Higher Standards

reg Woods will never forget the reaction when he first suggested to the Social Security Administration that it promise that everyone who called its 800 number would get through on the first try. Toni Lenane, then SSA's chief policy officer, minced no words: "That's insane!"

It was 1993, and Woods was the National Performance Review's new point man on strategies to improve customer service. He made the suggestion at a briefing he held to introduce the idea of customer service standards to a handful of large agencies.

When he recovered, Woods asked Lenane what she meant. "She said, 'You've got to understand, currently we can't answer anywhere near 100 percent of the calls,' " he remembers. "The idea that we would invest the money to add the system and operators to do this is just not feasible.' "

But Woods rarely takes no for an answer. He kept pushing, arguing that by using new technologies and learning from the private sector, the agency could provide a 100 percent response rate without hiring more staff. And in August, as the first NPR report went to bed, agency leaders finally agreed. In addition to three innocuous, unmeasurable policy statements promising courtesy, good information and referrals to other programs, the one on the 800 number stood out like a beacon: "You will reach us the first time you try on our 800 number."

Swallowing Hard

Lenane and her colleagues swallowed hard, then set to work gathering better data. What they discovered confirmed her initial reaction: It was impossible. At the busiest times of the year, half the calls got busy signals. And volume-already the highest of any 800 number in the world, with more than 60 million calls a year-just kept rising.

The agency had already considered contracting the phone service out to private firms and decided it wouldn't work. Fielding questions from anxious senior citizens about their retirement checks requires knowledge, patience and courtesy. SSA's teleservice representatives knew a tremendous amount and required significant training.

So the agency put together a team and began studying its dilemma. It launched a series of more detailed customer surveys, and when the team analyzed the new data, it was clear that Woods' instinct had been right: Endless busy signals on the 800 number damaged the agency's reputation with its customers. "Our data showed that access was the single biggest driver for customer satisfaction," says Lenane. "Getting through to the 800 number influenced the public's perception of our competency and knowledge and their overall satisfaction."

As they began work, they also negotiated to soften the standard. After lengthy discussions between the agency and NPR, they settled on a commitment that 95 percent of all callers would be served within five minutes. In 1995, Rep. John E. Porter, R-Ill., the new chairman of SSA's appropriations subcommittee, forced agency leaders to include that commitment in their Government Performance and Results Act (GPRA) goals.

Porter controlled the agency's purse strings, and money talks. His insistence "not only gave it teeth, but it made it a very, very clear and focused goal that everybody knew and everybody was shooting for," says Steve DeMarcos, an assistant regional commissioner in Social Security's mid-Atlantic region.

But with downsizing well under way, thanks to the 1994 decision by the Clinton administration and Congress to eliminate 273,000 federal jobs, the access rate was still declining. In fiscal 1995, only 73.5 percent of callers got through in five minutes, the lowest rate ever.

So after 18 months of study, including a benchmarking study of telephone service in the private sector, agency leaders finally bit the bullet. First, they agreed to replace teleservice representatives when they left, despite the downsizing effort. Then they converted two SSA Data Operations Centers, with about 700 people, to full-time telephone work.

SSA officials had already developed a new phone system with AT&T and had begun training employees who didn't ordinarily deal with phone calls to handle them when volume spiked. "We worked out a sophisticated call-routing system, with the ability to quickly bring these people on during peak calling periods," says Jack McHale, who was then in charge of the 800 service. "When calls fell off, they went back to their work."

Then, in January 1996, SSA expanded teleservice responsibilities to virtually every technical person who worked in a Program Service Center-back offices designed to process benefit applications, among other activities. They trained 3,700 people in six large processing centers, more than tripling their number of "spikers."

This required a long campaign to win the support of the American Federation of Government Employees, which represented affected workers. The key to success was job security: Automation was expected to do away with some of the work at the processing centers, so shifting to telephone service saved people's jobs. Once the union had agreed, SSA leaders and a union official spent two months visiting the centers that were going to have to change, talking with the employees, explaining the importance of customer service and working through the resistance.

Meanwhile, the agency restricted leave for teleservice representatives and spikers at peak times, accelerated the use of overtime and worked with employees to come up with changes in processes and rules that would improve performance. Because calls peaked after people received checks at the beginning of every month, it began staggering monthly payments for all new recipients.

The transition was hardly smooth. As the agency retrained thousands of people, customer service suffered-both in the processing centers and on the phones. November 1995 was the low point for 800-number access: Only 57.2 percent of callers got through within five minutes. Then, on the first morning back to work in January 1996, as the agency put all the new spikers on line to handle the post-holiday surge, AT&T's 800 system crashed. Hundreds of thousands of Social Security recipients called in, and for several days they all got busy signals. The media gave the agency a drubbing, and agency leaders lost a lot of credibility with their employees.

But once AT&T fixed the system, the numbers began to turn around. By February, the five-minute access rate was 92.1 percent, and the percentage stayed in the 80s and 90s for the rest of the fiscal year. In November 1996 it hit 95.9, and it has stayed above 95 percent every year since. (In 1997, the agency added a promise that 90 percent of all callers would get through on their first try.) Meanwhile, customer ratings of the courtesy and knowledge of teleservice reps remained more than 95 percent positive.

