Last updated · By the Clerk2Officer editorial team · Report a correction
1. Connect a standard with actual performance
Controlling is the management process of setting or using standards, measuring actual performance, comparing it with the standard and taking appropriate action. It helps an organisation understand whether its plans are producing the intended results and why differences occur. Control is not limited to catching mistakes after harm has occurred. Some controls prevent an error, some detect it and others support correction or recovery. A useful system connects information with an accountable decision rather than producing reports that no one acts upon.
Consider fictional Meridian Bank's Riverside branch. Management sets a goal to reduce waiting time for routine service while maintaining accurate processing and required checks. A month later, the branch's average waiting time is lower, but the number of corrected applications has risen. The manager should not automatically declare success or assume that employees are careless. The next step is to examine the measures, the cause and the consequences. A new process may have improved one part of service while creating a quality problem elsewhere.
Control applies at different levels. A frontline supervisor checks whether a transaction follows the required process; a regional manager reviews several branches; senior management examines whether strategy and resources are producing the desired outcomes. The scope and frequency differ, but the basic questions remain: what was intended, what happened, why did it differ and what action is appropriate? Control also provides feedback for future planning. A standard that proves unrealistic or measures the wrong behaviour may need authorised revision rather than repeated unexplained exceptions.
2. Understand the control process
The process begins with a clear standard or expected result. The standard may be quantitative, such as a specified service period, or qualitative, such as documented compliance with an approval procedure. It needs a definition, scope, responsible owner and measurement method. A target without a reliable source of evidence is difficult to control. The organisation should distinguish a mandatory requirement from an internal management target; missing a legal or policy obligation cannot simply be treated as an acceptable variance because another measure looks favourable.
Next, measure actual performance using consistent data and compare it with the standard. If the branch target is an average wait of fifteen minutes and the measured average is eighteen, the difference is three minutes above target. That arithmetic describes a variance, not its cause. Investigate whether demand was higher than planned, staffing was insufficient, the system was slow or the data definitions changed. Corrective action should address the relevant cause and be proportionate to the impact and urgency of the problem.
Finally, implement and follow up the action. A manager might adjust scheduling, improve instructions or escalate a system issue, depending on the evidence and authority. Record the owner and date, then check whether the result improved without creating another problem. If the original standard or assumptions were unsuitable, management may need to revise the plan through the approved process. Changing a target after seeing an unfavourable result solely to avoid reporting a variance is not the same as a reasoned revision supported by evidence.
3. Link planning and control
Planning defines intended objectives and actions; controlling examines performance against relevant standards and supports adjustment. They are distinct but interdependent. Without a plan or standard, the manager may struggle to decide what a result means. Without control, the organisation may not learn that a plan is failing or that its assumptions no longer fit. The relationship is a feedback loop, not a one-way sequence that ends once a report is filed.
Suppose Riverside plans to process 1,000 routine applications in a month with an agreed quality standard. Actual output is 900. The shortfall is 100 applications, or ten percent of the target. Before deciding that staff should work faster, the manager should examine incoming demand, complexity, absences, system availability and the accuracy of the count. If the branch received only 900 eligible applications, output alone may not show a capacity failure. A variance needs context, a defined denominator and a connection to the decision being considered.
Control information can improve the next plan. If a predictable salary-credit peak repeatedly causes delays, the branch may adjust schedules or capacity. If a target encourages employees to avoid recording corrections, the measure itself may need redesign. A plan should therefore include both the desired outcome and safeguards against harmful shortcuts. Customer service, control quality and staff workload should be considered together rather than allowing an isolated productivity target to become the organisation's sole definition of success.
4. Distinguish preventive, detective and corrective controls
Preventive controls aim to stop an undesirable event before it occurs. Examples include access restrictions, input validation and a required approval before a sensitive transaction is released. Detective controls identify an event or difference after it occurs, such as reconciliation or an exception report. Corrective controls support recovery, repair or remediation. These categories describe the control's primary purpose and timing; a complete process often uses several layers because no single control can prevent or detect every possible failure.
Imagine an employee attempts to send a payment to an unauthorised beneficiary. A validated approval workflow may prevent release. If an incorrect transaction still enters processing, a reconciliation or unusual-transaction alert may detect a mismatch. An authorised recovery or correction process can then address the consequences where possible. The timing matters. A detective control performed weeks after an irreversible transfer can reveal the event but may provide little opportunity to limit the loss. A control should therefore be assessed for both design and execution.
