Waste Sorting KPIs and Benchmarks: Recovery, Purity, Uptime and Labor for MRFs

Six KPIs describe a sorting plant: recovery rate, bale purity, availability, throughput per labor hour, energy per ton, and residue rate. Everything else drives one of those six. Use the ranges below as reference points for spotting a gap, not as targets, and compare yourself with plants that receive similar inbound material.

Every figure shows its basis. A superscript number links to a public source in the Sources list, and the sentence gives that source's scope. A figure marked no cited source is kept exactly as originally published because no public source stating it was found, so treat it as illustrative.

At a glance

Recovery rate

Recovery rate is the share of targeted recyclables that end up in a saleable product. It runs 80 to 85 percent at a typical plant and 92 to 95 percent at world class facilities using optical sorters with recirculation loops and robotic quality control (QC) no cited source.

Measure it from the other side, by weighing what the plant throws away. Residue containing 25 percent recoverable PET, aluminum, and fiber means real money on the floor; the best plants keep recoverables in residue below 10 percent no cited source.

The biggest single lever is burden depth control, since an overloaded optical belt can drop recovery by 10 to 15 points during peak hours no cited source.

How to measure

• Divide targeted recyclables shipped as saleable product by targeted recyclables received.
• Pull a 90 kg residue sample every week and weigh the recoverable PET, aluminum, and fiber in it.
• Trend the weekly audit results every month.

What moves it

• Burden depth control on optical sorter belts, especially at peak hours.
• Recirculation loops that give missed material a second pass.
• Robotic QC after the optics.

See how ejection accuracy and availability drive sorted tons with the Optical Sorter Throughput calculator.

Bale purity

Purity is what buyers pay for. Typical PET bales run 92 to 94 percent, world class 96 to 98 no cited source. Aluminum used beverage can (UBC) bales should exceed 98 percent no cited source.

Mixed paper holds under 3 percent outthrows at strong plants versus 5 to 8 at weak ones no cited source. For comparison, ReMA's ISRI grade (54) Mixed Paper specification allows outthrows of no more than 3 percent and prohibitive materials of no more than 2 percent¹, so a bale at the weak-plant level falls outside that grade.

For PET, ReMA's ISRI specification grades bales by the share of PET bottles they contain. These are trading thresholds a bale must meet, not measured plant results.

How to measure

• Run a bale breakdown audit for each commodity every week.
• Weigh outthrows and prohibitive materials against the sampled bale weight.
• Log every audit so you get a trend line instead of anecdotes.

What moves it

• Last chance manual QC positioned after the optics, the biggest purity lever.
• A dual pass optical configuration on containers.
• Film removal early in the process, since film that reaches the container line degrades every downstream sort.

Compare each audit with the specification or broker target in the Contamination Rate calculator.

Availability

Availability means scheduled hours actually running at rate. It sits at 80 to 85 percent for typical plants and 92 to 95 percent for the best no cited source. World class plants hold unplanned downtime under 3 percent of scheduled time no cited source.

Two specific losses matter here. The first is wrapper and tangler cleanout on screens: typical plants lose 30 to 60 minutes per shift, and best in class plants lose under 15 with non wrapping screen designs no cited source.

The second is optical sensor fouling. Ejection accuracy commonly drops 5 to 10 percent when cleaning intervals stretch past 8 running hours in dusty fiber applications no cited source, so set the interval on purpose rather than by habit.

How to measure

• Divide hours running at rate by scheduled hours.
• Split downtime into planned and unplanned categories.
• Log screen cleanout minutes per shift as their own line.

What moves it

• Non wrapping screen designs that shorten cleanout.
• A set optical sensor cleaning interval on dusty fiber lines.
• A daily downtime huddle on the previous day's stops.

The Optical Sorter Throughput calculator and the Baler Capacity calculator both take availability as an input, so you can see what each point is worth in tons.

