
Updated 16 October 2026. A practical Zama Engineering Systems guide.

A biometric time attendance system records identification events, but reliable attendance reporting requires context around those events. A clocking event does not explain a whole working day. The organisation still needs an approved roster, employee assignments, exception review and a clear decision about which records are ready for the receiving team. This work becomes especially important when shifts cross midnight, employees change departments or supervisors authorise work away from the usual clocking point.
The aim is a reporting process that people can explain. Supervisors should understand why a record appears in an exception queue. HR should know which policy decision was applied. Payroll should receive an identified, approved version rather than a spreadsheet that changes without notice. Employees should have an understandable way to raise a missing or disputed record. A terminal installation alone does not establish these operating arrangements.
This guide focuses on managing shifts and reviewing records. It does not prescribe payroll formulas or employment rules. Those decisions belong with the organisation’s responsible HR and professional advisers. Software functions vary by product, version and configuration, so the proposed system must be demonstrated against the actual requirements. ZES can discuss assessment, installation, configuration, testing, training and support within an agreed scope.
Keep original events separate from attendance decisions
Start by distinguishing what the device recorded from what the organisation concludes. An original event identifies a clocking action at a time and location within the system’s available data. A reported attendance outcome may also depend on the assigned shift, approved leave, authorised field work and a supervisor’s explanation. Combining these without a visible distinction makes later questions unnecessarily difficult.
Ask the supplier how original events are preserved and how authorised adjustments are represented. A correction should not simply erase the evidence needed to understand the original discrepancy. The chosen system may offer an audit function, an approval workflow or a different supported arrangement. Confirm the actual behaviour rather than assuming every product uses the same method or keeps the same information.
Define the minimum correction record the business needs: the affected employee and period, original issue, supporting explanation, authorised decision, responsible reviewer and date. If the application cannot support an essential requirement, address that gap before purchase or agree a controlled supplementary process. The desired outcome is an understandable history, not a screen that appears tidy because unexplained records were silently replaced.
Give every employee a stable reporting identity
Attendance review becomes unreliable when one employee has several active profiles or two people share an ambiguous identifier. Establish which employee number is authoritative and how it relates to HR and payroll records. Names help users read reports, but names alone are a poor way to reconcile changes, similar spellings and transfers between departments.
Before enrolment or migration, review the employee register for duplicates, inactive records and missing assignments. Identify who can create a profile and who confirms that it belongs to the correct person. If temporary workers or contractors use the system, define their records and reporting treatment separately through the organisation’s approved process. Do not infer employment status from a device category chosen for convenience.
Keep effective dates for relevant changes. When an employee moves to a different team, the historical record should still be understandable in its original context. Ask how the selected software handles transfers and whether old reports are recalculated when a current profile changes. This is a practical demonstration requirement because an apparently simple administrative edit can otherwise change how earlier periods are interpreted.
Build a shift catalogue in ordinary language
Write a shift catalogue before configuring rules. For each shift, describe its intended start and finish, the relevant working-day convention, breaks and the employees or teams that use it. Identify whether the pattern is fixed, rotating or assigned individually. Give each pattern a recognisable name so supervisors can discuss it without needing to remember a software rule number.
Keep policy decisions distinct from software fields. A screen may offer options for rounding, grace periods or automatic pairing, but their availability does not make them appropriate for the organisation. HR should approve the intended treatment before the implementation team translates it into configuration. Where a rule is uncertain, record the open question instead of accepting a default simply because it produces a total.
Use the catalogue to identify genuinely different patterns. Several departments may share one pattern, while employees in the same department may require different assignments. Avoid unnecessary duplication that makes maintenance difficult, but do not force different work arrangements into a single rule. The catalogue should represent actual operations with enough clarity that a backup administrator can maintain it consistently.
Match rosters to effective dates
A roster tells the attendance process what was planned for a particular period. Decide who prepares it, who approves it and when it becomes available for reporting. If supervisors change assignments informally after events have already arrived, reviewers may struggle to tell whether an exception reflects unusual attendance or an outdated schedule. A clear roster process reduces that uncertainty.
