by Yves Morieux
The key operational units that had to cooperate with each other included maintenance staff, train drivers, conductors, and station crews. Members of the maintenance unit, for example, could immediately inform the other functions about the nature of the repairs they were going to perform and how long they would take. In turn, platform managers could then announce the delays to passengers and position them on the platform so they were ready to board as soon as the repaired train arrived; a more efficient boarding process would make up for the delay that originated in the maintenance unit. Vice versa, whenever a train arrived at the station needing maintenance, the station manager could speed things up by sending it to the platform closest to the maintenance shed. The train conductors could help the platform team manage the flow of passengers once boarding had begun. In other words, cooperation would provide a wider set of choices so that, depending on the circumstances and specific nature of the unexpected issues, RapidTrain could find a superior solution to satisfy its conflicting performance requirements.
The operational units at RapidTrain, however, had not been cooperating in the ways that were necessary to solve the on-time arrival problem. What was all the more frustrating for senior managers (and employees) was that the organization would display exceptional cooperation during crises (caused by extreme weather, for instance). In a crisis situation, the teams would quickly find ingenious solutions that were considered wonders by the usual industry standards. Putting direct feedback loops in place was not an option to achieve the same level of cooperation on a day-to-day basis. RapidTrain personnel were insulated from reciprocal impacts because of the very nature of train traffic in a large network. Members of the functions worked at numerous sites throughout the day and interacted with the various functions differently in every situation. Rule five—extending the shadow of the future—was not applicable because a maintenance worker can’t become a conductor on the train he has just repaired in order to experience the consequences of his actions. A train conductor needs to be on the platform before the train arrives from maintenance, preferably with clean clothes and hands.
It was up to the management team to close the feedback loop. What performance evaluation criterion could it use to create a context within which cooperation became the best choice for team members? To answer this question, we worked with managers at RapidTrain, first to understand the existing work context. We discovered that the real goal of the unit managers was not so much to make sure the train was on time for the passengers as it was to avoid being found guilty of creating delays and getting blamed for them.
That goal had arisen as a direct result of one of the earlier measures RapidTrain had taken in an attempt to improve its on-time percentage—the creation of a new monitoring function. Whenever a train was late, the monitoring function pounced and undertook an investigation to determine which unit had caused the delay. Thousands of hours were spent in these investigations, as the monitoring function issued one finding after another, and the units rebutted them or tried to explain them away. An investigation would go on until it could reveal which unit was responsible for the technical factor that was the root cause of the delay. Had someone failed to replenish the oil stock? Had the wrong platform been announced? When the technical root cause was determined, the manager of the unit in which it occurred was pronounced guilty, with an impact on annual evaluation and promotion decisions. This approach is standard practice in the railway sector, as well as in many other businesses. Whenever there is a delay—such as a holdup in construction or a slip in the delivery of a piece of software—organizations determine the technical root cause and then place blame.
The main resources of a unit manager were the team and equipment under his or her direct control. Were the other units also a resource? Remember that nothing is a resource in itself. It depends on the actor’s goal as shaped by the work context. So, yes, the other units were a secondary resource—however, not because they could help, but because they were potential candidates to get the blame when something went wrong. There was little mutual help among the functions, except during a serious crisis. To get help, one needs to ask for help. But as soon as a unit asked for help, it was signaling that it was the root cause of the delay. So when a problem happened in a unit, its manager and team would try to accelerate and make up for the delay by themselves. Sometimes that worked but not regularly enough to improve on-time performance. Delays in one function would translate into delays in the others, delaying the train and affecting the network.
Of course, no one deliberately decides to be the technical root cause for a problem. Generally, if you delay others, you don’t make a deliberate choice to do so. However, when the others are delayed because they do not cooperate to help make up for the delays you have caused, they have made a deliberate choice not to cooperate. If someone is to be blamed, who should it be—you or the others?
We have never encountered a situation in which people were not able to cooperate. Our constant observation is that people have room to maneuver in allocating their efforts to ensure either the protection granted by their measurable contribution or the nonmeasurable contribution to overall results. The typical barrier is not unwillingness to spend more energy—people usually fully spend it anyway in protecting themselves one way or another—but the risk involved in turning an effort into nonmeasurable cooperation. At RapidTrain, the analysis showed that people were making the greatest efforts in trying to catch up and accelerate on their own, but often in vain. To be sure, being transparent with others would have required much less energy. What if all these efforts and intelligence, instead of being used in vain, could be turned toward making the trains run on time, thanks to cooperation?
Based on this analysis, RapidTrain senior management made a major change in the evaluation criteria. The executives decided that once a unit told others it had a problem, the units that failed to cooperate in solving the problem would be held responsible for the delay. To realize how radical the change was, think of it this way: it was a bit like saying, “When another unit causes you to be late, you are going to be the one that takes the blame.” The key question was no longer: “Are you technically the cause of the delay?” It was: “Have you cooperated to solve it?” Evaluation was no longer hinging on technical criteria but rather on organizational ones.
