03 / Decisions & Accountability
When Internal Expertise Is Free, the Queue Becomes the Price
When asking the in-house expert costs the requester nothing, the queue is the only thing rationing her time — and a queue rations by patience, not by what the decision is worth.
The example below is illustrative.
A manufacturer's senior engineer gets 28 review requests in a week. Some are required release checks. Some seek advice on an unfamiliar design. Others ask her to redo work someone else could have done. Every request lands in the same inbox, and a time-sensitive change waits behind an incomplete submission because the order is set by timestamp, not by operating consequence.
To the requester, sending the question costs seconds. The cost appears somewhere else — in the waiting time and broken concentration of everyone who depends on her. A queue can signal a true shortage. It can just as easily be rationing a free service that has no visible conditions for using it.
Measure the load in hours that can actually be used
Log intake over two normal weeks. For each request, record the decision it enables, the real deadline, whether the review is mandatory, the preparation required, the repeat visits, the specialist time, and the downstream wait.
Then do the arithmetic the inbox hides. Suppose the log shows 12 required release checks at 45 minutes each, eight prepared advice requests at 30 minutes, and eight incomplete requests that each return twice at 25 minutes a visit. That is nine hours of mandatory checks, four of prepared advice, and six hours forty minutes of repeat handling of work that arrived unready — 19 hours 40 minutes against an engineer who has maybe 14 hours for review after her other commitments. The queue exceeds capacity by nearly six hours a week, and a third of the load is pure rework.
Now separate the fixable from the structural. Pre-screening the incomplete submissions might cut the repeat handling to one visit of 25 minutes each — workload falls to 16 hours 20 minutes, still over capacity by two hours twenty. That residual is the real signal: some of the overload came from process design and vanishes with a fix, and some persists even after a plausible fix and has to be solved with capacity. Before treating any of it as a hiring case, check whether some submissions never needed her at all — and what happened to those decisions when they waited.
Test eligibility and allocation
For optional reviews, define what actually needs specialist input, what the requester must prepare first, and where routine questions can go instead. Offer a visible fast path for genuinely urgent issues, priced by the cost or risk of waiting — not by who shouts. Protect the mandatory approvals with a clearly identified route and enough reserved capacity to meet their obligations. Run it on one request class before expanding, and start the service clock when an eligible request arrives — but also record the total time the requester waited before it became eligible, so a new intake process cannot flatter its own statistics by moving the wait earlier.
A shorter queue achieved because people stopped asking necessary questions is a failure, not a win. Watch for rising defects and unrecorded workarounds alongside the turnaround numbers.
Decide whether supply is still binding
If well-prepared, eligible work still arrives faster than qualified people can clear it, quantify the shortfall and the delay's effect, then weigh adding or training reviewers, reserving predictable capacity, or shifting work to a validated alternative — against the cost and timing of a hire. A specialist who stays indispensable for necessary assurance can still be freed of avoidable interruptions.
This is one of the mechanisms behind apparent understaffing, catalogued in Six Reasons a Company Can Look Understaffed — where "we need another engineer" is often really "we have no conditions on using the one we have." And the capacity-landing test in Faster Isn't Sooner shows how to check whether a receiving function can turn faster arrivals into completed work; apply it after you change intake, following the revised requests through to actual release so a cleaner inbox cannot hide a later wait.
Take this to your team. Which decision is waiting right now because scarce specialist time is being spent on work someone else could do?