Record the effective date of a shift change and the reason where appropriate. Ask how the application applies a new assignment to future records and what happens to earlier ones. Some products may allow bulk changes or recalculation, but the exact supported behaviour needs checking. A convenient editing feature should not become a way to rewrite a closed period without review.
When a roster changes at short notice, use a named approval route. The supervisor should identify the affected employees and dates, and the administrator should confirm the change was applied as intended. Keep a record of the previous assignment where the process requires it. This allows the reviewer to explain why the final report differs from the originally published plan.
Separate the calendar date from the work period
An overnight shift can begin on one calendar date and finish on another. The reporting process needs an agreed way to associate those events with the intended work period. Without that agreement, a normal clock-out after midnight may appear as an isolated event on the next day, while the preceding day looks incomplete. The error is in interpretation, not necessarily in the employee’s clocking action.
Ask the supplier to demonstrate a complete cross-midnight example using the proposed configuration. Review the assigned shift, original event dates, calculated result and exported fields. The organisation should be able to explain which work period receives the record and how reviewers can still see the actual timestamps. Do not accept a total alone as proof that the date relationship is correct.
Include changes around the overnight shift. A late roster revision, an additional punch or an approved early departure may expose weaknesses that a perfectly ordinary example misses. The purpose is not to test every imaginable case at once. It is to confirm that the system and reviewers can handle the real patterns the organisation expects before relying on the reports operationally.
Work through a hypothetical overnight example
Consider a hypothetical employee assigned to a shift beginning at 22:00 on Monday and ending at 06:00 on Tuesday. The device records an entry before the scheduled start and an exit shortly after the scheduled finish. These times are illustrative, not a payroll calculation or a recommendation for working hours. The reviewer needs to see the Monday-starting work period alongside the original Monday and Tuesday events.
Now add an extra identification event during the shift. Ask whether the application treats it as another attendance boundary, a duplicate candidate or an event needing review. The correct operational interpretation depends on the organisation’s clocking procedure and the supported configuration. A supplier should demonstrate the behaviour rather than promise that every extra event is automatically harmless.
Finally, remove the exit event from the test data. The resulting record should become recognisably incomplete under the agreed process, not silently treated as a confirmed full shift. The supervisor can then follow the missing-punch workflow. This three-stage test checks ordinary pairing, an additional event and an incomplete record using the same roster, making the effect of each change easier to understand.
Make break expectations explicit
Different organisations record breaks differently. Some expect specific clocking actions, while others apply approved schedule rules or use a separate process. The attendance configuration should reflect the actual policy and user instructions. An employee cannot reasonably follow a procedure that exists only in an administrator’s interpretation of a software setting.
Document which actions users are expected to take and how exceptions are reviewed. If a person forgets a required break-related event, identify the evidence and approval needed to resolve it. If the business uses planned break rules, confirm how those rules interact with unusual shifts or authorised changes. Do not infer that every interval between device events is automatically an unpaid break or approved working time.
Test a representative day containing the relevant break pattern. Review both the on-screen result and the export that the receiving team will use. Labels and units should make the interpretation clear. Where the chosen software cannot represent the policy reliably, address that limitation openly rather than compensating with untracked spreadsheet edits after every reporting run.
Create an exception queue with useful categories
An exception queue should help reviewers decide what to investigate next. Useful categories can include a missing entry, missing exit, unexpected location, roster mismatch or an unresolved employee identity. The exact categories depend on the supported application and business process. Avoid combining every unusual record into one vague error list that gives supervisors no practical starting point.
Assign an owner to each category or review stage. A supervisor may confirm operational facts, while HR decides how an approved policy applies. An administrator may investigate device or synchronisation issues. These are different responsibilities. Routing everything to one person can delay resolution and encourage decisions outside that person’s authority.
