Why Contractor Estimates Rarely Match Actual Waste Weight

Abstract#

Most commercial waste tonnages are not weighed. Contractor waste estimates are calculated from the number of containers lifted and their volume and an assumed fill level and an assumed density. This blog tests one question: how far can a volume based estimate of commercial waste move with its assumptions and which assumption moves it most. It applies the volume to weight conversion factors published by the US Environmental Protection Agency in 2016 which GRESB points participants to when tonnage is unknown. It sets them beside measured fill levels from a municipal sensor trial and three peer reviewed studies located through Google Scholar and platform data from Wastify AI. The finding is that one full container of 1100 litres of commercial waste can be estimated at anything from 36.5 kg to 195.8 kg depending on which published factor is applied: a spread of 5.4 to 1. Add the range of fill levels measured at emptying and the spread for a single lift reaches 7.1 to 1. The empty container itself weighs 58 kg. A weighed record removes every one of those assumptions.

Key Findings#

1. The US EPA volume to weight factors give commercial dry waste at 56 to 73 pounds per cubic yard and all commercial waste uncompacted at 138. Mixed municipal waste including commercial waste uncompacted is given at 250 to 300. Applied to one full container of 1100 litres those factors give between 36.5 kg and 195.8 kg: a spread of 5.4 to 1 [2].
2. GRESB permits a tonnage estimate from bin count and volume and fill level and collection frequency. Where fill is unknown it asks for a worst case assumption that every bin was full and it points participants to the US EPA factors [3]. A municipal sensor trial measured average fill at emptying of 75% to 84% for mixed waste and 36% to 94% for plastic [4].
3. Combining the lowest published factor with the lowest measured fill for mixed waste gives 27.4 kg for one lift of a 1100 litre container. The highest factor with the full bin assumption gives 195.8 kg. The same lift can be reported at values 7.1 times apart [2 3 4].
4. The national figure rests on the same problem. Defra estimates that commercial and industrial sectors in England generated around 32 million tonnes of waste in 2024 and states that this remains extremely difficult to estimate owing to data limitations and data gaps [5]. A review of UK waste composition found insufficient data to compare the commercial element of municipal waste and called for further work to produce more reliable estimates [7].
5. In the Wastify AI container register a 1100 litre container has a tare of 58 kg: more than the 36.5 kg to 47.6 kg of contents that the EPA dry waste factors assign to it when full. Across 956000+ weighings the average recorded disposal is about 5.4 kg net of its container [1 2].

1. Introduction#

1.1 The Question#

A tonnage on a contractor report looks like a measurement. In most commercial buildings it is a calculation. The containers lifted are counted. Each is assumed to be a certain proportion full. A density is applied to turn litres into kilograms. The result is printed to one decimal place and carried into the service charge and the sustainability report.

This blog asks one question about contractor waste estimates. How far can a volume based estimate of commercial waste move with its assumptions and which assumption moves it most. The question can be tested with published numbers. If the official conversion factors and the fill levels measured in practice sat in a narrow band the estimate would be close to the weight. Section 3 shows that they do not.

1.2 Why This Matters Now#

Two developments make the gap between estimate and weight visible in a way it was not before.

First every receiver record now says whether a weight is estimated or actual. Under the Digital Waste Tracking (England) Regulations 2026 a receiver must state the total weight of controlled waste under each waste code with the unit of measurement and whether the weight is estimated or actual. It must also record the type and number of containers [6]. The regulations treat estimation as a fact to be declared rather than hidden.

Second reporting frameworks are discounting estimates. GRESB asks for a full bin assumption where fill is unknown [3]. Research Blog 05 in this series found that the EPRA guidelines ask for estimation to be carried out as little as possible and for the proportion of estimated data to be disclosed [8]. An estimate that was good enough for an invoice is now a disclosure risk.

1.3 Building on the Research Series#

This is the sixth blog in the Wastify AI research series [8].

• Research Blog 01 Reported measured densities for office waste streams from 73 to 172 pounds per cubic yard in a study of 2684 bags across 18 office buildings.
• Research Blog 03 Found that the tare of one large container weighs as much as about eleven average disposals.
• Research Blog 05 Found that the EPRA guidelines ask for estimation as little as possible.

This blog takes the next step. It applies the published factors to one container and shows how far the answer moves.

1.4 Scope and Structure#

This blog covers waste from multi tenant commercial buildings collected in wheeled containers. It proceeds in six sections.

