← ccPlanning Academy · Channel planning track

Email & processing at scale

Free visual lesson · about 5 minutes · short quiz at the end

ccPlanning academy · channel planning

Email & processing at scale

Planning a queue you’re allowed to keep waiting.

The big idea

Two clocks: arrival and completion.

For email and case work you track two flows — how fast items arrive, and how fast you complete them. Staffing is about keeping completion ahead of arrival enough to clear each item within its promised time. It’s production-line thinking, not Erlang.

The backlog

A buffer you manage, not eliminate.

Because email can wait, a backlog isn’t failure — it’s a tool. You can let it grow through a busy spell and clear it in the quiet, staffing to the average rather than the peak. The goal is a backlog that stays comfortably inside its SLA, not one at zero.

The metric

Age and SLA, not seconds.

The numbers that matter: the age of the oldest open item, the percentage cleared within the promised window (say 95% in 24 hours), and backlog size relative to daily capacity. Service-level-in-seconds is meaningless here; deadline conformance is everything.

The ageing trap

FIFO, or the oldest items rot.

Left to choose, agents cherry-pick quick, easy items and the hard ones age into breach. Without a deliberate work-allocation rule — oldest-first, or priority-weighted — your SLA quietly fails on a minority of items even while the average looks fine. Route the work; don’t let agents shop.

Forecasting it

Forecast arrivals and handle time, like anything else.

Email volume has trend, seasonality and day-of-week shape just like calls — often a Monday spike from the weekend. Forecast arrivals, forecast the handle time per item (which varies more than calls), and convert to the daily capacity needed to hold the SLA.

The hidden inbound link

A slow email backlog calls you.

Let email age and customers chase by phone or chat — turning one deferred contact into two live ones. The cheapest way to protect inbound is often to keep the email SLA healthy. Plan the channels together, not in silos.

The backlog as a tool

Staff the average, not the Monday

1,000 emails land Monday, ~500 a day after, SLA is 24 hours. Staff for Monday’s peak and you’re idle by Wednesday. Instead let the backlog absorb Monday and clear it across the week — you staff to the ~600/day average.

A backlog isn’t failure here; it’s the buffer that lets you smooth. The only rule: keep its oldest item comfortably inside 24 hours, and work oldest-first so nothing rots.

The takeaway

Throughput, backlog, deadline — and FIFO.

Track arrival vs completion, use the backlog as a buffer you keep inside its SLA, measure age and deadline conformance not seconds, allocate oldest-first so nothing rots, and remember a slow backlog generates inbound. Plan email like a production line with a promise to keep.

Now test yourself ↓

1 / 8

Slides done? Here’s the same idea in a bit more depth — the part worth keeping.

In depth: planning a queue you’re allowed to keep waiting

Email and case work run on two clocks — how fast items arrive and how fast you complete them — and staffing is about keeping completion ahead of arrival enough to clear each item within its promised time. That’s production-line thinking, not Erlang, and the mental shift it requires is the whole lesson: because the work can wait, a backlog isn’t a failure, it’s a tool. You let it grow through a busy spell and clear it in the quiet, staffing to the average rather than the peak; the goal is a backlog that stays comfortably inside its SLA, not one at zero.

The right metrics, and the trap

Service-level-in-seconds is meaningless here. The numbers that matter are the age of the oldest open item, the percentage cleared within the promised window (say 95% in 24 hours), and backlog size relative to daily capacity — deadline conformance, not speed of answer. And there’s a specific trap: left to choose, agents cherry-pick the quick, easy items while the hard ones age into breach, so without a deliberate allocation rule — oldest-first, or priority-weighted — your SLA quietly fails on a minority of items even while the average looks healthy. Route the work; don’t let agents shop.

Forecast it, and don’t silo it

Email still has trend, seasonality and day-of-week shape — often a Monday spike from the weekend — so you forecast arrivals and the handle time per item (which varies more than calls) and convert to the daily capacity needed to hold the SLA. The hidden link is that a slow backlog generates inbound: let email age and customers chase by phone or chat, turning one deferred contact into two live ones. Often the cheapest way to protect your phone queue is to keep the email SLA healthy — so plan the channels together, not in silos.

The principle to remember: throughput, backlog, deadline — and FIFO. Track arrival vs completion, use the backlog as a buffer kept inside its SLA, measure age and deadline conformance not seconds, allocate oldest-first so nothing rots, and remember a slow backlog generates inbound.

Email & processing at scale

Email & processing at scale

Two clocks: arrival and completion.

A buffer you manage, not eliminate.

Age and SLA, not seconds.

FIFO, or the oldest items rot.

Forecast arrivals and handle time, like anything else.

A slow email backlog calls you.

Staff the average, not the Monday

Throughput, backlog, deadline — and FIFO.

In depth: planning a queue you’re allowed to keep waiting

The right metrics, and the trap

Forecast it, and don’t silo it

Quick quiz