Set a review timetable that fits the reporting period. Daily or regular review may be appropriate where volume is high, while a smaller team may use another agreed rhythm. The essential point is to resolve known issues before the final export deadline where possible. A queue that is ignored until the last day creates pressure for poorly supported corrections and makes the whole process harder to trust.
Resolve a missing punch without inventing a record
A missing punch is a request for investigation, not a complete explanation of attendance. The employee may have forgotten an action, used an unavailable terminal, worked elsewhere or been assigned an incorrect schedule. The supervisor should establish the circumstances through the organisation’s approved process. Do not treat absence of one biometric event as the sole basis for a pay or disciplinary decision.
Define what information a correction request needs. It can identify the employee, date, missing action, explanation and relevant supporting reference. Keep evidence proportionate and restrict access to those responsible for the review. The process should not require employees to circulate sensitive personal information widely merely to correct an ordinary attendance discrepancy.
Record the decision as a reviewed adjustment, preserving the distinction from an original device event. If the explanation remains unresolved, use the organisation’s approved unresolved status or escalation route. A fabricated-looking precision, such as entering an unsupported exact time simply to close the queue, can create a misleading record. The system should make the decision understandable rather than hide uncertainty behind a neat total.
Investigate duplicate and repeated punches
Repeated identification events can arise for different reasons, including a user retrying an action or records being received through a recovery process. Reviewers need enough information to distinguish the cause. Ask which event identifiers, device details and timestamps are available in the proposed system and how they can be examined by authorised administrators.
Do not delete repeated events merely because they are close together. The organisation’s workflow may intentionally include several actions, and the software’s interpretation must be tested. If a duplicate-handling rule is proposed, ask what it does, which events it affects and whether the original records remain available. Keep the approved configuration and its rationale in the handover documentation.
Include a recovery scenario in testing where relevant. If previously unavailable events arrive later, check whether the system identifies records already received and whether the report changes. The intended result should be demonstrated with the chosen equipment and software. This avoids assuming that every device model or connection method prevents duplication in the same way.
Handle off-site work and unusual clocking locations
Some employees work away from the normal terminal. The attendance process needs an authorised way to represent that work without pretending a biometric event occurred at the usual location. Describe the relevant situations during discovery: field assignments, training, temporary transfers or work at another branch. The appropriate process depends on the organisation and the capabilities selected.
Define who authorises the assignment and how the reviewer receives the information. A supervisor’s confirmation should identify the relevant period and purpose through an approved record. If the software offers alternative attendance methods, assess their suitability and access controls before using them. Do not assume mobile or remote clocking is included simply because another product supports it.
Unexpected location records also deserve context. An employee may legitimately use another branch, or a profile may have an incorrect assignment. The reviewer should investigate before interpreting the event as misconduct. Keep location information limited to the operational purpose and applicable organisational requirements. The goal is a consistent attendance explanation, not unnecessary monitoring beyond the agreed scope.
Treat overtime as a decision, not merely elapsed time
Recorded presence before or after a scheduled shift does not automatically establish approved overtime. The organisation must define who authorises additional work and who confirms that the relevant conditions were met. HR and the responsible professional advisers should determine the applicable rules. The attendance software should implement approved requirements rather than create them through a default calculation.
Separate the observation from the approval. A report may identify time outside a schedule for review, while an authorised person records the decision. Ask the supplier how the selected system supports that distinction and what happens when approval is pending or rejected. If the software cannot provide the required workflow directly, define the controlled alternative before implementation.
Ensure the receiving team understands the exported fields. A column labelled additional time may not mean approved overtime, and a decimal value may not mean the same thing as an hours-and-minutes display. A sample review with payroll or HR can expose these misunderstandings early. Clear labels and documented meanings are more useful than a promise that the system calculates everything automatically.
Use a hypothetical overtime review to test responsibilities
Imagine a hypothetical team completing an authorised stock count after its planned shift. Device events show that several employees remained on site, but one employee stayed for an unrelated personal reason. The example is illustrative and makes no statement about legal entitlement or payment. It shows why presence alone should not replace the organisation’s approved review process.