Throughput per labor hour

Throughput per labor hour separates plant generations. Legacy manual plants process 0.5 to 0.8 tons per labor hour, modernized lines with optics run 1.0 to 1.5, and highly automated facilities exceed 2.0 no cited source.

On the pick line itself, a typical human sorter sustains 30 to 40 picks per minute, and top performers reach 45 to 50 no cited source. For robots, one manufacturer rates its belt-mounted Delta Compact at up to 65 picks per minute², a single model's rated maximum rather than a fleet average. Robots sort with better than 90 percent accuracy on well presented material no cited source.

Track safety beside productivity. In 2024, U.S. materials recovery facilities reported 5.9 recordable injury and illness cases per 100 full-time workers and a days away, restricted or transferred (DART) rate of 4.8, against 2.7 and 1.7 for all manufacturing³.

How to measure

• Divide tons processed by paid labor hours for the same period.
• Convert station counts and pick rates into tons per labor hour.
• Benchmark the result every quarter.

What moves it

• Stabilizing burden depth first, so every station sees well presented material.
• Automating the highest volume commodity second.
• Redeploying people to QC rather than cutting heads immediately.

Before choosing which commodity to automate, size what one optical unit sorts per shift with the Optical Sorter Throughput calculator.

Energy per ton

Energy intensity runs 8 to 15 kWh per ton at typical plants no cited source. Under 8 is world class, and under 6 has been demonstrated with full variable frequency drive (VFD) conveyance and demand controlled air no cited source.

Air handling and conveyance together usually account for 60 to 70 percent of plant draw no cited source, so that is where to look first.

How to measure

• Divide metered kWh by tons processed over the same period.
• Submeter conveyors and air systems separately where you can.
• Trend the figure monthly with the rest of the scorecard.

What moves it

• Variable frequency drives across conveyance.
• Demand controlled air on separators.

Break the load into addressable pieces with the Sort Line Conveyor Energy calculator and the Air Separator Energy calculator.

Residue rate

Residue is what leaves the plant as waste instead of product. In a November 2019 survey of seven U.S. MRFs by The Recycling Partnership, residue made up 13.7 percent of outbound material, valued at minus 51.62 dollars per ton⁴. That is a small sample, not a national statistic.

Inbound composition sets the floor. The same report estimates U.S. national average inbound contamination of curbside recycling at 17 percent by weight⁴. A facility receiving 25 percent contaminated feed will never post the residue numbers of one receiving 12 percent no cited source, so compare against plants with similar feed.

Screen efficiency

Screen separation feeds residue directly. Screen efficiency of 80 to 85 percent is typical, while well maintained and correctly loaded screens hold above 90 percent no cited source.

A screen slipping from 90 to 80 percent pushes misplaced material to every downstream sorter and quietly inflates both residue and labor no cited source.

How to measure

• Divide residue tons by total outbound tons for the same period.
• Measure screen efficiency on each screen every month.
• Record inbound contamination beside each residue figure so like is compared with like.

What moves it

• Cleaner inbound feed.
• Screens that are well maintained and correctly loaded.
• Fewer recoverables lost to residue, which the weekly residue audit tracks.

Check each screen with the Screen Efficiency calculator, and work out each sample's contamination share with the Contamination Rate calculator.

How to build your own baseline

• Hold a daily downtime huddle on the previous day's stops.
• Run contamination and recovery audits every week.
• Score all six KPIs on a full monthly scorecard.
• Compare against plants with similar inbound composition, not the best plant you know of.

Improve in this order

• Availability first, because a point of uptime is pure throughput with no quality tradeoff.
• Recovery second, since material already paid for should leave as product, not residue.
• Purity third, which raises price per ton rather than tons.
• Labor productivity and energy fourth, which compound the earlier gains.

For example, moving from 80 to 88 percent availability adds roughly 10 percent annual tons at the same rate, because 88 / 80 = 1.10.

Plants that run this cadence for 12 months typically move two of the six KPIs from the typical band into the top quartile; plants that only measure annually move none no cited source.