A written policy describes expectations but does not prove that a control operated. Evidence might include dated approvals, access reviews, reconciliations and records of resolved exceptions. The manager should ask who performed the check, what information was available, whether the check occurred in time and how problems were escalated. A tick in a checklist can be useful when it is supported by meaningful work; a repeated tick without review creates the appearance of control while leaving the underlying risk unchanged.
5. Understand feedforward, concurrent and feedback control
Feedforward control acts before an activity, using information about inputs or anticipated conditions. Concurrent control occurs while the activity is taking place. Feedback control examines results after the activity and uses them to improve future performance. These categories focus on timing. They complement the preventive, detective and corrective view rather than replace it. A bank can use a pre-processing validation, live monitoring and a later review together to manage different points in the same customer-service process.
For Riverside, checking that employees are trained and authorised before they process a new product is a feedforward measure. A supervisor observing the process during a pilot and responding to an emerging queue provides concurrent oversight. Reviewing a month's waiting times and error patterns provides feedback for the next plan. None of these examples guarantees success. A trained employee can still face a confusing system response, a supervisor can miss an event and an average report can hide a severe individual customer problem.
Timing should match the risk and decision. An annual review may be useful for a broad trend but too late to prevent repeated payment errors today. Continuous monitoring can generate excessive alerts if thresholds are poorly designed or no one owns them. Select frequency based on the consequences, available evidence and capacity to act. A control that produces information faster than anyone can interpret it may create noise rather than protection, especially when important exceptions are buried among low-value alerts.
6. Choose meaningful measures and standards
A measure should reflect the objective, be defined consistently and be capable of supporting action. For waiting time, specify when the clock starts and ends, which customers are included and whether the figure is a mean, median or another statistic. For quality, define a correction or error and how it is recorded. Compare like with like across periods. A lower average can coexist with a group of customers waiting much longer, so a manager may need more than one measure to understand the actual service experience.
Financial measures can include budgets, costs and income; operational measures can include processing time, rework, unresolved exceptions and system availability. Customer feedback and staff observations provide additional evidence. Each measure has limitations. A low count of reported errors can indicate better quality or poor reporting. A high count after a new reporting channel is introduced may reflect improved visibility. Interpret changes in context rather than treating every movement as proof of a single cause.
Measures influence behaviour. If staff are rewarded only for the number of applications completed, they may feel pressure to overlook accuracy or customer suitability. A balanced control system considers speed, quality, compliance and customer outcomes together. Some requirements are boundaries that cannot be traded away for a better average score. The manager should explain why the measure exists, who owns it, what action follows a warning and how staff can report a problem with the measure or the process.
7. Analyse variances before acting
A variance is the difference between actual performance and a standard. Its sign and meaning depend on the measure. Spending more than budget may be adverse, but a higher customer-satisfaction score may be favourable. Always state the direction. If budgeted cost is INR10 lakh and actual cost INR12 lakh, the variance is INR2 lakh above budget, or 20% of the budget. The calculation does not establish whether the extra cost was avoidable or whether it produced an important benefit.
Investigate material differences using relevant evidence. The extra cost may reflect higher demand, a one-time system repair, incorrect coding or an unplanned activity. A single total can combine several causes. The manager should separate volume, price and process effects where the available information permits, then decide which cause can be controlled. An automatic instruction to cut cost could harm service or controls if the increase came from a necessary activity. Conversely, explaining a variance is not a substitute for action when a preventable problem remains.
Materiality and frequency matter. A small repeated error can accumulate or indicate a weak process, while a large one-time event may require immediate escalation. Thresholds should support attention without encouraging employees to split or conceal issues below a reporting limit. Investigate trends, ageing and unresolved exceptions as well as current totals. The objective is to identify the decision needed, not to attach a favourable or adverse label without understanding how the result arose and what its consequences are.
8. Use control techniques appropriately
Budgetary control compares planned and actual financial or resource amounts. It helps managers identify differences, but a budget is based on assumptions that may need revision when demand or circumstances change. A fixed budget should not prevent an authorised response to an unexpected critical need merely because the original forecast did not include it. The organisation needs clear approval and reporting for exceptions. Equally, repeated unapproved overspending should not be excused indefinitely as an unforeseeable event if it reflects a known planning weakness.