"There was a lot of internal celebration [when we met] the goal in 1997 and 1998," DeMarcos remembers. "When you do it one year, two years, three years, it really becomes a part of your culture. Internally, if you were to talk to all the people on the network. . . . They know the '95 and five' commitment. That's what drives us."

McHale had been lobbying for many of the changes for a long time. "We had to improve, because we were so bad," he says. "But I don't think we would have stretched as far as we did. As someone who was asked what we needed to do to meet this standard, I would come up with these recommendations, and I could always lean on the standard. I could always do calculations of how short we would be. Had the standard not existed, we would have improved, but not as much."

A Mixed Bag

The Social Security story demonstrates the power a good customer service standard can have to force agencies to do the unthinkable. SSA endured tremendous expense, dislocation, pain-and even failure-to meet its standard.

That's why President Clinton issued an executive order in 1993, at the NPR's recommendation, requiring all agencies that deal directly with the public to survey their customers and establish customer service standards. By fiscal year 1998, 570 agencies had published more than 4,000 standards, according to the NPR.

Predictably, they were a mixed bag. Some, like those at Social Security, were specific, measurable and meaningful to customers. But too many were vague and unmeasurable. ("We will maintain the facility in a clean, safe condition," "We will answer your questions or refer you to additional sources of assistance," and "We will display schedules of programs and activities available throughout the park"-all from the National Park Service-were fairly typical.) Performance was measured on only 2,800 of them, and few brought any consequences for success or failure.

So in 1998 the NPR shifted its strategy to focus on the real outcome it wanted: higher customer satisfaction. It commissioned the producers of the American Customer Satisfaction Index (ACSI) to survey customer groups selected by most of the 32 "high-impact agencies." It then compared their satisfaction levels with ACSI data on customers of similar private services and published the results.

NPR leaders also pushed the high-impact agencies to include customer satisfaction in their performance goals. And they worked with the President's Management Council to fashion a policy statement urging that all senior management performance bonuses and awards depend on success measured against a "balanced scorecard" that included measures in three areas: customer satisfaction, mission effectiveness and employee satisfaction.

"Instead of focusing on the development of customer service standards, we shifted our focus to ask, 'Are our customers satisfied?' " explains John Kamensky, longtime NPR deputy director. "We jumped straight to the outcome of customer satisfaction."

This was a wise move, and it has had a real impact. A quick review of performance goals at the high-impact agencies shows that customer satisfaction is now taken quite seriously in most. And the new PMC policy should make it more important to top managers.

But now that more agencies are taking customer satisfaction seriously, it may be time to reintroduce them to one of the most powerful tools they can use to increase it: customer service standards.

Lessons From Experience

If the next administration were to launch a renewed customer service standards initiative, how could it make the new effort more successful than the first? The experiences of government agencies with this tool-both successes and failures-in the United States, Great Britain and Canada suggest a number of lessons:

1. Involve customers in the creation of standards.
If you don't, you won't know what is important to them. "Never assume what people want," says SSA's DeMarcos. "Have some concrete reason and basis for what you're setting as a service goal. Find out from the customers. You can kill yourselves to do something that you find out people think is a yawner."

In both the United States and abroad, many organizations have neglected this step-and paid the price with worthless standards. "The place agencies have had the most trouble is the idea that they have to ask their customers what they want and whether they're getting it," says NPR's Woods. "They revert to the Washington mentality: 'We figure it out in D.C.' So they always get it wrong."

Customer surveys are useful, but face-to-face contact is even more important. "Get the people who are charged with doing the improvements to look at people they're serving square in the face and ask the questions," recommends Laurie Ohmann, who has helped several organizations develop standards. Ongoing customer councils are perhaps the best tool, though you can use many others-focus groups, panels, interviews and so on. You can even videotape the sessions and show them to the rest of your organization.

The obvious reason to do this is to generate standards that are meaningful to your customers. The less obvious reason is to persuade your employees to take them seriously. They have to be "connected to what the customer has actually told you," says Janice Warden, who led the successful effort at SSA, then moved on to the NPR. "You have to be able to point to that, to make it credible to the entire organization." Otherwise, she says, employees will say, "Are we doing this because it's important or because Vice President Gore told us to?"

2. Make your standards specific, measurable commitments, not vague statements.
If you can't measure it, you can't tell whether you're meeting it, and you can't reward success. And if you aren't specific, you will only feed the cynicism of your customers. As a guide published by the Canadian government says, " 'We guarantee' is better than 'We will,' which is better than 'We aim to.' "

Candace Kane, who succeeded Woods as head of the NPR customer service team, says she saw more than her share of the "We're going to be responsive to you" variety of standards. The agencies "discovered that responsiveness to me is different than responsiveness to you. You really want to get specific: 'We will resolve the problem within 24 hours.' "

3. Publicize your standards-and your results.
If people don't know about your service standards, they will have far less effect. The U.S. Postal Service, for example, publicized its first-class on-time delivery standards (three days within the continental United States, one day locally) and reported quarterly on its performance. The results generated front-page newspaper stories, creating useful urgency within top management. But the Postal Service was silent about another standard it committed to in 1993: "You will receive service at post office counters within five minutes." If you look hard next time you go into a post office, you may find a tiny, 4-inch-by-5-inch sign announcing the standard. But you've probably never noticed it. As a result, it meant nothing to you, the customer. Nor did it seem to have any impact on employee behavior. It was, sadly, a wasted opportunity to win over the public.

