Sanjay Verma has stood over a dumpster holding three hundred faultless loaves, ruined by nothing but timing. Not a burst oven, not bad flour, not a broken line. Just a shift change that went wrong, dough that sat past its proving window, and a batch of perfectly good bread that became a batch of perfectly good waste.
Sanjay is group operations director at Granary Fields Bakeries, which runs three plants baking around the clock. Ask most people what costs a twenty-four-hour bakery money and they will point at line speed or yield. Sanjay had already optimised both to the edge. The real leak was somewhere nobody was measuring at all.
I've stood over three hundred faultless loaves that were ruined by nothing but timing. There isn't an operations director in the country who enjoys that sight.
Where the money was actually going
Each plant runs three shift changes a day, roughly ninety seconds at each seam where one team hands the operation to the next. That is where the trouble lived. A changeover planned by the early shift would not reach nights cleanly, and the incoming team would run the old plan for forty minutes before anyone caught it. Or cover for a sick operator didn't hold the right ticket, and a line sat idle while managers scrambled.
On a bakery line, idle is never neutral. Dough left proving past its window turns from stock into waste, and the bakery had already paid for the flour, the labour and the energy before paying again to bin the result. Across the group, Granary Fields was writing off a couple of hundred trays a week to nothing more than botched handovers, a cost nobody had ever put a number against because nobody had ever looked at the seam itself.
Building the fix around the shift change, not the line
Granary Fields brought in Northpeak, starting with its scheduling tool, Crew, which checks that every operator assigned to a shift is actually signed off to run the product planned for that line. That single rule removed a whole class of two-in-the-morning problem before it could start. Handovers themselves became scheduled and visible across the seam, so an incoming team walked in already knowing the plan rather than reconstructing it from a note left on a board.
Fresh, Northpeak's monitoring tool, tracks proving and bake windows on every batch and flags anything drifting toward its limit, catching problems before they become write-offs rather than counting them after. Pulse then pulls all three plants onto one screen for Sanjay, replacing three separate reports that used to land a day late.
The rollout was not smooth. The first version of Fresh was tuned too tight and flagged batches that were genuinely fine, and a shift leader will only tolerate false alarms so many times before switching off. Sanjay sat down with the most sceptical shift leader at the Wakefield plant, retuned the thresholds using his judgement, and within a week the system caught a genuine drift on an overnight run that would have cost a full batch.
The sceptic always has to become the advocate, or it doesn't stick. Nothing I say from head office carries the weight of one respected shift leader telling the next plant it earns its keep.
What changed, and what Sanjay won't overclaim
Changeover and handover write-offs have dropped from around two hundred trays a week across the group to a level low enough that a single incident now stands out. Overtime has fallen too, because staffing follows the actual plan instead of covering gaps with the same stretched people. Sanjay is careful about attributing all of it to software; some of the improvement, he says, is simply that measuring a thing makes you manage it better.
The number he stands behind without any caveat is line uptime, because it comes straight from mechanical run data rather than judgement. Lines now start their shifts ready instead of scrambling to catch up.
I'd rather give you a number I believe than a number that sounds good. The uptime is just the system doing what it said it would.
Pulse changed Sanjay's own week as much as anything on the floor. Instead of stitching together three late reports, he now compares waste, overtime and uptime across all three plants live, making decisions on what is happening rather than what happened yesterday, which matters on a product that doesn't wait.
A quieter kind of proof
The moment that stuck with Sanjay wasn't dramatic. He stood on the Wakefield floor at a two in the morning handover, the hardest of the day, and watched the night team walk in already knowing the plan. No firefighting, no skip, no phone call. His advice to other operations directors is blunt: stop staring at line speed and go watch a shift change instead, because the production problem you assume you have is often a much cheaper communication problem in disguise.