The sampling requirement is rising while the material itself is changing. The MF Regulations, widely known as the MRF Regs, have demanded more frequent sampling across more material categories since October 2024. Simpler Recycling has been live for households since March, although film and flexibles collection has moved to 2030 in England. Deposit Return Scheme (DRS) is scheduled for October 2027, taking PET bottles and cans out of the kerbside stream.
At RWM 2026 in Birmingham, our CEO Mikela Druckman chaired a panel with four people dealing with that adaptation problem directly: Pedro Faraldo García (Senior Technical Manager, FCC Environment), Sophie Lerway (Performance Improvement Director, Amcor), Jody Sherratt (Managing Director, Cheshire West Recycling) and Patrick Brighty (Head of Recycling Policy, ESA).
Each of them came at it from a different point in the chain. What they had in common was the moment the data stopped being a reporting exercise and started changing decisions on the line.
Pedro described what MF sampling looked like at FCC's Severn Waste Services site before the project started: collecting material, coning and quartering, then manually sorting a portion into categories, for inputs and outputs alike. Done properly, that method still sees less than 1% of what moves through the facility. The Greyparrot Analyzer, now running across more than 300 installations in 20 countries, measures the composition and mass of every stream it covers, continuously.
FCC had already been running Analyzer units on intermediate fractions to understand flows ahead of an upgrade. Then the question became what else the existing data could do. The Analyzer on the mixed plastics line allowed to run the two methods side by side, manual sampling against continuous AI monitoring for compliance reporting. In the first quarter of 2026, FCC and Biffa became the first UK operators to submit AI-derived composition data to the Environment Agency.
Patrick set out the load behind that. Advanced waste sampling, introduced in October 2024, feeds the 2028 efficient and effective assessments and informs how EPR payments are distributed to local authorities, with roughly 190 notified sites now reporting across England and Wales.
It's really positive to see the regulator acknowledging the efficacy of AI-based sampling."
– Patrick Brighty, ESA
He was equally clear about what has not been solved. Reporting remains spreadsheet-based, so continuous data still has to be reshaped into an Excel return. The ESA is pushing for API-based submission, which Patrick expects is a few years off.
Amcor's site takes polypropylene in and reprocesses it into an FDA-approved food-grade pellet. Sophie described the shift from throughput to quality as the line stabilised.
We started out focused on whether we could actually get the throughput. Now we can run stably, and what we need to focus on is quality. Can we make food grade, consistently?"
– Sophie Lerway, Amcor
The Analyzer data showed the team where target material was being lost into the reject stream, and at which point on the line it was going. That fed an 18-month redesign of the sortation plant.
If I compare the first six months of this year to the first six months of last year, there's a 12% improvement in our yield."
– Sophie Lerway, Amcor
The second shift was outward. As a reprocessor, Amcor's output quality is set largely by what arrives at the gate, and Sophie's view is that the data is more useful shared with the suppliers producing that feedstock than held inside the plant.
Our suppliers can help us and we can help them. It starts with actually talking about the data flowing between us."
– Sophie Lerway, Amcor
For Cheshire West Recycling (CWR), a council-owned company running collections for more than 160,000 households and one MRF at Ellesmere Port, the gap was granularity.
On the collection side I have data on everything. Then it becomes a mass of material at the back end."
– Jody Sherratt, Cheshire West Recycling
Jody's example on how waste intelligence provides that visibility, was the material CWR loses inside its own process, and what it takes to justify fixing it.
Getting the percentage of aluminium that's hitting the plastic line gives you your investment decision. We can work out how much loss is hitting that line, and from there calculate what your total investment is in your eddy current separator."
– Jody Sherratt, Cheshire West Recycling
DRS is the clearest example of a change no operator can sample their way into, because the material has not moved yet. For CWR, whose income depends partly on material sales, removing PET and cans is a material commercial event.
I'm going into business planning for next year, and my material changes massively in the last quarter of it. The data lets us forecast at a granular level: how many drinks cans, how much in-scope DRS material. Then we overlay return-rate scenarios and run the variance."
– Jody Sherratt, Cheshire West Recycling
We couldn't model that without the granular data. Otherwise you're guessing at the composition from tiny samples. Now I can tell you, for every load, exactly how much DRS material is in there."
– Jody Sherratt, Cheshire West Recycling
Jody went further into second-order effects: losing dense plastics from the stream changes abrasion inside collection vehicles, which changes replacement cycles, which lands back with the shareholder.
Patrick raised a related data problem still open across the sector. PRNs cannot be issued on DRS material, so operators issuing PRNs will need to discount the DRS-eligible fraction from October 2027, particularly in the early years when return rates are still scaling.
Asked what had surprised them, Pedro described the moment the project turned.
While we were validating the data for the EA, we saw anomalies. After investigation, it wasn't the Analyzer. It was a problem with equipment, or a bale that had been misplaced. That was the moment of realisation. Now I can trust the data, whereas before Greyparrot I couldn't."
– Pedro Faraldo Garcia, FCC Environment
Jody's example was quieter. CWR sees a reliable December spike in throughput that used to trigger overtime and hired vans, until continuous data showed the material was simply spiking around Christmas and would balance out by February.
My operations manager was getting stressed as the volume climbed. We could say: hold your nerve, the numbers say this is transferring between months."
– Jody Sherratt, Cheshire West Recycling
Pedro: gather the data now, before the composition change arrives, so you can measure the transition rather than react to it.
Sophie: bring the operational team with you, then take it to your suppliers. The value only lands if the people running the line, and the people feeding it, are using it.
Jody: audit your own assumptions.
Challenge yourself on how much of what you believe is based on evidence. The assumed position gets exposed by the actual. Your team don't always like it the first time, because it tells them something they've assumed for four decades might not be correct."
– Jody Sherratt, Cheshire West Recycling
Nobody on this panel set out to solve a compliance problem and stopped there. Each started with a reporting obligation or a process question and found the same data answered a different one: what the material is actually worth, where yield is being lost, and what next year looks like.
The closing exchange pointed at where this goes next. If a reprocessor can tell the facility upstream not just how good the bale was but what its process was losing, and the facility can act on it, both ends improve.
With Simpler Recycling live and DRS 18 months out, what arrives at the gate is going to change. Operators who can see their material will know within a shift. The rest will find out in a quarterly return.
Book a live demo to see what full-coverage composition data looks like on your own lines.