Ask an organisation what it knows about its waste and you will usually be handed a report with tonnages, containers, recycling rates and a trend line. It is real information, produced at some cost, and it answers a narrow question well: what left our sites, in what containers, and where did the contractor say it went.
It cannot answer the question most people are actually asking, which is why there is so much of it.
Output data cannot explain an input decision
A bin tells you a laminated pouch arrived at a site and had nowhere useful to go. It does not tell you that the pouch replaced a recyclable tray eighteen months ago, that the change saved a small amount per unit, or that nobody costed the disposal consequence. The decision lives in the specification and the purchase order. The evidence of its consequence lives in the waste stream, separated from its cause by months and by several departments.
This is why waste reduction programmes built only on waste data plateau. You can improve segregation, right-size containers and cut collections. Those are worthwhile and they reach a limit, because the material still arrives.
Three reasons the numbers will never simply match
Before anyone attempts to compare what came in with what went out, three effects need to be understood, because ignoring them produces confident conclusions that are wrong.
- Units. Purchasing records cases, rolls, pallets and each. Waste records tonnes and lifts. Converting between them needs component weights, and those weights need a source.
- Time. Material bought in March may be consumed in May and collected in June. Stock holding, shelf life and collection frequency all shift the apparent date.
- Fate. Not everything purchased becomes waste at your site. It leaves with a product, with a customer, or through a reverse logistics route you do not measure.
So do not mass balance a month
A predictable analytical error is a monthly comparison of packaging purchased against packaging in the waste stream, presented as a capture rate. The gap is then read as loss, leakage or contractor error, when a large part of it is simply timing and stock.
The honest approach is to compare over a longer period, to compare like categories rather than totals, and to state the assumptions used for conversion. A longer period reduces the distortion but does not remove it: an annual figure does not automatically wash out stock movements. If opening and closing stock changed materially, or if a product line launched or ended mid-year, reconcile opening stock, purchases, closing stock and outputs explicitly rather than assuming the year balances itself. The result is not a perfect balance. It is a set of differences worth investigating, ranked by size.
What the combined view is genuinely good for
When the two datasets are brought together carefully, three things become visible that neither shows alone.
First, avoidable material: items purchased in volume that consistently appear in residual waste because no recycling route accepts them. Second, specification opportunities: components where a small material change moves a large tonnage into an existing segregated stream. Third, behaviour: places where the right material is arriving but the capture point is in the wrong location or the wrong shape.
A tonnage report on its own shows none of that, although a waste composition audit can reveal part of it by describing what is actually in the residual stream. What a composition audit still cannot tell you is why the material was bought, by whom, under which specification. Procurement data does not tell you why either. Put the two together and you have enough to investigate the question properly.
Where to start if you only do one thing
Take your top twenty purchased items by volume. Add a component weight and a disposal route for each. Then look at your residual waste composition. The overlap between those two lists is a sensible place to look for avoidable cost, and it is a short piece of analysis rather than a programme.