Planning look-up tables
The tables planners keep reaching for — Erlang C staffing at the conventional 80/20 target, and the shrinkage uplift from “agents on the phone” to “agents on the schedule.” Printable, with the caveats attached so they travel safely.
Erlang C staffing — AHT 300 seconds, target 80% in 20s
Agents required per half-hour interval. ASA is the average speed of answer the staffed number delivers; occupancy is the proportion of agent time spent handling.
| Calls / 30 min | Offered Erlangs | Agents required | ASA (s) | Occupancy |
|---|---|---|---|---|
| 10 | 1.7 | 4 | 13 | 42% |
| 20 | 3.3 | 6 | 17 | 56% |
| 30 | 5.0 | 8 | 17 | 62% |
| 50 | 8.3 | 12 | 14 | 69% |
| 75 | 12.5 | 17 | 11 | 74% |
| 100 | 16.7 | 21 | 16 | 79% |
| 150 | 25.0 | 30 | 15 | 83% |
| 200 | 33.3 | 39 | 13 | 85% |
| 300 | 50.0 | 57 | 11 | 88% |
| 400 | 66.7 | 74 | 12 | 90% |
| 500 | 83.3 | 91 | 12 | 92% |
| 750 | 125.0 | 134 | 11 | 93% |
| 1,000 | 166.7 | 176 | 12 | 95% |
| 1,500 | 250.0 | 261 | 10 | 96% |
| 2,000 | 333.3 | 345 | 11 | 97% |
Erlang C staffing — AHT 450 seconds, target 80% in 20s
| Calls / 30 min | Offered Erlangs | Agents required | ASA (s) | Occupancy |
|---|---|---|---|---|
| 10 | 2.5 | 5 | 23 | 50% |
| 20 | 5.0 | 8 | 25 | 62% |
| 30 | 7.5 | 11 | 23 | 68% |
| 50 | 12.5 | 17 | 17 | 74% |
| 75 | 18.8 | 24 | 15 | 78% |
| 100 | 25.0 | 31 | 13 | 81% |
| 150 | 37.5 | 44 | 15 | 85% |
| 200 | 50.0 | 57 | 16 | 88% |
| 300 | 75.0 | 83 | 15 | 90% |
| 400 | 100.0 | 109 | 14 | 92% |
| 500 | 125.0 | 135 | 13 | 93% |
| 750 | 187.5 | 199 | 12 | 94% |
| 1,000 | 250.0 | 263 | 11 | 95% |
Reading the Erlang tables honestly
Two things the tables show that slide decks routinely hide. First, the economies of scale: at 10 calls a half-hour you staff at 42% occupancy to hit target; at 1,000 calls you hit the same target at 95%. That’s why pooling queues and skills is usually worth more than any scheduling cleverness — and why occupancy targets copied from a bigger operation will break a smaller one. Second, the model assumes nobody abandons and nobody retries — in heavy load Erlang C overstates the agents you need, because real callers leave the queue. If abandonment matters to your numbers, use Erlang A.
Deeper: the Erlang white paper · Erlang to Excel · the Erlang explorer
Shrinkage uplift — from “on the phones” to “on the schedule”
Erlang gives you agents handling contacts. People take leave, train, coach, brief and go off sick — so the schedule needs more. Scheduled FTE = required ÷ (1 − shrinkage).
| Agents required (on phone) | @ 20% | @ 25% | @ 30% | @ 35% | @ 40% |
|---|---|---|---|---|---|
| 10 | 13 | 14 | 15 | 16 | 17 |
| 25 | 32 | 34 | 36 | 39 | 42 |
| 50 | 63 | 67 | 72 | 77 | 84 |
| 100 | 125 | 134 | 143 | 154 | 167 |
| 200 | 250 | 267 | 286 | 308 | 334 |
| 500 | 625 | 667 | 715 | 770 | 834 |
The two classic shrinkage mistakes
Multiplying instead of dividing (100 × 1.30 = 130 scheduled at 30% shrinkage — wrong; 100 ÷ 0.70 = 143 — right), and applying an annual average shrinkage to every interval. Shrinkage has a shape: training lands in blocks, leave clusters around school holidays, sickness peaks in winter. Forecast it by period like everything else.
Deeper: Shrinkage: the planner’s hardest input · the shrinkage calculator · Shrinkage, Honestly (white paper)
Want different assumptions?
These tables fix AHT and the service target so they fit on a page. Your operation almost certainly differs — that’s what the Erlang C calculator, the staffing-sensitivity tool and the forecaster are for. For typical ranges of the surrounding metrics (shrinkage, occupancy, attrition, adherence), see the planning benchmarks page.