Operational controls include procedures, segregation of duties, reconciliations, access controls, quality reviews and exception monitoring. These techniques serve different purposes. A maker-checker process can reduce an opportunity for unauthorised action, while reconciliation helps detect a mismatch. A process should be designed around the failure it addresses and the time available to respond. Adding many checks without clear ownership can delay work without improving protection, especially if every reviewer assumes someone else has examined the critical information.
Management information systems can collect and present data, support alerts and preserve records. Technology improves control only when data are complete, rules are appropriate and someone acts on the output. An automated report built from missing transactions can be consistently wrong. Access and changes to rules need authority, testing and audit trails. A digital dashboard should identify definitions, update time and exceptions rather than conceal uncertainty behind a green indicator. Human judgement remains necessary to interpret the cause and decide on a response.
9. Recognise limitations and unintended effects
Control systems have costs. Measurement, review and documentation use time and resources, and excessive controls can slow service or encourage workarounds. Controls can also fail through poor design, inconsistent execution, collusion or changes in the process they were intended to protect. The answer is not to remove every check. It is to assess the risk, design proportionate layers and test whether they work. A high-risk payment process and a low-risk internal enquiry may reasonably need different intensity of oversight.
Measurement can create unintended incentives. If a branch is judged only on the number of complaints formally recorded, employees may be tempted to discourage complaints instead of fixing problems. If a target rewards speed alone, quality may decline. A control system should make it safe to report issues, use complementary measures and review whether the measure is producing the desired behaviour. The manager should be alert to a sudden improvement that coincides with a change in definition or reporting channel rather than assuming the underlying process improved.
Control also depends on reliable information. A report may be delayed, incomplete or based on inconsistent definitions. A new system can introduce data errors at the same time that management increases reliance on automated measures. Independent reconciliation and testing help, but no control eliminates all uncertainty. Reports should identify missing data and significant limitations. Managers should be able to explain why a figure is suitable for the decision and what additional evidence is needed before taking a consequential action.
10. Coordinate accountability and escalation
An effective control framework assigns ownership. The employee performing a task should know the standard and how to report an exception. A supervisor reviews relevant results and can resolve issues within authority. Independent functions provide challenge or assurance according to their roles, while senior management and the board oversee the framework. The exact structure differs by institution. The essential point is that an exception must reach someone who can decide, with enough information and time to act.
Escalation should be based on materiality, urgency and the applicable rules. A small difference may be corrected through a routine process; a suspicious payment or serious control breach may require immediate containment and formal reporting. Staff should not wait for a monthly meeting when delay increases harm. Equally, managers should avoid creating so many low-value alerts that important issues are lost. A threshold needs a defined response and an owner, and repeated unresolved alerts should trigger review of the process itself.
Follow-up establishes whether an action worked. If the branch introduces a new checklist after repeated errors, review actual files and customer outcomes rather than assuming that issuing the checklist solved the problem. Assign a deadline and verify closure using evidence. Some findings require a change to a system, staffing or incentive rather than more reminders. Accountability means making the remaining risk visible and ensuring that a responsible person decides what to do, not merely transferring a report between departments.
11. Review the Riverside control case
Riverside's manager should begin by checking the quality and definition of the waiting-time and correction data. If the figures are reliable, investigate why corrections rose: perhaps a new form is confusing, employees lack training or the system displays an incomplete status. The appropriate response depends on the cause. A pilot correction should have an owner, a date and measures for both service and accuracy. The manager should communicate the change and ensure that required checks remain in force while the process is being improved.
For examination practice, distinguish the standard, actual result, variance, cause and corrective action. Identify whether a control prevents, detects or corrects a particular failure, and whether it operates before, during or after the activity. Calculate percentages using the stated base: a two-lakh excess against a ten-lakh budget is twenty percent, not two percent. Explain that a variance is a signal for investigation, not proof that a particular person caused the problem or that cutting a cost is automatically the right response.
The management cycle closes when evidence from control improves planning and execution. A control framework is valuable when it helps the bank achieve objectives responsibly, protects customers and makes problems visible in time to act. Its effectiveness depends on meaningful standards, reliable data, proportionate checks, clear ownership and follow-up. Used in this way, controlling supports rather than replaces the judgement and cooperation developed through planning, organising, staffing and directing.
Chapter practice
This practice is separate from timed mock tests. Answers and explanations appear after you choose an option.
Your reading progress
Sign in to save your progress across devices.