result, says Wu, IBM’s consultants saw their productivity grow 5%.

Many companies have also turned to blanket solutions to address another common productivity problem: the fear that employees are spending too much time on personal email, Facebook and other social networks.

Once again, experts suggest that overly broad policies — such as limiting Internet access for everyone — risk doing as much harm as good. “People don’t goof off on the ‘net if they’re engaged or are doing high-quality work,” notes Mark Ellwood, president of Pace Productivity, a consulting firm. “If you see abuses happening, the question is, ‘Why?'” Does the person’s job need to be restructured? Could they be given more challenging work? Do they simply need time management training? “Do some diagnostics — don’t just set a blanket policy,” he adds.

That’s particularly true given the importance of social media to many jobs in today’s economy. New research by Joe Nandhankumar, a professor of information systems at the Warwick Business School in the United Kingdom, demonstrates that workers encouraged to use Skype, Twitter and other social networks are among the most productive. He found that workers who used such sites could handle more calls from customers, and resolve their problems faster, than those who didn’t, thanks to their ability to tap social networks to quickly find new people and resources to address customers’ issues.

While some workers undoubtedly do waste too much time checking out Facebook at work, the lesson is clear: Teaching the low-productivity workers to better manage their time, while ensuring they have engaging work suited to their skills, will more effectively enhance productivity, says Ellwood.

The Productivity Punishment

Another difficult issue in managing a team with varying levels of productivity is balancing out the workload for those who boast better than average performance.

Consciously or not, these highly productive employees tend to become the “go-to” players: Managers know they can count on them to get a difficult or time-consuming project done. And as staffing levels shrink even as the workload at firms continues to grow, the impulse to pile more projects onto the most productive only increases. It’s a natural human tendency — but it can also leave the most productive people feeling burnt-out, warns Gino. Employees quickly come to feel as though they are being punished — by being given more work — rather than rewarded.

The risk for managers is that these high-productivity people will leave. To counter that risk, managers must improve the productivity of other team members so that they can pick up more slack. They must also monitor the top performers closely to ensure they don’t hit the tipping point.

“As a manager, you have to align the incentives for your high-productivity people — give them a bonus, or more responsibilities, so they have an incentive to stay,” notes Ellwood. “For a lot of knowledge workers, money is part of the equation, but the intrinsic desire to do something cool is just as important.” Assigning them projects that increase their status often works better than financial incentives, he adds.

It can be a difficult balance to maintain, especially when budgets remain tight and internal job movement is limited. But for companies intent on achieving optimal productivity across the workforce, managing highly productive employees carefully can be just as critical as improving the performance of those who lag behind.