Each week, the managers of the units sat down with their superiors to review the delays and to answer the new question. Station managers, who were present at some key moments of cooperation, also judged if units had contributed to solving problems. Of course, the senior managers also made it clear that if the same unit remained the cause of repeated problems because it did not engage in continuous improvement, it would take the blame. But there was no longer a direct link between technical causes and blame. It became in the interest of those who needed help to be transparent about it and in the interest of others to provide that help. The enhanced cooperation, in turn, allowed RapidTrain to bring further improvement to each unit’s work processes.
In just four months after initiating the new approach, on-time performance at RapidTrain jumped to 95 percent on the major lines where the change had been implemented. This was achieved without new equipment, new scheduling systems, or additional trains or teams. A side benefit was that people no longer had to spend thousands of hours in the root cause investigations, as it was in everybody’s interest to be transparent about his or her own problems—and also because there were fewer delays.
Another outcome was also important. As part of the evaluation of the change initiative, one team launched an employee survey and interviewed the unit members to probe how they felt about the new approach. The feedback was that they felt happier than they had before. Three factors seemed to account for this greater satisfaction at work. First, the teams in contact with customers were now able to provide more helpful answers in case of problems, while taking mitigating actions. Relationships with customers had changed for the better. Second, hierarchical relationships had improved. The managers coul
d now help units get the cooperation they needed from others. Third, there was also some pride in breaking the records. Indeed, this is what happens when efforts are turned toward cooperation. (See the sidebar “Use Performance Evaluation to Enhance Cooperation.”) Interestingly the maintenance supervisor told us when we met him again that his real role was “to make teams cooperate so that passengers, not only maintenance, arrived on time.” His mind-set had changed, and without psychoanalysis or bureaucracy therapy. In fact, the context had changed, behaviors had intelligently adjusted, and mind-set had evolved as a consequence.
SIMPLE RULES TOOLKIT
Use Performance Evaluation to Enhance Cooperation
Don’t punish or blame people for results but encourage in-depth knowledge of how results are obtained and who helped out.
Managerial presence and feedback loops capture how each individual contributes to the effectiveness of others, making it more difficult to pass the buck to the weak roles that bear most of the adjustment cost. Helping others with their results then becomes attractive and, in turn, fuels transparency about performance.
Do not confuse the following three terms: business steering (which requires many metrics and KPIs to identify and anticipate the trajectory of the business), performance management (evaluating people’s performance by using some metrics and also your qualitative judgment, providing advice for improvement), and rewards (the way to gratify contributions). The balanced scorecard is often a sophisticated form of confusion among these three elements, where remuneration ends up being a direct function of the score achieved by people in performing against the weighted average of the multiple business KPIs. The system has computed your score, and as your manager I am very happy to tell you that this year you have achieved a great 4.81—I am not surprised—for the rest, please see the accountant who has your check ready, and don’t forget to celebrate! If the manager is not surprised, what was the point in having the score? If the manager is surprised, what was the point in having the manager? It is rare that managers say they are surprised, which simplifies discussions but does not simplify the balanced scorecard, let alone the achievement of overall results.
All too often, a company chooses performance assessment criteria to link things that go well or badly in operational processes to specific areas of responsibility. The more direct, accurate, and clear the link, the more it thinks it has the right assessment system. But the proper goal of an evaluation system is not to be technically right in this link. Rather, it is to elicit full engagement and cooperation. So, in using an evaluation system, ask: “Are we trying to be true to the technicalities of the job description or trying to generate engagement and cooperation?” This question often causes managers to change the criteria, move away from technicalities, and emphasize the parts of work that make a difference.
Make Those Who Don’t Cooperate Bear the Cost
One way to apply this simple rule is to adopt the principle established by Jørgen Vig Knudstorp, CEO of the LEGO Group: “Blame is not for failure, it is for failing to help or ask for help.”1 When this is the rule, people become much more transparent about their weaknesses, uncertainties in their business forecasts, and opportunities they have for improvement.
An organization is much more resilient when people know that it is in their individual interest to help others and to be transparent than when people are judged and rewarded on their ability to avoid mistakes in their own area. When people work in this way, they will always have the help of others accountable for finding the solution to a problem. This is the strength of the Musketeers’ motto: “One for all and all for one.”
This approach puts the vital issue of tolerance for failure in a useful perspective. Intolerance for failure is bad. It can lead to risk phobia: “Don’t take initiatives. Don’t try new ideas. Hide your mistakes!” Issuing a decree of zero tolerance for mistakes will not prevent them from happening. It will only cause people to hide the mistakes that do occur.
