One Model, 20 Supermarkets: 14 Months to Pay Back — 111 at the Smallest Six

Standardising an estate on one cleaning robot is the right call and the estate number is a liar. Twenty stores, one PUDU CC1 Pro each, A$656,640 of capex — and a blended 14-month payback that no individual store in the estate actually experiences. Three stores get there in eight months. Six take 111.

The brief

The chain runs 20 single-level supermarkets in three formats. Every store closes by 10pm, the night crew works to a five-hour window before the 5am restock, and floors are washed six nights a week — 313 cleans a year. Night labour is the line under pressure: the loadings that apply after midnight are the reason a nominal award cleaner costs a fully-loaded A$35/hour on this shift, and that rate is the conservative end.

The estate wanted a single answer, not twenty. One model, one map procedure, one spares kit, one training package.

Nightly mopping is the reachable area at the planner's manual wet rate of 400 m²/h. Note how little of a supermarket a wheeled machine actually gets to — roughly 56% of a standard store's floor:

The shortlist

Grocery floors carry dry debris — produce trimmings, packaging dust, spilled product — under the wet mess. That rules out mop-only machines, which turn dry debris into slurry and push it to the kerbs. We need a wet machine that lifts solids before it wets the floor, sized on the same basis for every brand.

*Effective rate = manufacturer rated coverage (midpoint of any range) × uniform derate × 0.85 window duty — the same engine basis for every brand. Vendors publishing a bare theoretical maximum take the tighter 0.45 derate (here, only the CenoBots L3's 2,016 m²/h); practical ranges take 0.60. Unit prices A$32,832 / A$30,096 / A$32,832 / A$70,589 / A$44,597 ex-GST.

Three things fall out of that table.

CC1 and CC1 Pro size identically. They share the same 700–1,000 m²/h cover-cleaning spec, so on our uniform basis both answer "one unit per store" and both run a standard floor in ~3 h 20 min. The Pro's extra spot-cleaning mode (1,500–3,000 m²/h) is real, but it does nothing for the nightly full-coverage job and we give it no ROI credit. The estate chose it anyway, for a daytime problem: a produce or dairy breakage during trading is a slip hazard someone has to walk to, and a machine already on site can be sent instead. That preference costs A$54,720 across 20 units and 1.2 months of payback — the CC1 fleet pays back in 13.0 months against the Pro's 14.2.

CenoBots L3 is the cheap trap. Same fleet capex as the CC1 Pro, a much higher derated rate — and it only mops and scrubs, on a floor whose first problem is dry debris.

Gausium's machines are genuine peers priced for a different job. The Scrubber 50 Pro brings lift integration the CC1 family lacks, worth real money in a multi-level site and nothing in twenty single-level ones — for A$966,907 more. The Phantas is the quietest wet machine we list at 60 dB against the CC1 Pro's 70, which would matter if these stores ran the nightly clean during trading. They don't.

The money

One CC1 Pro sustains ~434 m²/h, which across a five-hour window is about 2,168 m² a night — enough for every format. That is the whole battery, not a comfortable slice of it: the 2,100 m² large store's 4.8-hour run fits inside the ~5-hour scrubbing runtime with nine minutes to spare. No mid-window recharge, and no headroom either. Across the estate that displaces 15,830 hours a year net of supervision, or A$554,049 at A$35/hour: A$46,171 a month.

Now split it by format, and the average stops being useful:

A metro store's entire nightly mop is 50 minutes. Take half an hour of attended time out of that and 20 minutes a night is left — real, but not a robot's worth.

Half an hour is a punitive assumption on a 330 m² floor, so we re-ran the estate on our documented modelling assumption of 8 minutes of attended time per robot cleaning hour — 0.10, 0.45 and 0.65 h a night for metro, standard and large. Metro payback moves from 111 months to 50, standard to 11, large to 8, and the estate to 13.5. That halves the metro gap and still leaves those six at more than four times the standard-store payback, well outside the 12–28 months this chain treats as investable. Either way the blended number is arithmetically true and operationally misleading: it is the large and standard stores paying for the metro ones.

What the rollout actually taught them

1. Sequence by payback, not by geography. The three large-format stores went first, the eleven standard stores next, and the metro six were deferred to the next lease cycle rather than bought for the sake of a tidy estate. Same model, same procedure — just not yet.
2. The dock is the rollout's critical path, not the robot. The CC1 Pro's dock auto-refills water, drains the waste tank and doses detergent, which is exactly what lets it run unattended. It also needs a tap and a floor drain where you want it parked. Twenty stores meant twenty small plumbing jobs, and they took longer to schedule than the machines took to deliver.
3. Standardisation pays in the boring places. One model meant one map procedure, one spares kit, one training package — and a relief unit from any store fits any other. It also saves water: at a uniform 4 ml/m² against 100 ml/m² for hand-mopping, the fleet avoids roughly 728,000 litres a year.

Add a fourth trading hour and the whole plan moves: a four-hour window drops per-robot capacity to ~1,734 m², and the three large-format stores need a second unit each — 23 robots, not 20. Window length, not clean rate, is what sizes a grocery fleet.

Would it work for your estate?

The honest answer depends on your reachable floor, your window and your wage rates — which is what the planner is for.

• Run your stores through the Fleet & ROI Planner to see which formats clear the bar on your numbers.
• Or tell us about your estate and we'll send a vendor-neutral shortlist with indicative pricing.