• Section 2 Describes the method and the sources.
• Section 3 Sets out five findings.
• Section 4 Discusses what they mean for property managers and sustainability leads and contractors.
• Section 5 States the limitations.
• Section 6 Answers the question.

It does not assess the practice of any individual contractor. Level 0 is the competition: the problem is the absence of measurement rather than the conduct of any supplier.

Replace the estimate with the weight before it reaches a report. Wastify AI delivers waste intelligence for commercial buildings where every disposal is weighed and photographed at source.

2. Method#

2.1 Research Design#

The blog combines three evidence streams. The first is the published inputs to a volume based estimate: the US EPA conversion factors and the GRESB estimation rule and fill levels measured in a municipal sensor trial. Each factor was applied to one container of 1100 litres to show the range of estimates the same lift can produce. The second is peer reviewed research on the reliability of waste quantity and composition data. Three studies were located through Google Scholar and each abstract was read at an index or repository or library record. The third is platform data from Wastify AI.

2.2 Sources#

Every external source was opened directly on 12 October 2026 and each reference carries that access date.

Table 1: Sources consulted

Source: compiled for this blog from the references listed at the end.

2.3 Platform Data#

The container register holds tare weights for eleven bin types from a 2.5 kg slim bin to a 58 kg container of 1100 litres. Platform totals were read from the live counter on the Wastify AI homepage on 12 October 2026. The counter showed 5141000+ kg weighed at disposal and 956000+ verified weighings. It showed 2000+ tonnes of CO2e avoided and 380+ companies reporting. The same page lists £660000+ recharged on verified data. Public totals are rounded down. One weighing is one disposal placed on the integrated scale of the kiosk and recorded net of the container.

2.4 Calculations#

Pounds per cubic yard were converted to kilograms per cubic metre by multiplying by 0.5933 and then multiplied by 1.1 for a container of 1100 litres. The factor of 56 gives 33.2 kg per cubic metre and 36.5 kg per full container. The factor of 300 gives 178.0 kg per cubic metre and 195.8 kg per full container. The spread is 195.8 divided by 36.5 which gives 5.4. The lowest combined estimate is 36.5 kg multiplied by a fill of 75% which gives 27.4 kg and 195.8 divided by 27.4 gives 7.1. A full bin assumption for a container measured at 36% full overstates it by 1 divided by 0.36 which gives 2.8 times. The average weighing is 5141000 kg divided by 956000 which gives 5.38 kg. Every derived figure was recomputed from the stated inputs before publication.

3. Findings#

3.1 How a Volume Based Estimate Is Built#

The GRESB Real Estate Assessment sets out the method most estimates follow. A tonnage may be calculated from four data points: the number of bins and their volume and their filling level and the frequency of collection. Where the fill level is unknown the participant should apply a worst case scenario and assume every bin was full at each collection. GRESB does not prescribe conversion factors and encourages participants to refer to the US EPA guidelines [3]. Table 2 sets out each input and where it usually comes from.

Table 2: The inputs to a volume based estimate

Source: GRESB estimation rules [3] and US EPA [2] and the Wastify AI record structure [1].

Only the first input is counted. The second is a catalogue figure. The third and fourth are assumptions. The estimate inherits the uncertainty of its weakest input and the weakest inputs multiply.

3.2 The Density Term#

The EPA published its updated conversion factors in April 2016. It states that primary data collection was not performed and that the update drew on more current secondary measurements [2]. Table 3 applies the factors for commercial and mixed waste to one full container of 1100 litres.

Table 3: One full 1100 litre container under the US EPA factors

Source: US EPA Volume to weight conversion factors April 2016 [2]. Metric conversions computed for this blog.

The same full container can be reported at 36.5 kg or at 195.8 kg depending on which line of the same official table is used. The spread is 5.4 to 1. Within single categories the ranges are wide too. Mixed office paper loose is given at 110 to 380 pounds per cubic yard: a spread of 3.5 to 1. The organics lines range from 135 pounds per cubic yard for commercial organics to 1000 for source separated commercial organics [2]. Research Blog 01 reported a measured range from 73 to 172 pounds per cubic yard across office streams in the same type of building [8].

The research on pricing garbage by the bag shows why density cannot be treated as fixed. The study followed 75 households in Charlottesville in Virginia before and after a price per bag was introduced. The volume of their garbage fell by 37% while its weight fell by only 14% [9]. People compressed what they threw away. Volume and weight move independently and any factor that links them is a snapshot of one behaviour at one time.

