When Performance Measures Go Horribly Wrong

They have a powerful influence. But unrealistic, unreachable goals can produce unethical behavior.

After the scandal erupted over the creation by Wells Fargo employees of more than 2 million bogus bank accounts, for which customers were charged over $2.5 million in unwarranted fees, the bank’s CEO claimed that it never wanted the accounts created, that it had fired 5,300 employees who were involved, and that he was “fully committed to … fix the issue and strengthen our culture.”

But here’s the problem: What happened at Wells Fargo wasn’t about culture. Nor was it about unethical employees, or about one senior executive who oversaw the program that led to these abuses (and walked away with an exit package worth over $100 million). No, this was a case of super-aggressive daily sales goals that were almost impossible to reach and where failure to reach them could lead to firings. As one expert on white-collar crime asserted, “This wouldn’t have happened without pressure from the top.”

We’ve seen this movie over and over in the public sector. Teachers and administrators in several public school systems have been caught correcting students answers on high-stakes tests, as a reaction to the enormous pressure they felt to improve student performance. In one of the worst cases, in 2011, an investigation found that 178 teachers and administrators in Atlanta’s schools had corrected student responses on standardized tests. Several of them went to jail.

The same kind of unethical behavior occurred on a wide scale at the U.S. Department of Veterans Affairs, again because of high-stakes goals that were virtually impossible to meet. Supervisors directed front-line staff to cook the books to make it appear that all vets received a medical appointment within 14 days of requesting it — the department’s all-important performance measure. The huge number of troops who had been injured in the wars in Iraq and Afghanistan made it impossible for the VA to treat all vets so quickly. But there was intense pressure to do so, and that led hundreds of department staff and supervisors to falsify appointment records.

The old saying, “what gets measured gets managed,” is true. Performance measures have a very powerful influence on people’s behaviors. And that raises the key question: How do we know if the measures we use are appropriate? Here’s an example of one person who answered that question in a creative way:

In 1995, an excellent public-sector leader named Bill Leighty was named director of the Virginia Retirement System. He took over a deeply troubled agency. It was in turmoil because of a major scandal, support in the state legislature was almost nonexistent, and employee morale was in the basement. This agency had some of the most rigid silos I’ve ever seen. Leighty knew he had to make major changes quickly to raise morale and performance.

One of his early steps was to create a “dumb rules” contest. He invited employees to identify rules, measures and regulations that were getting in their way and leading them to do stupid things. After reviewing dozens of employee recommendations, Leighty noted that “we found about a third of these were outside our control, a third were rules we had written ourselves, and a third weren’t rules at all!”

By the end of this exercise, the agency staff had been relieved of rules and performance measures that wasted their time and, even worse, led to behaviors that lowered performance and customer service. And Leighty actively involved the employees whenever they developed new measures to ensure that those measures were meaningful and had no unintended consequences.

So when you look at your organization’s performance measures, ask yourself and your team a few simple questions:

• Which of these measures provide staff with operational information that’s truly useful?
• Which of them simply waste people’s time?
• Which ones promote collaboration and information sharing among staff?
• Which lead to unanticipated consequences that are harmful to customers?
• And perhaps most important, are the measures achievable? Do they focus on factors that staff can actually control?

Albert Einstein once said that “not everything that counts is countable, and not everything that’s countable counts.” Do your organization’s performance measures count?

When the Best Intentions Lead to Disaster

The VA scandal has its roots in two related management failures. Government leaders everywhere should keep them in mind.

As the Department of Veterans Affairs begins the long, difficult and expensive process of addressing the problems that led to its scandal over falsified wait times for veterans seeking medical appointments, government managers who want to keep their own enterprises out of the same kind trouble would do well to look at the elements that brought the VA down.

At the heart of the VA scandal is the falsification of records in the face of a huge surge in veterans needing care and insufficient resources to serve them. One study by federal auditors found that, while official VA reports claimed that some vets waited 24 days for an appointment, the average wait time was actually 115 days. And there is another, equally troubling aspect to the scandal: the harsh reprisals to which VA workers who tried to report wrongdoing were subjected.

At the core, I think the VA’s problems were driven by two factors: the misuse of performance measures and an agency culture that didn’t respect candor.

On the performance measures problem, VA leaders made an understandable and common mistake. To get employees’ attention, they put a huge emphasis on a single measure: All vets will have an appointment at a health clinic within 14 days of their request. Given the mismatch of demand and resources, there was no way the staff could achieve that laudable goal. Nor did the goal make sense: Vets with serious problems should be seen immediately, while it’s reasonable for those seeking an annual checkup to wait several weeks or more. Further, many supervisors’ bonuses depended on the performance numbers. Put these factors together — one all-important but unreachable goal, with one’s pay riding on the outcome — and it was a disaster waiting to happen.

To some, the candor issue might have been surprising. After all, former Veterans Affairs secretary Eric Shinseki spent three weeks in the field each year holding dozens of meetings with employees and supervisors at which he insisted on candor. But the bad news never got to him. It’s not hard to see why. A VA scheduler had to decide whether to listen to the supervisor sitting down the hall who said to game the system or to the department’s top leader who meant well but was thousands of miles removed. Tragically, all too many followed their supervisors’ orders, and those who didn’t were often punished.

So how do you turn this situation around? Provide mandatory ethics training? Fire the supervisors who told staff to cook the books? Bring in a whole new team of leaders and managers?

Management consultants often say that “what gets measured gets managed.” But when you have a single, impossible-to-achieve, high stakes measure, nobody should be surprised when some people game the system. A far smarter approach is to use a version of “the balanced scorecard,” which captures data on four key performance areas: financials, customer satisfaction, internal operations and employee learning/growth. Many public and private organizations have used versions of the scorecard with good results. There is no one meaningful metric that captures all that matters in an agency. Adopting a few (emphasize few) measures is far more realistic and effective.

In addition to using a balanced set of measures, managers need to involve employees and supervisors in devising those measures. The measures need to be ambitious but not impossible; they need to focus on things that staff has the power to control. They need to give managers and supervisors (as well as external stakeholders) data that they find useful. The VA’s all-important metric — a health-care appointment within 14 days — failed each of these tests.

As for the second issue, the VA story demonstrates that candor is vital to achieving the mission. Shinseki, a public servant of great integrity, was sincerely interested in getting honest feedback from front-line staff, but staff didn’t feel it was safe to speak up. What can agency leaders and managers do to create an open, candid environment? Here’s a starter list:

• Model candor at the top. When leaders acknowledge mistakes that they or their agency have made, it sets the right tone.

• Talk about the reasons that candor is so critical. The point isn’t candor for its own sake (although that’s a good thing). It’s to continually spot problems and opportunities for improvement.

• Craft a simple narrative that highlights the cost of stifling candor. NASA’s Challenger tragedy wouldn’t have happened had NASA’s managers been open to the engineers who tried to warn them that it was too cold to launch the spacecraft. If you say, “We can’t have another Challenger disaster” at the space agency today, people quickly understand the message.

• Work closely with your middle managers; they are the key. When senior managers frequently meet with those in the middle, they should ask them what’s going well and what isn’t, and make it safe and rewarding to talk openly about problems. When middle managers experience the power of operating in an open environment, they let their guards down and realize that candor improves agency performance (as well as their own careers).

The Department of Veterans Affairs is in serious trouble. So are other public agencies that relentlessly track one high-stakes measure and that allow retaliation to replace candor. We should all take a lesson from the VA.

You may use or reference this story with attribution and a link to:
http://www.governing.com/columns/smart-mgmt/col-department-veterans-affairs-scandal-misuse-performance-measures-reprisals.html