'I said I wanted to dispose of it'
Tell the lab a container is near mint and you want it gone, and it refuses to recommend disposal. We would rather buy it from you at top band. People find this annoying and then they find it useful.
Yard log · 29 January 2026
Turning a grader's judgement into a tool
For thirteen years the decision about what a container should become lived in two people's heads. Writing it down as something a customer could run themselves turned out to be harder, and more revealing, than building it.
- Sheet
- 44
- Revision
- 2026-10
- Scale
- 1:1
- Drawn at
- Peoria IL
3Questions
4Possible paths
13 yrOf judgement encoded
0Data sent anywhere
Every container that comes through the gate gets a decision: reuse, rebottle, repurpose or recycle. Until this year that decision was made by two graders using eighteen checks and a lot of accumulated pattern recognition. Writing it down exposed how much of it was not actually judgement at all.
What turned out to be rules
- Last contents set the ceiling, not the path. Solvent exposure caps a container permanently regardless of how it looks, and undocumented history caps it at Grade B. Neither of those is a judgement call; they are facts with consequences.
- Condition sets the path. A sound bottle goes back to service, a spent one gets rebottled, a failed cage gets baled. Also not a judgement.
- The customer's goal picks between valid paths. The same container can be cash, inventory, a rain harvester or regrind, and all four can be correct.
- Only two things were genuinely judgement: how severe staining is, and odour. We fixed the first with a published four-level scale and the second with a two-grader rule.
That was the surprising part. Thirteen years of expertise decomposed into about forty lines of decision logic plus two checks that need a human nose and a human eye. The expertise is real — it is just concentrated in a much smaller place than it felt like from inside.
The verdicts customers push back on
'But the bottle is cracked'
A split bottle sounds terminal and is not. The cage, pallet and hardware are about 60% of the container by weight and are usually fine. That is a rebottle, and it is roughly half the price of replacement.
'It is perfectly clean though'
A solvent container can be spotless and still permanently unsuitable for product service, because aromatics absorb into the polyethylene wall. The lab says so instead of selling you a wash cycle that cannot help.
'Why does it drop a grade for paperwork?'
Because we cannot sell a history we cannot prove and nobody serious will buy one. Undocumented containers cap at B. This is the single most common objection and the least negotiable.
We built the thing so that it would sometimes argue with the person using it. A tool that always ends in 'contact us' is a form with extra steps, and we had enough of those already.
What we deliberately left out
- Prices. Bands move with resin and season, and a number that goes stale makes the whole tool untrustworthy.
- Anything that leaves the browser. The lab runs entirely in the page and sends nothing anywhere, which is why there is no 'enter your email to see results'.
- A grade guarantee. The lab predicts the grade a container will earn; a grader still earns it, with eighteen checks and good light.
- Regulatory determinations. It will not tell you whether a residue is RCRA-regulated, because that is not a question a decision tree should answer.
The one addition we did make is the ticket code. Every verdict generates a short code you can paste into the contact form, and the grader reading your request then knows exactly which container you are describing before they reply. It is a small thing that removes an entire round of email.
Try it on something you own
Three questions about a container you already have, and the honest verdict with the environmental arithmetic attached.