Explore

Estimating the climate impact of an annual Mill Membership
Estimating the climate impact of an annual Mill Membership

Background

At Mill, we’re building an entirely new system to outsmart waste, starting with food.

Uneaten food is the

most common material in landfills⁠

and most of it comes from

households⁠

. When that food rots, it produces methane, a greenhouse gas that has about

80x⁠

the global warming potential of carbon dioxide over a 20-year period.

With a Mill Membership™, members receive a kitchen bin that dries and grinds kitchen scraps and turns them into compact, nutrient-rich Food Grounds™. Once the bin is full – which takes a few weeks – members can send their Food Grounds to Mill. From there, we can keep your Food Grounds in the food system by turning them into food for chickens.

We use a life-cycle assessment (LCA) to quantify the climate impact of an annual Mill Membership. Based on our preliminary study, a household can avoid about a half-ton of greenhouse gas emissions per year with a Mill Membership.

By publishing our initial estimates, we want to invite more people to help shape how we think and how we outsmart waste. In the future, we will be able to report on impact in more detailed ways as we gather data from households with Mill Memberships. We anticipate our numbers to change as members reduce the kitchen scraps they generate and as sending food to landfills becomes less common in the U.S.

We’re taking a phased approach (

see below⁠

) to account for the ability to move from estimates to actual data.

Scoping LCA

Our Scoping LCA is a kitchen-scraps-to-feed gate comparison of the Mill Membership (that includes the Mill kitchen bin and Mill’s operations, where a feed ingredient is produced and shipped to farms) to a counterfactual scenario without the Mill Membership (where kitchen scraps are landfilled and some are composted).

We describe emissions avoided by an annual Mill Membership, where a year’s worth of a household’s kitchen scraps is the functional unit, and global warming potential, measured in kilograms of CO2 equivalents (kg-CO2e), is the environmental indicator.

Cradle-to-feed-customer-gate describes the upstream and downstream extents of the system boundaries, specifically inclusive of extraction of raw materials all the way through production of a feed ingredient.

⁠

UPDATED 02.22_System diagram for Scoping LCA.png

⁠

Figure 1. System diagram for Scoping LCA of an annual Mill Membership

External validation

Mill worked with external independent parties to gather feedback on inputs, assumptions, and methodologies used in the Scoping LCA. The goal of this process was to make the model as fair and unbiased as possible. These parties included:

⁠

Franklin Associates⁠

: a division of Eastern Resource Group (ERG) who we are working with to inform our LCA model and to guide future developments and results benchmarking.

⁠

Synapse Product Development⁠

: a professional engineering services firm with expertise in sustainable design practices who built a SimaPro life-cycle assessment to account for the carbon emissions of the materials and manufacturing of the device and its components.

⁠

Energy Impact Partners⁠

(Mill investor): a strategic investment firm with a team of LCA experts who reviewed our model structure and transportation calculations.

⁠

Dr. Michael Walsh⁠

: a decarbonization researcher who reviewed our landfill and compost calculations.

Approach

We carried out a series of literature reviews to identify the most reputable reference (

see below⁠

) available for impact accounting methodologies, assumptions, and emission factors. We then evaluated four scenarios:

Scenario

Operations

Assumptions

Scenario 1

Near-term

Conservative

Scenario 2

Near-term

Optimistic

Scenario 3

Long-term

Conservative

Scenario 4

Long-term

Optimistic

There are no rows in this table

⁠

Here are the definitions for each scenario:

Near-term operations

Refers to less optimized end-to-end operations projected for 2023 where Mill’s resource efficiency, energy efficiency, transport distances, and feed ingredient manufacturing process yield have room for improvement with time and product/process refinements.

Long-term operations

Refers to more efficient end-to-end operations projected for 2027 where Mill deploys memberships at greater scale, optimizes bins and processes for greater efficiency and yield, and source energy from a cleaner grid.

Two examples of parameters anticipated to become more efficient over time:

Mean daily electricity use of Mill kitchen bin: reduced through both hardware and algorithm design optimization

Feed ingredient manufacturing process yield: increased through process and capital improvements

Conservative assumptions

Selected to emulate a more pessimistic scenario, where processes are, for example, assumed to be inefficient, energy-intensive, further apart, and products are shorter-lived and heavier.

