Outsourced CompanyPowered by StealthAgents

Philippines staffing research · Published · Updated

How a second approver changes an outsourced queue's cycle time

Adding an approver looks like a control. It also adds a wait, a handoff, and a new place for cases to stall. Measure the decision path before changing the queue.

Two colleagues reviewing a queue on a shared screen in a quiet office

Research findings

A second approver is usually added after a mistake. The logic is simple: another person checks the work, so fewer errors reach the customer. That logic is sound for risk, but it ignores the effect on flow. Every approval adds a waiting period, a handoff, and a new opportunity for the case to lose its owner. The queue changes even when the work itself does not.

The first change is visible in the cycle time. A case that once moved from ready to done now moves from ready to waiting, then to review, then to done. The work is the same, but the clock now includes a period when nobody is actively handling the case. The average rises, and the cause is not a slower worker. It is a longer path.

The second change is a shift in the queue mix. Items that need approval accumulate while items that do not need approval continue. The approved queue looks busy and the waiting queue looks quiet. A dashboard that counts only open items will miss the growing share that is blocked on a decision. Segment the queue by state before drawing a conclusion.

The third change is a change in the reviewer behavior. An approver who receives a large batch at the end of the day cannot give each case the same attention. They may approve in patterns, reject in patterns, or defer the ambiguous cases. The quality of the approval falls precisely when the volume is highest. That is the opposite of what the control was meant to achieve.

The fourth change is a subtle transfer of ownership. When a case moves to an approver, the original coordinator may stop watching it. The approver may assume the coordinator is still responsible for the customer. In that gap, nobody answers the customer. The case is not lost, but it is also not owned, and the customer experiences the silence as neglect.

A second approver can still be the right choice. The question is not whether to add a control, but where the control creates the most value for the least wait. A high-risk decision may deserve a second signature. A routine correction may not. The queue should distinguish the two rather than applying the same path to every item.

Discussion

Start by mapping the actual decision path for a sample of cases. Record each state, the time spent in it, and the person who moved the case forward. Do not use the documented process if it differs from the observed one. The map should show where a case waits for a person rather than for work. Those waits are the candidates for redesign.

Then measure the decision wait separately from the total cycle time. Decision wait is the time a case spends waiting for an approval, a policy answer, or a judgment call. It is often the largest and least visible part of the delay. Once it is separated, the team can see whether the second approver is the cause or whether the queue was already waiting for something else.

Set a decision range for the approver instead of a blanket review. The range can say that corrections under a stated value or risk level are approved by the coordinator, while higher-risk cases go to the second approver. The range should be written and applied consistently. A range that changes with the reviewer mood is not a control.

Give the approver a dedicated review block and a way to clear the queue in one pass. A reviewer who checks cases between meetings will batch them, delay them, and rush them. A short block with a clear start and end is easier to protect and easier to measure. The block should be long enough for the expected volume and short enough to survive a normal week.

Implications and limits

Define what happens when the approver is unavailable. A named backup, a decision range, and a maximum wait are the minimum. If the backup cannot approve the case, the queue needs an escalation path rather than a longer wait. The absence of an approver should not turn into a customer-facing delay without a message.

Review the design after one month using the same sample method. Compare the decision wait, the error rate, and the rework count before and after the change. If the error rate fell but the wait doubled, the trade-off is real and should be stated. If neither improved, the second approver is adding cost without adding control.

The final test is whether the approver can explain the decision rule without reading a document. A control that lives only in a policy is applied inconsistently under pressure. A control that the reviewer can state in one sentence is more likely to hold. Write the sentence, test it on real cases, and revise it when the queue changes.

Mapping the approval path

Draw the path as a sequence of states rather than a list of roles. Each state should have an entry condition, an owner, and a maximum expected duration. A case that exceeds the duration is a signal, not a failure of the person. The map is useful when it shows where the case waits, not just who touches it.

Validate the map with three recent cases, including one that went smoothly and one that stalled. Compare the recorded path with the map and note where they diverge. The divergence is usually the informal step that makes the process work. Decide whether to formalize it or remove it before adding another approver. Where the map and the case disagree, trust the case and revise the map.

Measuring wait instead of volume

Use four measures: arrival count, active work time, decision wait, and total cycle time. Active work time is the period when someone is actually handling the case. Decision wait is the period when the case is waiting for authority. The difference between total and active is the friction the queue carries. Collect them for the same cases so the comparison does not mix different work.

Report the measures by segment and by week. A total average hides the segment that changed. If the approval queue grew while the routine queue stayed flat, the average may barely move. Segmenting the data makes the effect of the second approver visible and gives the review a specific question to answer. A weekly view also shows whether the wait is stable or getting worse.

Designing a decision range

Write the range in terms of the decision, not the job title. State what can be approved, what evidence is required, and what must be referred. Use examples for the boundary cases so the reviewer can apply the rule without a debate. A range with clear examples reduces both delays and inconsistent decisions. The reviewer should be able to apply it without asking who owns the case.

Test the range on past cases before publishing it. Count how many would have been approved, referred, or rejected under the new rule. If most cases fall in the referred group, the range is too narrow to reduce the wait. If almost none are referred, the control may be too loose for the risk. Adjust the boundary until the split matches the risk the business is willing to carry.

When a second approver helps

A second approver helps when the cost of an error is high, the decision is infrequent, and the reviewer has enough context to add information. It helps less when the work is routine, the error is easy to correct, and the reviewer only repeats the first check. The design should match the control to the risk. When those conditions are absent, a simpler check is usually better.

The strongest version of the role is a reviewer who can change the outcome, not just sign it. That means authority to approve, reject, or request a different remedy within a stated range. A reviewer without authority adds a step without adding a decision. The queue feels the cost and the customer sees the delay.

A review checklist for the approver

Give the approver a four-item checklist for each case: the decision requested, the evidence attached, the customer promise, and the authority to act. The checklist should take less than a minute and should be visible in the case. It prevents the reviewer from reconstructing the request and keeps the approval focused on the decision. Keep the checklist beside the case so the reviewer does not have to switch tools.

Review the checklist monthly and remove any item that does not change the decision. A field that nobody reads adds time without adding control. Keep the evidence requirement specific to the risk, so a routine case does not carry the same burden as a high-value exception. The checklist should make the approval faster as well as safer.

Questions

How do we know the second approver is worth the wait?

Compare the error rate and the decision wait before and after the change. If the error rate does not improve, the extra step is cost without control.

Should every case go to the approver?

No. Use a decision range so routine cases stay with the coordinator and higher-risk cases go to the approver. A blanket review creates a queue for work that does not need it.

What is a reasonable decision wait?

Set it from the customer promise and the risk of the case. A safety issue and a routine correction need different limits, and the limit should be written.

Who covers the approver?

A named backup with the same range and evidence requirements. Without a backup, the approval step becomes a single point of failure.

Philippines staffing intake

Define the position before hiring begins.

Share the tasks, tools, schedule, and approval limits for your Filipino team member. The intake turns those details into a practical staffing brief.

Contact Us