Plans fail less often because the actions were wrong than because the assumptions underneath them were never tested. These five questions take a morning to work through and determine whether what follows is a plan or a wish.
1. What are we actually trying to change?
"Reduce waste" contains at least four different objectives, and they pull in different directions:
- Reduce cost — fewer collections, lower gate fees, less material purchased
- Reduce volume — less leaving the site, regardless of cost
- Increase diversion — the same volume going somewhere better
- Satisfy a reporting requirement — defensible numbers, on a schedule
A programme optimised for cost can lower diversion. One optimised for diversion can raise cost. Name the objective first, because it decides which actions count as success.
2. Who owns the material at the moment it becomes waste?
Every stream has a point where it stops being product, stock or packaging and becomes waste. Someone is standing there. Usually it is not the person who will be asked to deliver the reduction.
If the decision happens on a production line at 3am and the plan is owned by a sustainability manager who works days, the plan needs a mechanism to reach that moment. Most do not, which is why they produce posters rather than change.
3. What have we already tried, and what did we learn?
Nearly every site has attempted something — a segregation scheme, a supplier conversation, a recycling contract. Many were quietly abandoned. The reasons are the most valuable information available, because they are specific to your operation and nobody else can supply them.
Ask what was tried, when, who led it, and what happened. "It didn't work" is not an answer. "The bins were twenty metres from the line and the shift couldn't spare the time" is.
4. What constrains us that we do not control?
Some streams cannot currently go anywhere better, because the processing capacity does not exist within a sensible distance, or no buyer wants the material in the condition it leaves in.
Identifying these early matters for two reasons. It stops effort being spent where it cannot succeed. And it redirects that effort upstream — if a material cannot be diverted, the remaining lever is generating less of it, or generating it in a different condition.
5. How will we know whether it worked?
Decide the measure before the intervention, not after. Specifically: what is measured, by whom, how often, against what baseline, and expressed how.
That last point causes the most trouble. An absolute tonnage that falls during a quiet quarter has not demonstrated anything. Waste per unit of output, per shift, or per order survives the question "was this just lower production?" — and that question will be asked.
What the answers produce
Worked through honestly, these five produce a short document stating the objective, the person at the decision point, the history, the constraints, and the measure. That document is worth more than a fifty-page strategy, because every line is specific to your site and can be acted on.
It also makes any external conversation dramatically more useful. A consultancy arriving to these answers starts at the real problem rather than spending the first two weeks discovering them.
Where most teams get stuck
Question three. Not because the history is unknown, but because writing down what failed feels like cataloguing mistakes. It is the opposite — a documented failure is a tested hypothesis, and it is the cheapest learning in the entire programme.
Which of the five is hardest to answer at your site? That answer usually points at where the work needs to begin.
Written by ezero-waste. Any measured result carries its own scope and evidence.
Common questions
Why does the objective matter so much?
Because a programme optimised for cost can lower diversion, and one optimised for diversion can raise cost. Naming the objective decides which actions count as success.
What if we have tried before and failed?
That history is the most valuable input available, and nobody else can supply it. A documented failure is a tested hypothesis, not an embarrassment.
Do we need a baseline before starting?
You need one before setting a target. A target set before a baseline exists is a number someone invented.