Optimistic assumptions

Selected to emulate a more optimistic scenario, likely possible to achieve in the timelines modeled in the near- and long-term scenarios.

Considerations

We conducted our Scoping LCA using both 20-year and 100-year

global warming potentials ⁠

(GWP20 and GWP100). This means that we applied emission factors (EF) with 20- and 100-year time horizons to convert the mass of non-CO2 greenhouse gases (GHGs) into CO2-equivalents (CO2e). Using both GWP20 and GWP100 to calculate avoided CO2e emissions allows us to understand our climate impact on both short-term and long-term timescales.

The Intergovernmental Panel on Climate Change (IPCC) classifies methane emission factors as either biogenic emissions or fossil emissions. Biogenic emissions are the emissions produced from recently-created natural sources, such as methane emitted from anaerobic digestion. It is common in LCAs to omit biogenic CO2 emissions from materials of recent biological origin, such as CO2 emissions from composting, because it is considered to be a part of the short-term carbon cycle. Fossil emissions result from the combustion or oxidation of fossil fuels, such as burning jet fuel. In this model, biogenic CO2 is not counted towards Mill or Counterfactual emissions.

GWP20

Focuses on contributions from short-lived gases, such as methane.

Uses EF 80.8 for biogenic methane with 20-year time horizon (

IPCC AR6 Ch7 Table 7.15⁠

GWP100

Strikes a balance between short-term and long-term climate impacts.

Uses EF 27.2 for biogenic methane with 100-year time horizon (

IPCC AR6 Ch7 Table 7.15⁠

It is common practice for ISO-compliant LCAs to use the 100-year time horizon.

The stark difference in methane GWPs over different time frames is due to the fact that methane is short-lived in the atmosphere, sticking around for roughly a decade (as compared to over a century for CO2). During its short atmospheric life, methane traps significantly more heat than CO2, thereby increasing the risk of

passing potentially catastrophic climate tipping points⁠

. Even though CO2 has a longer-lasting effect, methane sets the pace for warming in the near term.

In 2018, Balcombe and colleagues did a deep dive on the appropriate uses of the 20- and 100-year methane GWP emission factors. To paraphrase, they found that although reporting both is the safest approach, the 20-year factor was appropriate for assessing the impact of technologies targeted at mitigating methane emissions (

Balcombe et al., 2018⁠

⁠

UPDATED 02.22_Methane global warming potential emission factor over time.png

⁠

Figure 2. Methane (CH4) global warming potential emission factor (GWP EF) over time. Methane is a potent greenhouse gas in early years and because it is a short-lived climate pollutant, its impact relative to CO2 diminishes over time (since CO2 retains a nearly constant radiative forcing for >100 years). 20-year and 100-year GWP EF’s reference

IPCC AR6 CH7⁠

. Figure produced from public data:

AR6 CO2 CH4 Response by Chris Smith, University of Leeds⁠

Findings

It is important to re-emphasize that this is a preliminary study, and that we made a best-faith effort to build a holistic model with fair assumptions based on our knowledge today. Our inputs will change in the future as we gather operational data from households with Mill Memberships.

When taking the difference in the CO2e impact between the counterfactual and Mill, we can quantify the emissions avoided by a Mill Membership (in units of kg-CO2e/year). Ranges, showing the values between conservative and optimistic scenarios for both near-term and long-term cases, frame uncertainty in assumptions.

According to our Scoping LCA, a counterfactual world without Mill appears to incur a much greater cost to our climate than a world where Mill is in operation. This means that the emissions avoided by Mill are greater than the energy and resources invested for Mill to operate. Landfill emission avoidance is the largest contributor toward the net environmental benefit of a Mill Membership.

Here is a summary of our findings:

Operations

Assumptions

Avoided emissions (kg-CO2e per annual Mill Membership) using GWP20

Near-term

Conservative

471.5

Near-term

Optimistic

570.6

Near-term

Mean

521.1

Long-term

Conservative

565.9

Long-term

Optimistic

702.8

Long-term

Mean

634.4

There are no rows in this table

⁠

Here is a waterfall breakdown of the results for our near-term scenarios:

⁠

UPDATED 01.17_Estimated annual Mill Membership  emissions generated and avoided_Near-term.png

⁠

Figure 3. Waterfall plot showing values for near-term scenarios using GWP20.

