Section 1
The artifact: the rota redesign worksheet
Run your current rota through four passes, in order. Each pass targets a different source of per-head waste. Do them on paper against next week's actual rota, not in the abstract. Pass 1: Find the thin slices List every shift by person and hours. Flag every worker whose total weekly hours are so low that consolidating them into a single fuller shift, for one person instead of two, would cover the same demand. The classic thin slice is two people each doing a short mid-week shift that one person could do as one longer shift. Consolidating two thin heads into one fuller head removes a head from the NI count while keeping the hours worked, and often the remaining worker welcomes the fuller, more predictable shift. This is band consolidation: you are moving hours off marginal heads and onto fewer, better-utilised ones. The discipline: consolidate the slices that were only ever thin because it was administratively easy, not because the demand needed two separate people. Keep the slices that exist because two people genuinely have to be in two places. Consolidation is not about firing; it is about stopping the practice of splitting one job across several heads when one head would do it better and cheaper. Pass 2: Cross-train to remove a shift, not a person The second source of per-head cost is single-skilled roles that force you to roster a separate head for each task. If your quiet Tuesday needs someone on the till, someone in the kitchen, and someone on the floor, but the volume only justifies one and a half people, cross-training lets one person cover two of those roles during the trough. That removes a rostered head from the quiet period without removing the capability, because the remaining staff can flex across tasks. Cross-training is the highest-return lever here because it cuts cost in the troughs while improving resilience in the peaks: a cross-trained team covers absence and rushes without you calling in an extra head. Map it explicitly. For each role, list who else can already do it and who could learn it in a week. The gaps are your training plan. The goal is that no single quiet-hour task needs its own dedicated head. Pass 3: Decide where automation removes a shift versus moves the cost Automation is a lever here, but only when it removes a shift rather than relocating the cost to a subscription. The test is simple and worth being strict about. A booking system that lets one host cover what used to need two during service removes a shift, real saving. A self-order kiosk that replaces a taken order but still needs the same floor staff to run food and clear tables moves the cost from wages to a lease and often does not remove a head. Before you buy any tool sold as a labour saver, ask: does this let me roster one fewer head in a specific shift, or does it just add a fixed monthly cost on top of the same rota? Only the first is a response to a per-head charge. The second is a new fixed cost wearing an efficiency label.
Section 2
The artifact: the rota redesign worksheet (continued)
Pass 4: Re-cost the redesigned rota Rebuild the rota's labour cost on the new NI rules: for each remaining head, 15 percent of pay above 5,000 pounds, plus wages at the current National Living Wage of 12.21 pounds an hour, up 6.7 percent in April 2025 (GOV.UK). Compare total labour cost and total heads before and after the three passes. You are looking for two things: fewer heads carrying NI, and the same or better service coverage. If you have removed heads but service coverage dropped, you have cut, not redesigned, and you will pay for it in overtime or lost sales. Redesign holds coverage while reducing the head count that carries the charge.
Section 3
Why this beats cutting hours or pay
Cutting hours across the board shrinks capacity everywhere, including in the peaks that make your money, and it usually pushes the removed hours into overtime or worse service. Cutting pay, where it is even legal given the wage floor, damages retention in a labour market where replacing a trained worker costs weeks of productivity. The redesign avoids both by targeting the actual cost driver: the number of heads in the taxed band, not the total hours or the pay rate. You end with roughly the same people working slightly fuller, more predictable shifts, cross-trained to flex, and a lower per-head NI bill, without gutting either capacity or morale. Check the Employment Allowance before you assume the redesign is urgent. It rose to 10,500 pounds in April 2025 with the old eligibility cap removed, so many smaller employers offset a large share of employer NI and some pay none (The Access Group). If your allowance absorbs most of the threshold effect, the redesign becomes an optimisation you do at your own pace rather than a rescue you do this month.
Section 4
The models under the redesign
The worksheet rests on two disciplines. The first is comparative statics: identify the single variable the policy moved, heads in the taxed band, and change only that, holding service level constant. This is what stops the redesign from sliding into a general cut, because the constraint "hold coverage" is built into pass four. Its limit is that people are not variables; a consolidation that looks clean on the sheet can fail if the remaining worker cannot actually cover the combined shift, so test the redesign against real people, not just hours. The second is a mechanism-design read: the policy was structured to charge per head, so the rational response is to reduce heads-in-band without reducing labour. But mechanism design cuts both ways. If you optimise purely to dodge the charge, you can end up with a rota that is technically cheaper and operationally brittle, one absence away from collapse. The cross-training pass exists to counter that, building slack back in as capability rather than as spare heads.
Section 5
What the redesign cannot fix
It cannot fix a business whose margin does not reach breakeven at any realistic volume on the new cost function. If the rebuilt unit economics are underwater no matter how efficient the rota, that is a pricing, financing or footprint decision, not a rota problem, and reshaping shifts only delays the real one. The redesign also cannot manufacture demand; a fuller, leaner rota still needs enough covers to justify the heads on it. And it says nothing about the next policy move. It makes your labour model efficient on today's rules, not immune to tomorrow's.
Section 6
The fitness test
You have redesigned rather than cut if, after the four passes, your rota covers the same service with fewer heads carrying employer NI, your team is cross-trained enough to flex across the troughs, and any automation you added removed a specific shift rather than adding a fixed cost. If you can show that, you have answered a per-head charge with a per-head fix. You have merely cut if you have fewer people but thinner coverage, more overtime, or a rota that breaks the first time someone calls in sick. That trades a visible NI saving for hidden costs that usually exceed it. The operators who absorb a labour-cost shock without hollowing out the business are the ones who redesigned the shape of the work, because the policy taxed the shape, and the shape is the thing they could actually change.