Research question: what escalation load would a proposed Philippines-based work lane create for its manager? Count each case routed for a decision, not each message sent. Classify the reason as missing information, authority boundary, sensitive data, policy exception, dependency, or consequence.

Use a complete observation period and preserve the population: arrivals, ordinary completions, escalations, unresolved cases, and owner response. Report escalation rate by work type and show the highest-consequence cases separately. Rare cases matter even when they barely affect a percentage.

Escalations are not necessarily a defect. A Filipino specialist who stops at a clear authority boundary may be operating correctly. The useful question is whether the escalation contains the source record, what was checked, the uncertainty, and the exact decision requested. Incomplete escalations add manager work without improving safety.

NIST’s govern, detect, and respond functions give a vocabulary for connecting signals to action. The owner should define which condition triggers a pause, who receives it, and what evidence is retained. “Use judgment” is not a substitute for an example of the decision that remains internal.

Analyze repeated reasons. If most escalations ask the same definitional question, improve the SOP or source record. If they cluster around permissions, redesign access. If they concern customer promises or money, preserve the owner boundary instead of optimizing the rate downward.

The SBA’s emphasis on responsibilities supports separating preparation from authorization. A support contributor can collect facts, draft options, and update a permitted status while the manager decides the exception. This division is especially important when the client and Filipino specialist work in different time zones.

A short period can miss low-frequency high-impact cases, and recorded escalations can undercount unsafe silent decisions. Review a small sample of ordinary completions for missed triggers. State what the evidence does not prove before using the rate to plan manager capacity.

Conclusion: expansion is plausible only when the owner can absorb the observed escalation load, the highest-consequence triggers are explicit, and repeated low-value escalations are converted into clearer inputs or boundaries.