Verona AI
Waste and giveaway are two of the few costs in a food factory you can measure to the gram. Most businesses still don't, because the data sits in five places and nobody has time to join it up. This guide shows where to look and what to fix first.
The scale of it
- UK: WRAP estimates UK food manufacturers wasted around 706,000 tonnes of edible food in 2021, worth about £0.85bn. Manufacturing accounts for 13% of all UK food waste by weight (WRAP).
- Ireland: the food and drink manufacturing and processing sector generated an estimated 169,000 tonnes of food waste in 2024, 24% of the national total (EPA).
- What's coming: EU rules set a target to cut food waste from manufacturing and processing by 10% by 2030, which applies to manufacturers in the Republic of Ireland (EPA, as above).
Those are national numbers. The one that matters is yours, by line and by week.
Where the money goes
Most waste in a food factory comes from small, repeated losses that nobody adds up.
| Where it happens | Where the data usually is | First fix |
|---|---|---|
| Giveaway: packs filled above declared weight | Checkweigher logs, often never exported | Report average fill against target, per line, per shift |
| Start-up and changeover | Production sheets, sometimes paper | Track kilos lost per changeover; schedule to reduce them |
| Out-of-spec product | QA records and hold logs | Link rejects to line, shift, recipe and raw material batch |
| Overproduction and short-dated stock | Orders, forecasts and stock in separate systems | Plan from actual order history, not last year's spreadsheet |
| Yield loss on raw materials | Purchase data and production records | Compare kilos bought with kilos packed, per product |
Giveaway: the cost hiding in plain sight
Under the average quantity system, packs must meet their declared weight on average, with tight limits on underweight packs. To stay safe, many lines run well above target. Every gram over is product you've paid for and given away.
Here's an example of how fast it adds up (illustrative figures, not a client's):
- A 400g pack running at an average of 412g gives away 12g per pack.
- At 20,000 packs a week, that's 240kg a week.
- At £4 a kilo in ingredients, that's £960 a week, or about £50,000 a year, on one product.
Most factories already collect the checkweigher data needed to see this. Few report it in a way the people running the line can act on during the shift.
Why you can't cut what you can't measure
Most waste projects stall for the same reason: nobody trusts the numbers. Purchase data says one thing, production records say another, and the yield figure in the middle depends on whose spreadsheet you open.
The usual culprits are dull but costly: ingredients recorded in different units from one week to the next, credit notes counted as spend, and the same week's data loaded twice. Each one distorts every cost-per-kilo and yield figure built on top of it, without anyone noticing.
So the first job in any waste project is to get purchases, production and dispatch data joined up and checked automatically every time new data arrives. Only then can you see where the kilos are really going.
Where AI helps, and where it doesn't
- Helps: forecasting demand from years of order history, so you make closer to what sells. Spotting patterns in rejects across lines, shifts and raw material batches that a person would take weeks to find.
- Doesn't help yet: if checkweigher and production data aren't captured reliably, start there. A clear daily report often saves more than a clever model.
A first step this week
Pick your highest-volume product. Pull one week of checkweigher data and work out the average fill against declared weight. Multiply the difference by your weekly volume and your cost per kilo. That one number usually decides whether waste and giveaway deserve a project.
Sources
- UK Food Waste & Food Surplus – Key Facts, WRAP, July 2025 (2021 data)
- Food Waste Statistics for Ireland, Environmental Protection Agency, July 2026 (2024 data)