Still, tolerance for failure is not always good either, because it can often just lower the bar. The proper intent of tolerance for failure is not to provide greater leniency or make requirements less demanding. Just because people are given the right to make mistakes does not mean the eraser should be allowed to wear out faster than the pencil.
The right way to administer tolerance for failure is to use criteria that place demands where people can create the greatest impact for the organization and its performance—where individuals have a margin of maneuver that could be combined with that of others to make a big difference. By using criteria that reward decisions that would otherwise be risks for people individually, even when good for the company, pressure can be focused on the points that foster cooperation and therefore have more impact than pressure within silos. This kind of tolerance for failure makes the resulting system more tolerant of failure—that is, more robust. The result is reliability without the need to multiply control mechanisms. Attitudes toward risk are often described in terms of culture or mind-set. Our people are risk-adverse. Our culture is not tolerant enough of risk. This is wrong. In fact all these issues usually have nothing to do with people’s particular psychology or mind-set. Most often they are organizational and practical matters of cooperation. Risk is not a goal in itself. What matters is the effect on organizational performance and individuals. Risk-taking is a good thing only when there is cooperation. Only cooperation can make risk-taking a rational strategy for the individual. People take personal risk when they know they can count on the cooperation of others—to compensate, relay, absorb, or provide a safety net in case things go wrong. And then risk also becomes fruitful for the company.
How does your organization administer the right to fail? Is it completely intolerant of failure? Does it verge on leniency? Or do you approach failure in a way that generates resilience?
Changing the Managerial Dialogue: Make Transparency a Resource, Not a Constraint
An important ingredient of evaluation is the dialogue that managers have with their teams. By evaluation, we mean the conversations and judgments about whether people have given their best, what helps them, and what obstructs them. (We are not referring to the boxes that have to be filled in as part of the standard annual evaluation form.) The way managers frame this conversation is important.
“What Do People Say When They Complain about You?”
A CEO we know used to begin the evaluation of his manager of shared services by asking, “I have heard some complaints from country operations about your responsiveness. What’s going on?” The CEO was on a fault-finding mission. So, for the manager, it became a game of defense, countering each complaint with positive testimonials he had received from other quarters.
This approach doesn’t make the best use of people’s intelligence or of the available information. The CEO, who was most removed from operations, had little choice but to focus his attention on second-hand information about the manager’s performance and supposed problems. The manager, in turn, had no choice but to use his intelligence to justify his performance, attack his attackers, and pump up his successes. It was not a conversation, but rather a sequence of thrusts and parries.
The CEO came to realize this approach was not leading to performance improvement and decided to change how he handled the performance evaluation. He entered the conversation differently, by asking, “What frustrations do you cause your internal clients? What do they complain about? How can I help you solve these problems? Only one thing would be inexcusable: that I hear from other people about issues involving you and I have not already heard about them from you. This would mean that I know your internal customers better than you do.”
By evaluating people on their knowledge of what is not working in their area, rather than encouraging them to play personal defense, the logic of the conversation is reversed. The responsibility for supplying information and for acting on it is put on those who are best placed to do so.
This kind of dialogue can’t rely on in
formation gathered through internal customer satisfaction surveys and the like. These may provide a rank-ordered list of complaints, but they don’t tell you what really matters. How, for example, does the shared services manager view these complaints? What does he make of them? Only he can understand the implications of these findings for his own processes and operations, and only he can determine what he needs to do in order to improve.
“What Personal Risk Are You Taking in All This?”
The way the performance cycle starts in many companies explains why evaluation is often performed entirely without reference to cooperation and without any understanding of context, and therefore drives underperformance. The performance cycle starts when targets are set at the start of the year.
Typically, once a year, the executive team asks line managers for their performance forecasts. This process can be repeated when there are specific needs for cost savings or top-line growth, in which case there is usually an overall ambition and performance improvement target already in place.
The negotiations begin. The executive committee sets very high targets, knowing that people will give performance forecasts lower than those they know they could actually attain. They do this to keep some concessions—something in reserve—for the next round of negotiations. They know there will be a second round, because all their colleagues will also offer low performance expectations in the first round, so the total will never add up to what the executive committee wants. The gap between ambition at the company level and the results of all the bottom-up commitments is fully anticipated.
After the second round, people raise performance expectations a little, but still set them short of what is achievable. They all know that committing to improvements, and delivering them, is going to make things even tougher next year. When the process begins again, they will have to start from a much harder place and go up from there, promising to deliver even more performance improvements, such as even more savings and even higher growth. So people keep as much as they can in reserve. Another reason people do not reveal or aim for the fullest possible improvement they believe they could attain is the risk of what others will do. Whatever you are in charge of—sales, Asia-Pacific sites, logistics, client accounts—the chances are you have some interdependencies with others. No one has total control of the outcome. If you can’t count on the cooperation of others, then setting even a reasonable target makes you hostage to fate. Better to play it safe.