Fifteen Business Days
You lead a two-analyst financial review supporting a grand jury that convenes in 15 business days. Investigators flagged 1,284 transactions. A deduplication pass removes 22% as duplicate entries. Each analyst completes a documented review of 40 transactions per business day, a rate that includes write-up. Eight percent of the surviving transactions require second-level review by a supervisor who is available 6 days and can complete 25 per day. A recurring system outage is forecast to cost 3 of the 15 available days. No overtime is approved. One contract analyst could be added, but requires 5 days of onboarding before producing at half the standard rate.
1. Is the deadline achievable? Show the arithmetic for the base case and the outage case.
Read the worked answer
Volume after dedupe: 1,284 x 0.78 = 1,001.5, so round up to 1,002 transactions.
Base case (15 days, two analysts): 2 x 40 = 80 per day. 1,002 / 80 = 12.5 days, so the work clears in 13 working days with about 2 days of slack (capacity 1,200, headroom 198).
Outage case (12 usable days): 80 x 12 = 960. Shortfall = 1,002 - 960 = 42 transactions, roughly one analyst-day and a fraction (42 / 40 = 1.05). The deadline fails by about half a day of work.
Second-level review: 1,002 x 0.08 = 80.2, so 81 items. Supervisor capacity = 25 x 6 = 150. Sufficient, with 69 slots of slack - the supervisor is not the binding constraint.
The binding constraint is first-level throughput, and it misses by 42 items, not by a comfortable margin and not by a catastrophic one. That distinction drives the remedy.
2. Given the shortfall, rank three remedies and state what you would not sacrifice to close it.
Read the worked answer
1. Add the contract analyst now. Onboarding days 1-5, producing days 6-15 at 20 per day = 200 items, which clears 42 with room to spare and absorbs a further outage. Cost: onboarding supervision and a slower start. This is the only remedy that creates durable capacity.
2. Defer a defined, documented slice of low-value items. 42 items - transactions below the reporting threshold, outside the charging period, or already captured by a cooperating witness's testimony. Document the criteria and the count before deferral, not after, so the decision is reviewable. Cost: a defense argument that you cherry-picked.
3. Both analysts work one additional day. 80 more items clears 42. Cheapest in volume, but it collides with "no overtime is approved" and, if done informally, it is the option most likely to produce sloppy write-ups.
What I would not sacrifice: items that prove an element of the charge, and anything the government is obliged to disclose. Deferring a transaction because it is boring is triage; deferring one because it undercuts the theory is a discovery problem. I would also not close the gap by lowering the documentation standard per item - the review's value is the write-up, not the checkbox.
3. Which assumption in this plan is most fragile, and how would you test it?
Read the worked answer
The 40-per-day rate, compounded by the uniform 22% dedupe. Two things can break it at once. First, duplicates are usually the easy items; if deduplication removes the simple matches, the surviving 1,002 are harder than the average that produced the 40-per-day figure, and real throughput falls below 40. Second, the 8% second-level rate was measured on the pre-dedupe pool in all likelihood, so 81 may be an undercount if the duplicates were disproportionately the clean ones.
Test: take actual, measured throughput at the end of days 3 and 6 - items completed, and hours consumed - and re-forecast rather than defending the original number. Simultaneously sample 100 surviving items and time them against 100 deduplicated items to check whether the rate holds on the post-dedupe pool.
The failure mode to avoid is discovering in day 12 that the rate was wrong, when the only remaining options are overtime that is not approved or a deferral nobody had time to document.