4. Make service standards consequential, by creating rewards for fulfilling them and penalties for failing.
Consequences give this tool its teeth; they lend urgency to the often difficult challenge of meeting the standards. At Social Security, the leadership understood that service on its 800 number was critical to customer satisfaction, but answering every call within five minutes-or even 95 percent of calls within five minutes-seemed impossible. If Woods had not used Gore's authority to cajole the agency into publishing a standard and reporting its performance every year, it would never have happened. If Rep. Porter had not demanded that SSA meet the standard or suffer budgetary consequences, the agency might never have gone to the extraordinary lengths it did to reengineer its work.

Without enforcement of consequences, in other words, standards pack much less punch. There are several ways to create consequences:

• Guarantees give customers who have not received the promised level or quality of service either their money back or free redelivery of the services. This costs the agency money-a negative consequence-while also building customer loyalty.
• Redress policies give customers some form of compensation-financial or otherwise-if the organization fails to meet a service standard. For example, British rail lines award discounts on future purchases when they miss their on-time standards.
• Performance awards or bonuses reward units that meet their standards. Performance against service standards is in the bonus formula for some top managers at Social Security. And many British agencies include key service standards among their annual performance targets, which affect top management's performance bonuses.
• Public comparisons of performance against the standards builds pressure. NPR's annual publication of SSA's results was a big part of what kept it from abandoning its 800-number standards. "They knew that they were going to be publicly humiliated if in fact they hadn't delivered what they needed to deliver," says NPR's Kane. In the United Kingdom, a Citizen's Charter led to the publication of comparative data on local governments, schools, hospitals and passenger rail lines. This, the new Labor government concluded after a review of the initiative, "was considered by many respondents to the consultation exercise to be a major success of the old Charter programme."

5. Keep pressure on from outside the organization to create meaningful standards and to meet them.
Setting meaningful standards and then fulfilling them sometimes takes almost heroic efforts, as we saw in the case of Social Security. Most organizations won't do that-or will only do it until the leader who drove them moves on. So you need some external force that keeps the pressure on-forever.

You also need an external body to review and approve standards, redress policies and so on. Otherwise, vague statements that cannot be measured and have no consequences attached-"We will do our best to provide timely, courteous service"-will be the norm. The review process should involve customers, ideally through an ongoing customer council.

A neutral reinvention office like the NPR can handle the review and approval process. But it helps to use an organization independent of the administration-such as the Government Accounting Office, a customer council or even a private entity like Consumer Reports-to bring outside pressure to bear. There are limits to how much pressure an internal arm like the NPR can bring on agencies. Imagine the media flap if Al Gore had publicly criticized Education Secretary Richard Riley's standards, and you can understand why.

Because some fudging is inevitable in any kind of performance measurement, an outside body also needs to audit definitions, indicators and measurement efforts to keep them honest.

6. Involve employees in creating and meeting standards.
If standards are simply imposed on employees, few will respond to the challenge. At Social Security, says Warden, "I don't think we could have done it without really engaging the employees in discussions, getting their ideas about how we could do this better. We talked with thousands of them."

There are many reasons to involve staff: to get their ideas and buy-in; to see if draft standards are feasible; to learn what work-process and other changes need to be made to reach them; to identify the training employees need to meet the standards; and to identify other offices or units whose cooperation will be necessary to meet the standards.

Similarly, if you fail to deliver the quality of service you have promised, front-line staff need to be able to make it right immediately. If they have to wait three weeks for management to make a decision, you will alienate your customers.

7. Beware of perverse incentives that actually hurt customer satisfaction.
"Watch out for how you set your standard up," warns Warden. "You've got to present a total picture of what the customer wants. We could have very well said '95 and five,' without continuing to emphasize the courtesy and accuracy" in other standards.

The result could have been teleservice representatives hanging up on customers before they were satisfied. As Warden reminds us, "What you say is what you get."

You have to watch closely for any backfires, adds Larry Massanari, SSA's regional commissioner for the mid-Atlantic region. "You don't always know; only the customer begins to feel it." This is yet another reason to keep measuring customer satisfaction. As the NPR emphasizes, it's important to keep your eye on the ultimate outcome. Customer service standards are powerful, but they are only a means to an end.

David Osborne, a partner in the consulting firm The Public Strategies Group, is co-author of Reinventing Government (Addison-Wesley, 1992) and Banishing Bureaucracy (Addison-Wesley, 1997). This article is drawn from material in his new book, The Reinventor's Fieldbook: Tools for Transforming Your Government, published this month by Jossey-Bass.