3.3 The Fill Term#

The fill level is the input most often assumed. GRESB asks that an unknown fill be treated as full [3]. A sensor trial run by the City of Espoo with Wastebook in three primary schools from 17 October 2023 to 15 February 2024 measured fill at the moment of emptying. The school averages for mixed waste containers ranged from 75% to 84%. For plastic containers they ranged from 36% to 94% [4].

A full bin assumption applied to a plastic container emptied at 36% overstates its contents by 2.8 times. Applied to a mixed waste container at 75% it overstates by a third. The trial is small and was run in schools. Its value is that it measured what estimates assume: fill varies by stream and by site and nobody knows it without a sensor or a scale.

3.3.1 Density and Fill Together#

The two terms multiply. Combining the lowest EPA factor with the lowest measured fill for mixed waste gives 27.4 kg for one lift of a 1100 litre container. The highest factor with the full bin assumption gives 195.8 kg [2 3 4]. The same lift can be reported at values 7.1 times apart without any error of arithmetic. Each figure follows a published rule.

3.4 The Container and the Composition#

Two further terms sit outside the volume calculation. The first is the container. In the Wastify AI container register a 1100 litre container has a tare of 58 kg [1]. That is more than the 36.5 kg to 47.6 kg of contents the EPA dry waste factors assign to it when full [2]. A weight recorded with the container still on it can therefore be more container than waste. A net weight requires the tare to be known for every bin type.

The second is composition. A tonnage is assigned to a stream and a stream is assigned an EWC code from the List of Waste. The review of UK municipal waste composition found good agreement between recent surveys of household collected waste. It found insufficient data to allow comparisons of the commercial element of municipal waste and concluded that further work was necessary to produce more reliable estimates [7]. The review of methods for household waste composition studies listed twenty methods and noted that no working international standard had been adopted [10]. If the methods for measuring composition are not standard the assumptions inside an estimate cannot be either. Correct waste classification starts with knowing what was in the container.

Table 4: Where the estimate and the weight part company

Source: [1] [2] [4] [7] and [9].

3.5 What a Scale Records Instead#

A weighed record replaces each assumption with an observation. Each disposal is placed on the integrated scale of the kiosk and recorded through weighing at the point of disposal with the tenant and the stream and a photograph [1]. The container tare register deducts the bin so the recorded kilogram is waste and not container [1]. Across the platform 956000+ weighings covering 5141000+ kg have been recorded. The average disposal is about 5.4 kg [1].

The collection side is measured the same way. With bin by bin collection weighing each bin is weighed and confirmed by stream and timestamped under a contractor PIN before it leaves the building [1]. The building then holds its own figure to set beside the contractor figure and the receiver record. Measurement duties already exist further down the chain. A sorting facility that receives 1000 tonnes or more of recyclable waste a year carries a duty of materials facility reporting. The building is where the waste is produced and it is usually where the record holds no weight at all.

An estimate follows a rule. A weight follows the waste. Wastify AI gives every building its own record of both disposals and collections.

4. Discussion: What This Means for the Reader#

4.1 For Property Managers#

A contractor tonnage is a calculation with at least four inputs and only one of them counted. When it is used to divide a service charge between occupiers every assumption inside it is passed on. The practical test is simple: ask which conversion factor and which fill level produced the figure. If the answer is a contract default the figure is a planning number rather than a record. Checking contractor invoices against weight turns the question from a negotiation into a comparison.

4.2 For Sustainability Leads#

An estimated tonnage carries its uncertainty into every figure built on it. Under the GHG Protocol waste generated in operations is one of the fifteen categories of Scope 3. The line multiplies tonnes by a factor for each route so a lift that could be 27.4 kg or 195.8 kg produces a carbon figure with the same spread. The same is true of a recycling claim. Receiver records under digital waste tracking now state whether each weight is estimated or actual [6]. A report that relies on estimates will sit beside a government record that labels them as such.

4.3 For Contractors#

Nothing in this blog suggests that any contractor calculates wrongly. Each estimate in Table 3 follows a published rule. The issue is that the rules allow a spread wider than any invoice can carry. Contractors that receive weighed collection data from a building can price on it and report on it. Weight measured at the building benefits the recycling chain too: the reprocessor receives material whose stream and quantity were recorded at source.

5. Limitations#

This blog has five limitations that a careful reader should weigh.

