The economics pillar published in this cluster models the per-can dollar figure that reclamation technology has to clear to win capital review. The model is sound. The inputs the model depends on — measured residual percentage, facility-specific reclaim yield, the SKU distribution of trapped product — are not derived from the model. They are what the model is supposed to reconcile against.

Field-proof is the leg of the cluster that produces those inputs. It is the discipline of converting a published 12–15% industry figure into a measured number on the specific SKUs the facility runs, on the specific line, against the specific product mix the schedule produces that month. Without field-proof, the economics layer is a study. With field-proof, it is a capex case.

This pillar picks up where the economics pillar and its primer left off. The model itself is stakeable on aggregate inputs — that is the discipline those pieces walk. The field-proof tier stakes the model on the specific inputs a capex committee can interrogate. A reclamation case that ships field-proof is a case the committee can defend against measured data. A case that ships without it is a case that has to be taken on vendor trust.

What Field-Proof Means — And Why Aerosol Measurement and Validation Has to Lead It

Field-proof is not a number. It is a package.

The package contains the SKU-level residual baseline, the line-side pilot yield, the throughput-band yield curve, the operating-data support, and the regulatory-reconciliation layer that ties the residual reduction to the disposal-fee and EPR exposure. The whole package is what converts the economics pillar's model into a defended number for a specific facility.

Aerosol measurement and validation is the discipline of building the package. Measurement is the data layer — what's flowing through the lines, what's left in the cans, what the SKUs actually carry. Validation is the methodology layer — pilots that confirm yield at facility-specific mix, throughput-band work that confirms stability, operating-data audits that confirm the levers the model is tuned against. The two layers fail together or hold together. Measurement without validation produces data the committee cannot underwrite. Validation without measurement produces a yield figure the committee cannot defend.

The reason this tier exists as its own guide, distinct from the economics pillar: it is owned by operations and pilot teams, not by finance or strategy. The economics pillar's model can be written by a finance team working off cluster-published ranges. The field-proof package cannot be written off-cluster. It is built line-side. It is built shift-to-shift. And it takes a minimum of three months of SKU-level data plus ninety days of pilot operation before any number in it is publishable. That timeline is non-negotiable. The capital committee cannot compress it. The vendor cannot compress it. Only the calendar can.

SKU-Level Residual Measurement — The Data That Has to Exist First

End-of-line residual measurement is the first foundation. The instrumentation is straightforward — weigh-out stations that capture residual mass at the SKU level across enough production volume to be statistically meaningful. Most facilities that start measuring find data they did not have.

The variance between the best and worst SKUs in an instrumented facility typically sits at 4–7 percentage points of residual. That spread almost always exceeds the gap between the headline 12–15% figure and the facility's actual measured residual in aggregate. The why-aerosol-waste-is-invisible primer reports this spread as a feature of operations where residual measurement is being introduced for the first time. In practice, the spread holds for facilities regardless of measurement maturity, because the spread is driven by propellant ratio, valve configuration, and overfill margin — not by whether the facility has been measuring for years or hours.

What three months of SKU-level data typically produces:

- A weighted baseline residual profile by SKU. Reconciled against the same product mix the line expects to run in the capex year — not against last year's mix, not against an industry mix. The mix is part of the input set. - A distribution, not a mean. The mean residual percentage on a thirty-SKU product line is rarely the number that drives capex. The long tail of high-residual SKUs is what drives both the recovery opportunity and the eco-modulation exposure. - A timing pattern. Some SKUs run during the day, some during nights, some in the changeover window between campaigns. The residual profile shifts by hour, not just by SKU. A facility's measured number is the time-weighted number, not the SKU-averaged number. - A class of accepted-but-suboptimal cans. Some percentage of cans are accepted that fall below the residual profile percentile that matters, but above spec on fill weight and pressure. These are cans field-proof identifies as cost-bearing that ISO/GMP and the existing QA framework do not surface.

Without SKU-level data, the next layer of analysis — the line-side pilot — is missing its baseline. A pilot running against an aggregate residual figure produces a yield figure that has to be re-baselined against SKU-level data the moment it is available. The committee will recognize the re-baselining because the payback figure changes with it.

End-of-Line Residual vs. Fill-Head Scrap — Why the Visible Layer Is the Smaller Layer

Operators tend to equate the salvageable product layer with fill-line scrap — rejected cans, changeover flush, fill-head residue. These are visible. They sit in containers. They get written into rejects reports and waste handling manifests.

The visible layer is almost always the smaller layer. The larger layer is residue in accepted cans: product that passes every spec, ships to distribution, and then sits trapped at point of disposal because the propellant-to-product ratio or valve geometry does not deliver it to the consumer. End-of-line residual measurement captures the larger layer. Fill-head scrap reports capture the smaller one. Models that target only the visible layer leave the majority of the recoverable value on the table.

The line between the two layers is the gap that the Aerosol Fill-Line Optimization Guide addresses — on the operations side, the levers that move trapped-product volume up or down are propellant ratio drift, valve-to-product mismatch, overfill margin stacking, and changeover flush discipline. That guide links to the economics pillar for a reason: the per-can recovery figure grows fastest when trapped-product volume grows fastest, and trapped product is the larger of the two layers.