Here is a waterfall breakdown of the results for our long-term scenarios:

⁠

UPDATED 01.17_Estimated annual Mill Membership  emissions generated and avoided_Long-term.png

⁠

Figure 4. Waterfall plot showing values for long-term scenarios using GWP20.

Here is a chart of the results for each scenario using GWP20 and GWP100:

⁠

UPDATED 02.24_Total potential avoided emissions across different scenarios.png

⁠

Figure 5. Comparison total net potential avoided emissions of an annual Mill Membership based on using GWP20 and GWP100 emission factors for methane.

We organized the results of Mill Membership scenario in the following groups:

Mill kitchen bin production & refurbishment:

Scope: Includes Mill kitchen bin manufacturing (extraction and processing of raw materials, component manufacturing, assembly), packaging, energy to transport the bin to the member’s home, reverse logistics of bin refurbishment, and decommissioning at end-of-life.

Result: Around 80% of the climate impact in this process can be attributed to the annualized greenhouse gas emissions from manufacturing the kitchen bin. Most of the remaining impact comes from refurbishment.

Mill kitchen bin energy usage & maintenance:

Scope: Includes electricity consumption during normal bin use and reverse logistics of charcoal odor filter replacements.

Results: Electricity consumption is the primary source of emissions for this process group. It is also the primary source of emissions for the Mill Membership overall, although still small relative to the counterfactual).

Food Grounds shipping:

Scope: Includes the emissions invested to make Food Grounds mailer boxes and liners, to ship multipacks of mailers to member homes, and to ship Food Grounds-filled mailers to Mill.

Result: Although the impact of shipping is non-zero, it is fairly small relative to the bin’s electricity use. By leveraging trucks that are already on the road, shipping transport emissions are measured as nominal ton-miles – the energy to carry the additional mass of Food Grounds from the home to the feed ingredient manufacturing facility. Based on the ton-mile emission factors for medium- and heavy-duty vehicles, one year of a household’s Food Grounds could be transported 15,000–20,000 miles before the carbon cost outweighs the benefits of a Mill Membership.

Feed ingredient production & operations:

Scope: Includes the energy to turn Food Grounds into a feed ingredient, as well as bulk pack and ship the feed ingredient to the customer (formulator or farm). Also includes the management of the process co-product (material ineligible for feed) and process residual (plastic or metal contaminants). Composting of the co-product in the near-term scenario factors transport to compost facility, composting process, pack out, land application, fertilizer avoidance, and post-application degradation in the soil.

Result: The impact of this is related to (specifically, inversely proportional to) feed manufacturing process yield. With process optimization, the impact of this group of processes will decrease.

We organized the results of the counterfactual scenario in the following groups:

Compost collection & processing avoided:

Scope: Includes transport of scraps from the home to a compost facility, the process of biogenic and fossil emissions during composting, product packaging (for the portion of product sold into bagged markets), transport to customers, land application of compost, and fertilizer avoidance.

Result: Most of the impact results from compost process fugitive emissions, non-CO2 compounds emitted from the compost pile during the composting process.

Landfilling avoided:

Scope: Includes collection and transport of scraps to landfill, process biogenic methane, and process fossil emissions from landfill equipment.

Result: This is the largest impact in the whole system. Because of the potency of methane, the fairly small mass of methane emitted from landfilling one household’s kitchen scraps results in hundreds of kg of CO2e, on the order of 10–20 kg of methane. This is based on EPA WARM v15 figures on the properties of “Food Waste” and their national-average behavior in U.S. landfills, and accounts for landfill gas capture and methane oxidation in the near-surface of the landfill. See landfill methane emissions generation rate section below for more information.

Feed production avoided:

Scope: Includes grain agriculture, grain transport, grain processing into a feed ingredient, and transport to a customer.

Result: Grain agriculture accounts for most of the climate impact of this process, while transport and feed ingredient production account for a relatively small portion.

Future opportunities

As an early-stage company, there are numerous opportunities to drive down the greenhouse gas emissions associated with the device and device use, as well as optimize the process.

Below are a few key areas considered in our long-term scenarios that can improve the climate impact of an annual Mill Membership over time:

Improving efficiency of Mill kitchen bin production and operation (e.g. software improvements).

Improving logistics related to Food Grounds shipping and feed ingredient distribution (e.g. facility siting).

Improving energy efficiency of our facilities (e.g. renewable energy mix).

