Planning Foundations
A free five-day email course in the fundamentals of contact centre workforce planning. The course starts next Monday — then one short, practical lesson each weekday, Monday to Friday: the core ideas a good planner actually uses, with a calculator or article to go deeper on each. No fluff, no sales pitch, unsubscribe any time.
Starts next Monday · delivered via Buttondown · we’ll never share your address · you’ll also get the fortnightly newsletter.
What you’ll learn
Prefer to read it now? The five lessons are below in full. Sign up now and your five lessons begin next Monday, one each weekday, plus the ongoing newsletter.
Why you can’t just divide work by people
The first thing planning teaches you is that staffing isn’t arithmetic. Two hundred calls an hour at five minutes each is about 16.7 hours of work — the “offered load” — but you can’t staff 17 agents and expect a good service, because calls arrive at random, not in a tidy queue. Some intervals get a rush, some go quiet, and a queue that’s busy “on average” spends a lot of its time underwater. Erlang is the maths that turns offered load into the agents you actually need to hit a service target — and it always asks for more than the raw work, with the cushion growing as your service target rises and shrinking as the operation gets bigger.
Net versus gross: the people who aren’t on the phones
Erlang tells you how many agents you need on the phones. It doesn’t know that your people also take breaks, go to training, attend meetings, take holiday and call in sick. That’s shrinkage, and it’s the bridge between the “net” requirement Erlang gives you and the “gross” headcount you actually have to roster. At 30% shrinkage, every 10 agents you need on the phones means roughly 14 on the books. Get shrinkage wrong — or treat it as a single flat number that only ever drifts upward — and a roster that looks fully covered turns into an understaffed morning.
It’s not how many people — it’s when
Having enough people on the day is only half the job; having them at the right times is the other half. Demand isn’t flat — it has a shape across the day, with peaks and troughs — and a roster that’s the right size overall can still be painfully wrong interval by interval, over-staffed at lunch and drowning at ten. The art of scheduling is laying shifts over that requirement curve so coverage traces it as closely as people’s real lives allow. Most service problems that look like “we need more heads” are really shape problems: the same heads, in the wrong intervals.
A forecast is wrong — the question is how, and how much
Every forecast is wrong; that’s not failure, it’s the nature of predicting the future. What matters is measuring the error honestly so you can act on it. The trap is the wrong metric: a forecast that’s spot-on at the daily total can hide big interval misses that wreck the roster, and a single accuracy number can flatter a forecast that’s consistently biased one way. The measures worth watching tell you not just how big the error is but which way it leans — because a forecast that’s always low is a different, more fixable problem than one that’s just noisy.
The best plan in the world, ignored, helps no one
The final foundation isn’t a technique — it’s credibility. A planning function that produces brilliant forecasts nobody trusts, or rosters operations quietly overrides, might as well not exist. Credibility is built the unglamorous way: forecasting in honest ranges rather than false precision, owning your misses before someone else points them out, explaining the “why” behind a number in language the operation speaks, and being the person in the room who can size the prize of a decision. Planning is half maths and half trust, and the trust half is the one that decides whether any of the maths gets used.
Ready to start?
Pop your email in and your first lesson lands next Monday. Want the whole craft, not just the foundations? Browse the full article library, the calculators, or the free Academy.