What this means for a field-proof package: the operating data the fill-line guide discusses is an input to the field-proof package, not an output of it. The fill-weight distribution, the valve configuration audit, the propellant calibration record, the changeover flush log — these are inputs the model is tuned against. Without them in the package, the field-proof has measured residual but no validation against what is driving it. The committee can fund the residual reduction, but it cannot fund the operational discipline that maintains it.

A reclamation case that asks the committee to fund a system against measured residual, but cannot show which operating lever moved and which did not, is a case that will have to fight the same fight twice — once for capex and once for sustaining the gain during operation.

Line-Side Pilot Methodology

After the SKU-level residual baseline is established (typically at three months minimum), a line-side pilot begins. The pilot answers a different question than the baseline measurement does: at facility-specific product mix, at facility-specific throughput band, what is the reclaim yield?

The minimum pilot design is a ninety-day run, structured in three phases:

- Days 1 through 30: stabilization. The pilot is running on a real production line against a real product mix. The first four weeks are noise — operator learning, throughput adjustment, throughput-band testing, transition handling. No yield number from the first month is publishable. Anyone using a one-month pilot number as input to a capex decision is using noise as input. - Days 31 through 60: throughput bands settle. The pilot can now run a representative distribution of SKUs at throughput levels the line will sustain in steady-state operation. The yield figure begins to converge — though it still drifts. - Days 61 through 90: facility-specific reclaim yield. By the end of the third month, the pilot has data across enough SKUs and enough production hours to model a yield number the capex committee can underwrite. The number converges within a tighter band. Stability across shifts becomes visible.

What the pilot has to measure, in addition to the reclaim yield headline:

- Trapped-product volume at facility-specific mix. This is what the SKUs are carrying into the pilot. The published reclaim yield — 60–85% of trapped product, per the Aerosol Reclaim Economics primer — only becomes a usable model input when it is run against the facility's measured trapped-product volume, not an industry aggregate. - Reclaim yield at facility-specific mix. The output the pilot produces, weighted by production volume across the SKU distribution the line will run in the capex year. - Throughput band at which the yield holds. A yield figure measured at pilot throughput is a number only at pilot throughput. The committee needs the throughput range over which the yield holds, because the line is not always running at peak. The yield curve usually degrades at the throughput extremes — the model has to specify where. - Instabilities by SKU. Some SKUs cause pilot yield to degrade substantially. The pilot log has to identify which, because those SKUs are also typically the highest-residual SKUs, which means they are the SKUs the recovery case rests on. A pilot that runs without SKU-level yield decomposition is hiding the highest-value layer.

A four-week pilot is not a field-proof instrument. It is a calibration run. Any payback figure derived from it is a model on noise, and the committee will see that on contact.

Validating Yield at Facility-Specific Product Mix

The 60–85% reclaim yield in the economics pillar is a published range — calibrated across multiple facilities, against varying product mixes, on different valve configurations. It is a starting point. It is not the right input for a specific facility's model.

The right input is what the pilot has measured on the facility's own SKUs, through the facility's own valve configurations, against the facility's own propellant-to-product ratios, at the facility's own throughput bands. The economics primer walks through why the published range is the starting point but not the answer; this pillar is where the answer gets assembled.

A facility's measured reclaim yield will usually sit inside the published range, but it will not be the midpoint. The high-residual SKUs that drive trapped-product volume — and therefore the per-can recovery number — often carry valve geometries or propellant ratios that impede yield. The low-residual SKUs that contribute volume but little per-can return often run at higher yield. The averaged facility number is what closes the case, not the published midpoint.

What gets reconciled in this layer:

- Pilot yield, weighted by SKU and weighted further by production volume in the SKU-then-time pattern the line is expected to run. - Operating-cost profile of running the pilot at industrial scale. The pilot itself runs on a smaller footprint. The cost curve is not linear — pilot economics and steady-state economics are not the same number. - Throughput band at which the weighted yield holds, including shift-to-shift stability. The committee will ask whether the yield holds across an entire 24-hour day or only during the day shift when the most experienced operators are running the line. - A second-window pilot run on the same data set to confirm yield has held without operator-side compensation. Some yields degrade within a quarter because operators have learned to compensate — the recovered number is a stroke of operator practice, not the system's intrinsic capability.

The validation package at this layer is what converts the economics pillar's model into the facility's own model. Without it, the economics pillar is a study. With it, it is a defended number on a defended yield.

Reconciling EPA Universal Waste Exposure Against Measured Residual

The Aerosol Reclaim Economics primer walks through the $0.07–$0.09 per landed can EPA Universal Waste layer in detail. This pillar is where that layer reconciles to measured residual data.

Two facts hold simultaneously, and the field-proof package has to hold both:

- The disposal fee scales linearly with volume, not with residual content. A can with 2% residual and a can with 15% residual incur the same per-can fee, because the fee is about disposal, not about how much product is left in the can. - A facility with measurable reduction in end-of-line residual can show a smaller EPR eco-modulation surcharge — the 10–20% layer — in the same fiscal year the measurement begins. The surcharge is keyed to measured residual content, not to volume.

