When a Goal Becomes the Only Scoreboard: Sears' Auto Centers in 1992

Vague goals feel like letting yourself off the hook. “Read more” drifts for months, while “a book a week” comes with a tracker, a streak, and a small lift every time the count goes up. If a goal has ever felt more real to you the moment it became a number, goal researchers would say that instinct is sound. There is a second feeling you may know too. It arrives in November, when your count is behind and the thinnest book on your shelf starts to look unusually appealing. You never decide to cheat. The number just starts doing the choosing for you.

So when does a hard, specific number push you toward what you wanted, and what single check stops it from replacing the reason you set it? A 1992 investigation of Sears’ auto centers in California, and a public argument between goal researchers in 2009, answer most of that.

Keep the number, and give it a witness

A hard, specific goal tends to raise effort. Even the researchers who attacked goal setting in 2009 argued that its benefits were overstated, which is different from saying they don’t exist.

The trouble starts when the number becomes the only thing anyone checks. What happens then? The cheapest way to move the number wins, whether or not it serves what you were after. Your step count rises just as well from pacing the hallway as from a hill walk.

So the answer has two parts. Use a number when effort is what you’re missing. Then pair it with one check that looks directly at the thing the number was meant to stand for: the thinking you wanted from the books, the fitness you wanted from the steps, the customer a sale was supposed to help. The second part sounds like bookkeeping. The Sears case shows why it is more than that.

Inside a repair bay paid by the part

According to a Harvard Business School summary of the case, Sears management introduced a pay system with minimum work quotas, moving staff from hourly wages to base pay plus productivity incentives. Service advisers, the people who tell you what your car needs, had product-specific sales quotas and commissions. Those who missed their quotas could be transferred or have their hours cut.

California’s regulators then sent in undercover cars. In June 1992, the state’s Bureau of Automotive Repair said Sears employees had recommended unnecessary repairs in 34 of 37 cases, in UPI’s wording. Agents alleged overcharges of as much as $550 for needless work. A USA Today wire report put the average charge at $233 for repairs that were unneeded or never performed; another report from the time gave $223. The state said employees were paid by how many parts or services they sold, and it sought to close all 72 Sears tire and auto centers in California.

Sears fought back. It called the investigation flawed and said replacing worn parts before they fail was accepted industry practice. Chairman Ed Brennan’s open letter allowed only that “mistakes may have occurred.” On 22 June 1992, the company said it would take service advisers off commission and put them on straight salary. In September it settled with the state, agreeing to repay $3.5 million in investigative costs and attorneys’ fees. It also agreed to fund at least $1.5 million of auto-repair training at California community colleges and to run a national $50 coupon program for customers who had bought certain parts.

A lab experiment looks at the same pressure from another side. Schweitzer, Ordóñez and Douma found in 2004 that people who failed to reach specific goals were more likely to act unethically than people told to do their best. The effect held with or without money attached, and it was strongest when people fell just short. It is one laboratory experiment, and its abstract does not claim to test workplaces. What it does suggest is where the pressure peaks: close to the line, when one more unit would finish the job.

In 2009, Ordóñez, Schweitzer, Galinsky and Bazerman widened the argument in a paper titled “Goals Gone Wild.” They listed side effects: a narrow focus that neglects everything outside the goal, distorted risk-taking, more unethical behavior, less learning, damaged culture, and weaker intrinsic motivation. Their advice was to handle goal setting like a prescription-strength medication, dosed with care and closely supervised.

The first side effect on that list describes the repair bay. Under pay tied to parts sold, the most useful thing an adviser could tell a driver, that the car needed nothing, counted for nothing.

The case against

Goal researchers did not let the 2009 paper pass. In the same issue, Locke and Latham published a rebuttal titled “Has goal setting gone wild, or have its attackers abandoned good scholarship?”, defending goal-setting theory against the critique. Ordóñez and colleagues replied, calling for more research on the full range of what goals do. The exchange did not end with a winner, and you shouldn’t read it as settled here.

The evidence above also has hard edges. The lab result comes from one experiment; it shows that goals can raise misreporting under test conditions, not how common that is at work. The Sears case is a regulator’s allegation followed by a settlement. Sears disputed the probe, and even the average overcharge differs between reports.

Does that make the number the villain? Not on this evidence. Even if every charge were accurate, the case could not show that goals by themselves produce dishonesty. The more specific suspect is the design around the goal: quotas tied to commissions, lost hours for anyone who fell short, and nothing that measured whether a repair was needed.

That design point also marks where the short answer stops applying. When the number and the purpose are nearly the same thing, such as a filing deadline you simply have to meet, a second check adds little. The trouble lives in the gap between what is counted and what is wanted, and some goals have almost no gap.

A reading goal with a second column

A yearly book count has plenty of gap. If you’re behind pace in November, you still have a real choice. You can keep the tally and add one question beside each new title: would this book make the list if the count didn’t exist? If you see a run of no’s, the tally has started choosing your books. And what would the count say about a long, difficult book that took all of October? Only that it was one.

The same column works for a step counter. Laps of your kitchen late at night move your number; a question like “did I get outside today?” catches them.

At work, the check can be one paired measure. If you run a shop that counts parts sold, you could also track how many customers return with the same complaint. Jerry Z. Muller’s The Tyranny of Metrics argues that tying rewards to numbers often invites gaming, and that metrics work best alongside professional judgment. None of these paired checks was tested in the studies above. They follow from the shape of the Sears case, so take them as reasonable guesses rather than findings.

Your first step

The smallest version of that second column fits on one line. Choose one goal you are tracking with a number right now. Next to it, write what the number is supposed to stand for, in plain words: “reading that changes how you think,” “stairs without getting winded,” “customers who come back because they trust the shop.” Then write one yes-or-no question about that purpose that the number cannot answer.

Ask it at the moment the lab study points to, when you are just short of the target and the shortcut looks most reasonable.

A hard number earns its place when effort is the missing piece; it gives you a reason to start and a way to watch progress. It turns on you when it is the only thing you watch. Sears changed how its service advisers were paid after the undercover cars came through. A single question written beside your number is a far smaller change, and it lets the count keep measuring effort while something else keeps an eye on why you began.