A customer service manager reports a successful month. Calls are shorter, employees are handling more of them, and the department has met its productivity target.
Then someone asks how many customers are calling back.
In this hypothetical example, that answer changes the picture. If customers keep returning because their problems were not resolved, shorter calls may have created more work. The department met its target, but the business may now be spending more to deliver a worse experience.
This is a problem executives can easily miss. A report does not have to contain incorrect information to give the wrong impression. Sometimes the numbers are accurate. They simply leave out what matters.
Performance metrics help leaders make sense of complicated operations. They also tell employees where to direct their effort. Once a number influences compensation, promotions, or staffing decisions, people have a reason to prioritize it.
The question is whether improving that number also improves the business.
Wells Fargo offers a documented example of what can happen when the connection breaks down. In February 2020, the bank agreed to pay $3 billion to resolve criminal and civil matters concerning its historical sales practices. According to the US Department of Justice, unrealistic sales goals and management pressure contributed to employees providing millions of accounts or products under false pretenses or without customer consent. Wells Fargo admitted wrongdoing as part of the resolution. [1]
The Securities and Exchange Commission found that Wells Fargo promoted the average number of products held per retail banking household, known as its cross-sell metric, to investors even though it included accounts and services that were unused, unneeded, or unauthorized. The SEC’s $500 million penalty was part of the combined $3 billion settlement. [2]
Wells Fargo involved serious misconduct and failures of oversight. Sales targets do not inevitably produce such behavior. But the case exposes a weakness worth examining in less extreme settings: a measure of activity can be presented as evidence of value long after the two have stopped moving together.
Often, the problem is less visible. Nobody needs to falsify a report.
Imagine a sales manager whose bonus depends on revenue and an operations manager responsible for controlling costs. Sales wins a contract by promising extensive customization. Operations stays within budget by limiting the resources available to deliver it.
Both managers can explain their decisions. Both may meet their targets. The company still has to deal with the extra work, disappointed customers, and reduced margin.
Revenue alone will not reveal that problem. Neither will a departmental budget viewed in isolation. Someone has to follow the transaction through the business and account for what it took to fulfill the promise.
Research on compensation shows how incentives can change the way people divide their work. In a 2013 NBER working paper, Ann Bartel, Brianna Cardiff-Hicks, and Kathryn Shaw examined an international law firm that shifted incentives toward leadership activities. Team leaders increased their non-billable hours and shifted billable work to team members. [3]
That finding raises a practical question for anyone evaluating managers: does the scorecard recognize the work of helping other people perform?
A manager who delegates effectively may produce less individual output. Someone who trains an inexperienced employee may have less time for immediately billable work. Those contributions deserve examination, even when they are harder to count. The law firm study does not establish that every increase in leadership time improves profitability, but it demonstrates that compensation can change how responsibilities are distributed.
Edward Lazear’s study of Safelite Glass Corporation, published in the American Economic Review in 2000, examined a change from hourly wages toward pay based on output. He reported a 44 percent increase in output per worker. The increase reflected both productivity changes among workers and changes in workforce composition, rather than a pure increase in individual effort. Profits also increased with the compensation change. [4]
The lesson is that design matters. An incentive tied to observable production may work well in one setting. A role involving judgment, collaboration, or outcomes that take years to emerge presents a different measurement problem.
Executives reviewing their own targets can start with one question:
Could someone meet this target and still leave the business worse off?
If the answer is yes, the next step is to understand how.
For a customer service team, employees might shorten calls by transferring difficult cases or ending conversations before customers have a workable solution. In procurement, a lower purchase price might come with more defects or unreliable delivery. In operations, reducing maintenance spending might improve this quarter’s results while leaving the next quarter with a larger problem.
These are possibilities to investigate, not assumptions about employee behavior. Staff who work with the targets every day can help identify where the pressure falls and what the reports miss.
The underlying objective should be clear before the target is defended. “Resolve customer problems reliably and economically” gives a team a purpose. “Reduce average call duration” gives it a number. The number is useful only to the extent that it supports the purpose.
This does not require abandoning the measure. It may require reading it alongside something else.
Call duration becomes more informative when managers also examine repeat contacts and resolution quality. Sales revenue becomes more informative alongside margin, cancellations, and delivery commitments. Procurement savings need to be considered with quality and reliability.
The aim is to make important trade-offs visible. Adding every available measure can make priorities harder to understand and increase reporting work without improving decisions.
Some requirements should remain firm boundaries. Customer consent, legal compliance, and safety cannot be offset by stronger sales or higher output.
Leaders also need to consider what employees can reasonably control. A difficult customer mix, equipment failure, or unexpected change in demand can affect results. Managers should be able to explain that context, with evidence. Consistent review matters here; otherwise, exceptions can become arbitrary and employees may lose confidence in the process.
A practical place to begin is with one metric that materially affects compensation, staffing, or investment. Look at recent decisions made to achieve it. Ask what improved, what became harder elsewhere, and whether any work or expense was pushed into a later period.
Then ask what evidence would change the conclusion.
That last question is important. A performance review is less useful when its only purpose is to confirm that a target was met. It should help management determine whether the result deserves to be repeated.
The historical cases discussed here cannot tell us how often these problems occur across businesses today. They do show why the decisions behind a number deserve as much attention as the number itself.
Before rewarding a result, leaders should be able to explain what produced it, who benefited, and what it cost the rest of the organization.
Meeting the target is the beginning of that assessment, not the end.
Evidence and scope
This article draws on official enforcement records, a peer-reviewed study, and an NBER working paper. The customer service and sales and operations examples are hypothetical. The proposed review process is the author’s practical framework, not a validated diagnostic instrument. The historical evidence illustrates how incentives can operate; it does not establish a current trend across industries.
References
- US Department of Justice. February 21, 2020. Wells Fargo Agrees to Pay $3 Billion to Resolve Criminal and Civil Investigations into Sales Practices Involving the Opening of Millions of Accounts Without Customer Authorization.
- US Securities and Exchange Commission. February 21, 2020. Wells Fargo to Pay $500 Million for Misleading Investors About the Success of Its Largest Business Unit.
- Bartel, A. P., Cardiff-Hicks, B., and Shaw, K. (2013). Compensation Matters: Incentives for Multitasking in a Law Firm. NBER Working Paper No. 19412.
- Lazear, E. P. (2000). Performance Pay and Productivity. American Economic Review, 90(5), pages 1346 to 1361.
Cite this article
Alazzawi, Ali. (2026). When the Numbers Improve but the Business Doesn’t. Elyonik Vision. October 31.