The supervisor should identify the authorised activity, the employees involved and the relevant supporting record. HR applies the organisation’s approved rules, and the final attendance output distinguishes the reviewed outcome from raw events. The process should also handle a worker whose event is missing but whose participation needs investigation. Neither automatic acceptance nor automatic rejection resolves that question responsibly.
During testing, ask the supplier to show pending, approved and unresolved cases. Verify who can change each status and what evidence remains afterwards. The receiving team should see only the interpretation it expects, with exceptions communicated clearly. This worked scenario tests authority and traceability, which are as important as the arithmetic displayed by the application.
Manage leave and calendar exceptions through approved information
The attendance calendar may need to reflect approved leave, rest days or other designated periods. The organisation should provide the authoritative information and decide who maintains it. A missing clocking event on an approved non-working day should not be interpreted using an ordinary shift rule without context. Equally, a calendar label should not be changed casually to make a discrepancy disappear.
Confirm how the selected system combines calendar information with roster assignments and original events. Use representative cases chosen by HR rather than making legal assumptions about particular dates or entitlements. If the software offers predefined calendars, review them before use. A supplied template is not proof that it matches the organisation’s approved requirements.
Record changes that affect a reporting period, especially after review has begun. Ask whether applying a revised calendar recalculates earlier results and how users can identify the change. The same control principle applies throughout attendance management: current administrative information should not silently alter a previously approved interpretation without a traceable review.
Plan employee changes across the reporting period
New starters, transfers, returning employees and departures can create reporting exceptions if identity and assignment changes are not coordinated. Define which team supplies the authoritative instruction and when it takes effect. The administrator should not infer the correct status from a manager’s informal message that lacks a date or an employee identifier.
For a transfer, confirm the new supervisor, location and roster while preserving the earlier context. For a departure, define the authorised process for disabling future use and handling historical records under organisational policy. Do not assume deleting a profile is the correct way to end access. Ask the supplier to demonstrate the distinction between deactivation, historical reporting and any supported removal functions.
Test a change partway through a reporting period. The report should let reviewers understand which assignment applied to each relevant date. If the application has limitations, document the approved handling method before production use. This helps avoid a monthly total that looks plausible while placing records under the wrong team or interpreting earlier work through a newly assigned shift.
Prepare for connectivity interruptions and delayed events
Confirm the proposed terminal’s behaviour when its connection is unavailable. Ask about supported event storage, relevant limits, recovery and the information administrators receive when communication resumes. These capabilities differ by model and configuration. A general statement that biometric devices work offline is not enough to establish how your complete reporting process will behave.
Maintain an approved contingency for times when normal clocking is unavailable. Record the affected location and period so reviewers can distinguish a system interruption from individual missing actions. The contingency should have clear ownership and proportionate evidence requirements. Staff should know whom to contact rather than improvising several incompatible records that later need reconciliation.
After recovery, check delayed and potentially repeated events before closing the period. Compare the affected population and dates with the interruption record. Where recovered events change an already reviewed outcome, follow the approved reopening process. Restoration of connectivity is only one stage; the reporting team still needs to establish that the final records have been reconciled appropriately.
Close reporting periods with a named approval
Define when a reporting period becomes ready for final review. The checklist can include the current roster, resolved exceptions, identified unresolved items and reconciliation of known interruptions. The exact process should fit the organisation, but somebody must own the decision that the report is ready to pass to the receiving team.
Identify the approved version with a period, generation time and responsible reviewer using the supported tools and document process. Avoid several files with similar names circulating as though they are equally final. If an exception remains open, communicate its status explicitly rather than quietly excluding the employee or inserting an unverified value to complete the export.
Ask how the application supports period closure and whether later edits are restricted or traceable. Where locking is unavailable, establish another controlled procedure. Do not claim that a file is protected simply because it has been exported. The control comes from the agreed access, approval and version process, together with the capabilities the selected system actually provides.
