
Human Resource Management — Scope and Limits of Application
1. HR contributes to performance; it does not explain every result
The preceding chapters concern recruitment, development, appraisal, compensation, motivation, leadership, and labor relations. Their common practical question is how an organization establishes rules for acquiring and coordinating human capabilities. That does not make HR the sole explanation for revenue, profitability, or survival.
Consider a business that recruits excellent employees into a product market with little demand. Faster hiring can help it execute a mistaken strategy more efficiently. Conversely, a favorable market can conceal weak supervision or an unsustainable workload. A graduate reviewing HRM should be able to distinguish the quality of a people system from the quality of the strategy it serves.
A useful conceptual map is market and strategy → required tasks → people practices → workforce behavior → operating outcomes. Feedback runs in the other direction: disappointing operating results may change budgets, hiring, and training. This is an analytical map, not a measured causal model. It reminds us to inspect product demand, task design, pricing, and resource constraints alongside HR decisions.
A. Vertical and horizontal alignment answer different questions
Vertical alignment asks whether people practices support the selected strategy. A quality-oriented strategy, for example, calls for capabilities and incentives that do not reward output volume at the expense of defects. Horizontal alignment asks whether hiring, training, appraisal, and compensation support one another.
A company that trains staff to collaborate but ranks them exclusively on individual output has an internal tension. Fixing that tension does not establish that its product strategy is sound. Nor does adopting a named bundle of practices prove a competitive advantage: implementation, fit, cost, and alternatives still require evidence.
2. Separate activities, workforce outcomes, and business results
A report can contain correct numbers while supporting the wrong conclusion. Read each indicator as a measure of a defined object rather than as a general score for “good HR.”
| Indicator | What it can establish | What it does not establish by itself |
|---|---|---|
| Training completion | Recorded completion under a defined rule | Learning, transfer to work, or profitable improvement |
| Turnover rate | Departures relative to a defined workforce denominator | Whether valuable employees stayed, why departures changed, or whether retention improved output |
| Time to fill | Elapsed time between specified hiring milestones | Quality of hire, subsequent performance, or absence of rushed selection |
| Appraisal average | Recorded ratings under a particular scale | A change in actual performance when standards or rater behavior changed |
| Revenue or margin | A business outcome under accounting definitions | The portion caused by HR rather than demand, pricing, technology, or other changes |
Define the population, time window, and calculation before comparing periods. “Turnover” might include all departures or voluntary departures only. A company-wide mean can conceal a serious loss in a critical occupation. A shorter time to fill can also reflect a change in the starting point used to count days.
Reliability concerns consistency; validity concerns the support for an intended interpretation. Counting attendance consistently does not make attendance a valid measure of skill. A well-defined rate can still be unsuitable for the decision under consideration.
3. A numerical case: improvement is not attribution
The following is a fictional one-year comparison, not an observed company result. Assume the definitions and reporting populations remain comparable.
| Measure | Previous year | Current year |
|---|---|---|
| Annual turnover | 12% | 8% |
| Training completion | Not provided | 95% |
| Market share | 15% | 9% |
| Revenue | USD 40 million | USD 28 million |
Turnover’s relative reduction is about 33.3%, because the four-point reduction is divided by the original 12% rate. Calling it a “4% reduction” confuses a percentage-point difference with a relative change. Likewise, market share falls by six points, a 40% relative decrease. Revenue falls by USD 12 million, or 30% of the original USD 40 million.
These arithmetic statements do not identify causes. Training completion is high in the current year, but no previous-year completion rate was provided. There is no basis for claiming it improved, let alone for claiming it saved the business.
The turnover decline could reflect better management, fewer attractive outside jobs, a change in the workforce mix, or several mechanisms. Revenue could fall because of lost demand or another operating change. Those are hypotheses to investigate, not facts supplied by the table. Improved HR indicators and deteriorating business results can coexist; neither cancels the need to investigate the other.
4. What would strengthen a causal argument?
Start with a specific claim: “This supervisor-training intervention reduces preventable departures among the eligible teams.” Record eligibility, exposure, the outcome definition, and the measurement period before examining the results. This is more testable than “training creates business value.”
A before-and-after comparison cannot distinguish an intervention from every change that occurred at the same time. A comparison group helps only if its relevant differences are addressed. If participating teams volunteered or were selected because of especially poor results, a simple average comparison is not enough.
| Evidence arrangement | What to examine before interpreting it |
|---|---|
| Before and after in one group | Concurrent pay, staffing, demand, and measurement changes |
| Participants versus nonparticipants | Selection into participation and differences in jobs or supervisors |
| Random assignment, where feasible and appropriate | Assignment procedure, contamination, missing outcomes, and whether the measured response matches the question |
| Follow-up operating measures | Time lag, unintended effects, and consistency of definitions |
Random assignment is a design feature, not a synonym for a large sample. NIST explains that completely randomized designs assign factor levels to experimental units randomly; its design-selection guide starts from the experimental objective and factors being studied. NIST: completely randomized designs; design selection.
The table applies these design questions to a hypothetical HR review; it is not a claim that every workplace intervention should be randomized. Employment rights, privacy, fairness, operational constraints, and the possibility of withholding necessary support may rule out a proposed arrangement. State what remains uncertain rather than assigning the unexplained part of performance to HR.
5. Avoid making the metric the entire objective
Suppose a hiring team is rewarded only for time to fill. It can improve that metric by changing the recorded start date or lowering selection standards. Suppose a manager is rewarded only for low turnover. Retaining people in unsuitable roles might lower departures without improving service quality. These are possible incentive conflicts, not evidence that a particular company behaved this way.
Pair the target with relevant safeguards: hiring speed with early performance and early departures; training completion with job-relevant learning and transfer; retention with role-specific departures, workload, and contribution. The safeguards also need clear definitions. Adding more indicators does not automatically solve the problem if all measure the same activity.
Legal compliance belongs among the constraints on organizational choices, not as a substitute for evaluating outcomes. In cross-country comparisons, check the local rules rather than importing one jurisdiction’s thresholds. Chapter 9 treats labor relations separately from performance measurement.
Practice: write a defensible management conclusion
Problem
Using the fictional table, write a short conclusion for a manager who wants to claim: “HR improved retention by 4%, achieved 95% training effectiveness, and prevented a larger revenue decline.”
Worked answer
“Annual turnover decreased from 12% to 8%, a four-percentage-point decline under unchanged definitions. Current-year training completion is 95%; the data do not measure training effectiveness or provide a prior-year comparison. Market share and revenue declined. We cannot attribute the turnover change or a hypothetical avoided revenue loss to HR from these figures alone.”
The next investigation should separate voluntary and involuntary departures by relevant roles, examine the intervention’s timing and participation, measure job-relevant learning or behavior, and inspect demand, pricing, and operating changes. The available evidence supports descriptive changes, not an estimated causal effect. This preserves the distinction between reporting activities and demonstrating contribution.
OIYO Editorial
Editorial DeskThe OIYO editorial desk researches money, law, lifestyle, and self-understanding topics against primary sources and public statistics. Every piece carries source notes and is reviewed on a regular cycle for accuracy and usefulness.