Designing for the average, not the curve
The right total in the wrong shape
The most common scheduling failure is the schedule that produces exactly the daily FTE the capacity plan said was needed, distributed flat across hours when the demand curve has a peak that’s 3× the trough. The operation runs over-staffed in the troughs and under-staffed in the peaks. The daily-average dashboard reports everything as on-plan. The intraday SL chart tells a different story.
The failure is counter-intuitive. A planner who delivered the right number of FTE-hours got the right answer in total — how can the schedule be wrong? It’s wrong because contact-centre service isn’t a daily total; it’s an interval phenomenon. The customer who phones at 10am wants service at 10am, not on average across the day.
Why flat coverage curves persist
Three forces conspire.
Shift patterns are coarse. If the available shift options are 8am–4pm, 9am–5pm, and 10am–6pm, the coverage you can produce is constrained by what those shifts add up to. A demand curve that peaks at 10am needs disproportionately many of the early shifts, which most operations don’t have enough of.
Team-leader pushback against staggered starts. Teams prefer to start at the same time. The TL team huddle works better when everyone is there. The cultural norm is that 9am is when the team starts, and the schedule that moves agents to earlier starts gets reported as “unfair.”
The daily total is what gets reviewed. The capacity plan operates monthly. The schedule review compares daily total FTE to daily total need. The intraday curve isn’t in the review pack. The misalignment doesn’t surface in the conversation that should catch it.
The design moves that fix it
A deliberate part-time layer. Three-hour and four-hour shifts sized to cover the peaks. A 4-hour shift from 9am–1pm sits exactly across the morning peak in most operations and is one of the cheapest forms of capacity. See the part-time layer.
Staggered start times. If the demand curve ramps up from 8am to 10am, the schedule should ramp up the same way. 20% of the team starting at 8am, 40% at 8:30, 40% at 9am gives a smoother coverage build than 100% at 9am. The TL pushback is real but manageable with the right rhythm (team huddle is still done at the same time; the early-shift agents start on contacts first).
A small flex pool. A handful of agents whose start time and shift length flex within a window each week, tuned to whatever the forecast shows. This absorbs the week-to-week curve variation without redesigning the whole rota.
A late layer for the closing trough. Covered in the separate schedule-past-closing article. A small group rostered to 15–20 minutes past official close fills the under-staffed final interval that almost every operation has.
How to see the gap
The diagnostic is mechanical: plot scheduled FTE by interval against forecast required FTE by interval for the next week. The gap is usually larger than planners expect. The classic shape: 30%+ under-coverage at the morning peak and 30%+ over-coverage in the late afternoon, on a schedule that perfectly matches daily total.
Most operations don’t produce this view because the WFM platform reports schedule efficiency at daily or weekly granularity. Building the interval view in Excel for one week takes an hour and produces an argument for redesigning the schedule that’s hard to dismiss.
The honest commercial case
The redesign costs something. Shift pattern changes affect agents, TL pushback is real, and the implementation work isn’t free. The commercial case has to be made specifically: at the current schedule the SL miss costs £X in [overtime / outsource / complaint volume / lost CSAT]; the redesigned schedule produces a 4-point SL lift at zero net FTE; the implementation work is one quarter of effort. That conversation gets a different answer than “we should redesign the schedule.”
Conclusion
The flat coverage curve is the most common scheduling failure and the one that’s most invisible on the metrics most operations actually look at. The fix is structural: a part-time layer, staggered starts, a small flex pool, and a closing-trough layer. None of it is technically difficult; all of it is operationally and culturally challenging. Operations that make the redesign find a 3–6 point SL lift at the same FTE; operations that don’t spend the same FTE delivering worse service.
Next in the series: The shrinkage assumption that’s always wrong.
Pair this with the part-time layer, scheduling past closing time, and multi-skill scheduling.