Reopen a period without losing its history
A late correction may be legitimate, but it needs a defined process. Record who requested the change, the affected employee and period, the reason and the authorisation. Review the original approved output before applying the correction so the team understands what is changing. A fresh export should not silently replace the version already used by another department.
Notify the receiving team through the approved internal process and identify the revised version clearly. Explain whether the change affects one record or a wider recalculation. Payroll and HR should decide how the revision is handled in their own controlled workflow. The attendance administrator should not assume that replacing a file automatically reverses every downstream action.
Retain the relationship between the original and revised results according to the organisation’s records policy. This supports later questions without encouraging uncontrolled duplication of sensitive files. Test the reopening procedure during implementation. It is better to discover a version-control problem using sample records than during a time-sensitive correction after a reporting deadline.
Validate reports with the people who receive them
A technically readable export can still communicate the wrong meaning. Review employee identifiers, dates, time formats, units, headings and status labels with the receiving team. Include ordinary days, overnight work and reviewed exceptions. Ask users to explain what they believe each field means; differences in interpretation often become visible through that exercise.
Reconcile a controlled sample back to original events and approved decisions. The sample should cover the important patterns rather than only the easiest records. Document who checked it and which questions were resolved. If direct integration is requested, specify the receiving product and version, supported interface and handling of rejected or missing records. Compatibility must be confirmed for the actual setup.
Keep attendance approval distinct from financial calculation. For further planning, use the ZES biometric attendance and payroll integration guide. The reporting handover should make the approved interpretation clear while leaving the receiving team’s authorised calculations and controls in place. A clean transfer is one that preserves meaning as well as moving data.
Train primary and backup administrators on real cases
Training should follow the work administrators and supervisors must perform. Include assigning an approved roster, locating original events, routing an exception, recording an authorised correction and preparing a reviewed report. The available actions depend on the product and user role. Avoid a demonstration in which only the installer operates the system while users watch menus move past.
Train at least a primary and backup administrator for the agreed responsibilities. Supervisors also need guidance on what evidence they provide and which decisions they can make. A short written procedure helps maintain consistency when people change roles. Sensitive information should remain accessible only to authorised users, with separate responsibilities where the organisation requires them.
Finish training with representative tasks performed by the customer’s users. Record unresolved questions and arrange the agreed follow-up. The objective is not simply attendance at a session; it is evidence that the team can operate its reporting process. Clear handover documents and support contacts make that capability more durable than a one-time explanation remembered by a single person.
Check timestamps and report boundaries during acceptance
Agree which time convention the installation uses and how it is displayed in the application and exported reports. Where several devices or locations are involved, ask the supplier to demonstrate consistent interpretation of a controlled test event. A difference between displayed timestamps can confuse reviewers even when the underlying event is present. The administrator should know whom to contact about an apparent discrepancy rather than changing device settings independently.
Test the reporting boundaries as carefully as the middle of a period. Include an overnight shift spanning a month end and an event received after the usual review deadline. Confirm which selection criteria the report uses and whether a date filter includes the intended complete work period. The receiving team should understand any distinction between event date, shift date and approval date in the selected output.
Keep these tests in the acceptance record with the expected and observed results. If a later software change alters a relevant display or report, the examples provide a practical basis for checking the new behaviour. This is particularly useful when an upgrade appears minor but affects date formatting or selection rules. Familiar sample cases let the organisation investigate the effect without using live payroll processing as its first experiment.
Prepare an implementation brief for ZES
Provide employee numbers, locations, current device and software details, shift patterns and a sample report with personal information removed. Include overnight work, breaks, field assignments and the exceptions your team struggles to resolve. Identify the HR owner, reporting reviewer and receiving system. These details support a scope based on operational needs rather than a terminal’s advertised employee capacity alone.
Read the ZES biometric attendance system guide for the wider implementation journey. When you request a ZES assessment, ask for a demonstration covering your difficult cases, a clear configuration record and a practical acceptance plan. The intended outcome is a biometric time attendance process whose shifts, corrections and approvals remain understandable from the original event through to the final reviewed report.