This is the asymmetry that drives the field-proof case. Volume-side exposure ($0.08 per can, applied to every landed can) goes down by an indirect and modest amount when residual measurement begins — the offset is the reclaim yield multiplied by the measurement-driven residual reduction, not the gross volume. The eco-modulation surcharge (10–20% of base program fees) goes down by a direct and material amount in the same year, because the surcharge basis shifts from the residual number the facility has been using to the residual number the facility is measuring.

On a 50-million-unit facility in an EPR jurisdiction:

- Volume-side disposal exposure at the $0.08 midpoint: $4.0M annually, reduced modestly by reclamation. Reclaimed residual no longer incurs the fee — the offset is the reclaim yield multiplied by the residual reduction, not the gross volume number on which the fee is calculated. - Eco-modulation surcharge under a heavy residual profile: $500K to $1M annually, reduced materially in the first year measurement begins because the surcharge basis is now measured, not assumed. The year-one reduction is the gap between the old assumed residual number and the new measured one — applied across the eco-modulation surcharge formula.

The field-proof package shows both layers explicit, weighted by the facility's own measurement. A package that shows the volume-side number alone overstates the reclaim-driven fee reduction; a package that shows only the eco-modulation layer understates the multi-year savings stack. Both layers go in.

What a Defensible Field-Proof Package Looks Like Before Capex Review

The package a capex committee funds is not the package a vendor presents. The package the committee funds is the package the committee can verify against measured data, line by line.

Three readers, three different questions:

- Operations. "Does the residual measurement hold in production, or does it require pilot conditions to replicate?" The package has to show the residual measurement running on the actual production lines, on the actual shift patterns, not on a pilot rig under elevated operator attention. - Finance. "Does the yield validated in the pilot hold at the throughput band the line will run in steady-state operation?" The package has to show throughput-band-yield curves with explicit shift-to-shift stability, not a single-point yield figure averaged across the pilot. - Capex committee. "Is the payback figure derived from inputs we can defend, or is it derived from vendor-typical figures?" The package has to show the per-can recovery at the measured (not headline) residual, the reclaim yield validated at facility-specific product mix, the disposal-fee avoidance at the actual disposal-fee band the facility operates under, the EPR eco-modulation surcharge exposure computed downstream from measured residual — not upstream from a vendor presentation.

What goes in the package:

- A three-month SKU-level residual baseline, weighted by the projected production volume in the capex year (typically the year following the pilot). - A line-side pilot yield, validated across at least ninety days and across at least the top decile of SKU volumes the facility runs — the SKUs that drive the case. - A throughput-band-yield curve showing the band at which the yield holds and where it degrades. - An EPA Universal Waste exposure reconciled against measured residual volume, not against headline residual. - An EPR eco-modulation surcharge computed against measured residual, with jurisdiction-by-jurisdiction exposure flagged. - A per-can recovery figure that uses the measured residual, the pilot yield, and the facility's actual disposal and EPR exposure — not industry averages.

What does not go in:

- Industry-aggregate residual percentages from market research charts. - Vendor-published reclaim yields without facility-specific validation. - Throughput assumptions rather than measured throughput bands. - Optimistic residual-reduction trajectories the field-proof package has not yet produced. - Per-can figures that use the industry average residual rather than the measured one.

The package is defended line by line. The committee funds what the package can defend; it does not fund what the package cannot. The field-proof discipline is the discipline of producing a package whose every line holds on interrogation.

Linking Field Proof Back to the Economics Pillar

The economics pillar's model is a five-layer stack — yield-to-dollar conversion, landfill-disposal-fee avoidance, EPR eco-modulation, per-can cost recovery, payback window. The primer walks the layer math on aggregate inputs. The economics pillar coordinates the discipline across the cluster.

Field-proof is the input side of that model. Every number the model depends on — measured residual percentage, facility-specific reclaim yield, the SKU distribution of trapped product, the throughput band the yield holds at, the disposal-fee band the facility operates under, the EPR exposure the facility ships into — comes out of the field-proof package. Without it, the model is a study. With it, the model is a defended capex case for a specific facility.

The pieces map. SKU-level residual measurement produces the denominator. The line-side pilot produces the yield multiplier. The throughput-band validation produces the operating-assumption side. The EPA and EPR reconciliation turns the recovered-product layer into a per-can dollar figure that holds in actual fee schedules, not industry assumptions. The package delivers the model a defined set of inputs to operate on.

What the field-proof tier closes: the gap between an aggregate economic study and a facility-specific capex case. What the field-proof tier does not close: the operating discipline of running the reclamation system post-installation. That is a separate layer, owned by the same operations teams that built the package, and it is the layer the field-proof tier hands off to once measurement and validation are in production.

Run the model on measured inputs. The cluster's calculator at `/resources/manufacturer-roi` runs the same five-layer model the pillars describe — and the field-proof package is what makes its outputs defensible on a facility-specific case.

Run the per-can math on your facility →

Subscribe for more guides →

FullCan helps aerosol manufacturers measure, reduce, and recover product waste. Built for operations teams who need numbers, not narratives.