1. The EPA factors are US figures drawn from secondary sources. UK waste may have different densities. They are used because GRESB points participants to them and because they show how wide published factors run.
2. The fill levels come from one trial in three Finnish primary schools over four months. Commercial buildings may show different patterns. The trial is used to show that fill varies by stream rather than to give a UK benchmark.
3. The 7.1 to 1 range combines the extremes of separate sources. Most real estimates sit inside it. The point is that the published rules permit the whole range.
4. The peer reviewed studies concern household and municipal waste and household behaviour. They are applied to commercial waste by analogy.
5. Platform totals are rounded down so the average weighing of about 5.4 kg is approximate. The tare of 58 kg applies to the container type in the Wastify AI register and other models may differ.

6. Conclusion#

A volume based estimate of commercial waste multiplies a counted number by a catalogue volume and an assumed fill and an assumed density. The official factors alone allow one full container of 1100 litres to be reported at 36.5 kg or 195.8 kg. Measured fill levels widen the spread for a single lift to 7.1 to 1. The container weighs 58 kg before anything is put in it. That is why contractor waste estimates rarely match actual weight: the rules that produce them allow too much room.

A building that wants to know how much of its own waste data is estimated can take the waste data self check in two minutes.

Put a scale where the estimate used to be. Book a demo with Wastify AI for a one hour session tailored to your buildings.

Frequently Asked Questions#

How do contractors estimate waste weight without weighing it?#

Contractor waste estimates count the containers lifted and multiply by the container volume and an assumed fill level and a density factor. GRESB describes the same four inputs as its permitted method where tonnage is unknown [3].

Which assumption moves the estimate most?#

Density. The US EPA factors for commercial and mixed waste give one full 1100 litre container between 36.5 kg and 195.8 kg [2]. Fill level adds further spread on top.

Why does GRESB assume bins are full?#

GRESB calls it a worst case scenario for cases where the fill level is unknown [3]. It is a conservative rule. A sensor trial found plastic containers emptied at as little as 36% full on average [4].

Does the container weight matter?#

Yes. A 1100 litre container weighs 58 kg empty in the Wastify AI register [1]. A net weight is only possible when the tare of each bin type is known and deducted.

Is the national commercial waste figure measured?#

No. Defra estimates around 32 million tonnes of commercial and industrial waste in England in 2024 and states that it remains extremely difficult to estimate owing to data limitations and data gaps [5].

Will Digital Waste Tracking fix the problem?#

It makes the problem visible. Each receiver record must say whether the weight is estimated or actual [6]. It does not weigh waste at the building or attribute it to an occupier.

Does weighing at source replace the contractor figure?#

It gives the building its own figure to compare. Disposals are weighed net of the container and collections are weighed before bins leave so both sides can work from the same record [1].

Can a building ask which streams carry the most weight?#

Yes. Living AI answers plain English questions about the weighings with the sources attached so the answer comes from the records rather than from a factor.

References#

[1] Wastify AI platform data and published material. Container register of eleven bin types from a 2.5 kg slim bin to a 58 kg container of 1100 litres. Waste tracking and Contractor Mode and coverage features as published. Platform totals of 5141000+ kg and 956000+ verified weighings and 2000+ tonnes CO2e avoided and 380+ companies reporting read from the homepage counter on 12 October 2026. £660000+ recharged on verified data as listed on the same page. wastify.co.uk.

[2] US Environmental Protection Agency. Volume to weight conversion factors. Office of Resource Conservation and Recovery. April 2016. epa.gov. Accessed 12 October 2026.

[3] GRESB. Data estimation rules. 2026 Real Estate Assessment. guides.gresb.com. Accessed 12 October 2026.

[4] City of Espoo. Monitoring of degree of filling improves waste management in Espoo schools. Published 5 June 2024. espoo.fi. Accessed 12 October 2026.

[5] GOV.UK (Defra). UK statistics on waste. Updated 23 September 2026. gov.uk. Accessed 12 October 2026.

[6] legislation.gov.uk. The Digital Waste Tracking (England) Regulations 2026: Schedule 1. SI 2026/729 paragraph 14. legislation.gov.uk. Accessed 12 October 2026.

[7] Burnley SJ. A review of municipal solid waste composition in the United Kingdom. Waste Management 27(10) pages 1274 to 1285. 2007. doi.org. Abstract read at the Open University repository. Accessed 12 October 2026.

[8] Wastify AI Research Blogs 01 to 05. October 2026. wastify.co.uk.

[9] Fullerton D and Kinnaman TC. Household responses to pricing garbage by the bag. American Economic Review 86(4) pages 971 to 984. 1996. ideas.repec.org. Accessed 12 October 2026.

[10] Dahlén L and Lagerkvist A. Methods for household waste composition studies. Waste Management 28(7) pages 1100 to 1112. 2008. doi.org. Indexed in PubMed. Accessed 12 October 2026.