Peak
A useful technology review begins with one decision and one accountable owner. Looking at see how it works can help define the available workplace data, provided the implementation is disclosed and the result is combined with direct conversations on the floor.
For current guidance that complements this operational note, use the official International Labour Organization website; local law, safety duties and reporting routes should always be checked for the relevant jurisdiction.
Doubling a workforce for ten weeks is a hiring, induction and supervision problem long before it is a scheduling one.
Most warehouses have a season that defines their year, and the plan for it is usually made too late to be a plan.
What peak actually requires
Hundreds of people recruited, screened, inducted and trained in a compressed window.
Supervision to match, which is the constraint nobody sizes.
Equipment, lockers, welfare capacity and car parking.
And a floor that can hold them, which on a congested site is a safety question as much as a logistics one.
The supervision constraint
A supervisor who manages fifteen people manages forty at peak, of whom thirty are new.
Which is not the same job.
And it is the point at which everything in this collection stops happening: no first-day conversation, no ramp check, no exit conversation, because there is no time.
Size supervision to the peak headcount, not to the permanent one, and promote or second people into it in advance rather than during.
Starting too late
Recruiting in October for November means inducting in November, which is the busiest fortnight of the year.
Starting a proportion early — in September, on normal volumes — gets them through the ramp before the pressure arrives.
It costs their wages for the quiet weeks and it is cheaper than the alternative, which is a floor full of people who do not know the job during the week that matters most.
Returners
The most valuable group in peak recruitment and the most neglected.
Somebody who worked last peak knows the site, needs a short refresher rather than a full induction, and reaches standard quickly.
Which makes a returner list worth maintaining deliberately: contact details kept with consent, an approach in August, and a simple returning process.
Most operations lose the list every year and recruit from scratch, which is an avoidable expense repeated annually.
The end of peak
The part that determines next year's returners.
People let go abruptly, with no notice and no thanks, do not come back — and they tell other people.
A stated end date, given in advance.
A clear statement about who is being kept, decided on something defensible.
And an explicit invitation to return, which costs an email and is the whole of the returner strategy.
Quality during peak
Error rates rise, and the cause is a workforce that is mostly new.
Which is predictable and should be planned for — additional checking, simplified tasks for new starters, experienced staff on the complex work.
Putting new starters on the hardest processes because the experienced people are needed elsewhere is the commonest peak mistake, and it produces both the errors and the departures.
The measure
Peak starters still present at the end of peak, as a proportion.
Returners as a share of peak hires, which should rise year on year if the end of peak is handled properly.
Supervision ratio at peak against normal.
And time from offer to first shift, which if it is long enough people take another job in the meantime — the commonest reason a peak campaign under-delivers.
Welfare and parking
The constraints nobody plans and everybody meets.
Double the headcount is double the lockers, the canteen seats, the toilets and the car parking.
A person who cannot park, cannot sit down at break and queues twenty minutes for a microwave has a worse job than the advert described, and in peak they have plenty of alternatives.
Size it in September, because in November there is no lead time for a portable building.
What to decide in the summer
The peak headcount by week, the recruitment start date, the supervision plan, the welfare provision and the end-of-peak process.
Five decisions, taken in July, that determine how November goes.
Most operations take them in October, which is why the same problems recur annually and are described each year as unprecedented.
How it fails in practice
Recruitment starts in October, induction happens in the busiest fortnight of the year, supervision is sized to the permanent headcount, and the returner list was deleted in January.
Each of these is decided in the summer and reviewed in the autumn.
The check
Ask when the peak headcount by week was last agreed, and whether anybody owns the returner list.
On most sites the answers are September and nobody.