Improving process yield of feed ingredient manufacturing facility.

Technical appendix

Phased approach

Today - Scoping LCA

An initial modeling effort to directionally indicate the impact of an annual Mill Membership and inform our decision-making.

We’re designing for climate impact, from the beginning, by using a Scoping LCA to identify “hot spots,” or areas of our business which are disproportionately impactful and should be targeted for improvements.

Future - ISO LCA

This is the most common form of life-cycle assessment and follows the steps laid out in the 140040 and 140044 ISO standards developed by the International Standards Organization.

These include quantitative methods for the assessment of the environmental aspects of a product or service in its entire life cycle. The process is commonly broken down into four stages:

Goal and scope definition

Life-cycle inventory analysis

Life-cycle impact assessment

LCA interpretation (final report)

The ISO LCA process takes time (often up to a year) as it requires significant detail of all inputs and variables. We plan to initiate an ISO LCA with an independent, external third party once we have the requisite months of operational field data.

Future - Comparative ISO LCA

This is the traditional gold standard tool in industry when the goal is to make comparative claims. These analyses take longer than standard ISO LCAs.

In addition to the Mill Membership, we would assess each product system that it will be compared against, including one assessment for sending kitchen scraps to landfill and another assessment for composting.

Exclusions

A number of activities are excluded from this assessment:

Emissions from the oxidation of biogenic carbon in the feed ingredient.

In the future, we plan to build out a model that assesses the consumption of the feed ingredient. This is a non-trivial assessment with many factors to evaluate, such as: formulation inclusive of Mill feed ingredient, studying feed conversion ratio of that ration, a poultry digestibility study, research on carbon assimilation into eggs and meat, fate of carbon in poultry manure as a function of manure management method, the assimilation and loss of carbon in eggs and meat consumed by humans.

Helping drive behavior change around source reduction.

Our model can be considered conservative because it does not account for source reduction. Our estimates will improve if data suggests that an annual Mill Membership helps reduce household kitchen scrap generation rates.

The climate impact of Mill corporate operations unrelated to device manufacture, Food Grounds shipping, or feed ingredient manufacture.

Land use and soil carbon implications of offsetting feed grain agriculture.

Reduced footprint of waste and compost collection resulting from less scraps wasted.

Indirect effects from Mill operations, such as economic benefits in localities of Mill offices or feed ingredient manufacturing facilities.

Key assumptions

As with any model, decisions must be made about what falls within the bounds of the assessment, and what fair assumptions can be made about material flows, process activities, and accounting methods. In order to provide transparency, the following section will walk through inputs and assumptions which the model is most sensitive to:

Device electricity use

Electricity consumption varies with usage behavior and the amount of kitchen scraps added. The model currently bases this assumption on power use observed during a 20-device 2-month-long field trial (cumulatively more than 1,100 non-zero power use device days).

Parameter

Near-term (2023) conservative

Near-term (2023) optimistic

Long-term (2027) conservative

Long-term (2027) optimistic

Device electricity use (long-term rolling mean, kWh/day)

0.852 (from field trial data)

0.75 (2023 target)

0.75

0.5 (2027 target)

There are no rows in this table

⁠

Local grid energy mix

The proportion of a Mill member’s local electric grid mix that is fossil-derived energy affects the emissions of using the Mill kitchen bin. Although the grid is becoming progressively cleaner over time, this is a slow process. Households do have the power to significantly reduce the environmental impact of the bin and of their home overall by purchasing renewable electricity.

For both the conservative and optimistic near-term scenarios, we used the 2021 national average grid emission factor, as reported by the Energy Information Administration (

EIA Frequently Asked Questions⁠

), because there is significant historical data supporting this factor.

For the long-term conservative scenario, we referenced EIA’s 2027 forecast (

EIA, Annual Energy Outlook, Table 8⁠

) because it assumes a business-as-usual approach and no climate action taken. For the long-term optimistic scenario, we used the International Energy Association’s 2027 forecast (

IEA⁠

) because it assumes an aggressive approach to achieve a clean energy transition. We used different values for our two long-term scenarios because of significant uncertainty in the future U.S. grid mix.

Parameter

Near-term (2023) conservative

Near-term (2023) optimistic

Long-term (2027) conservative

Long-term (2027) optimistic

Electric grid mix emission factor (kg-CO2e/kWh)

0.388 (U.S. 2021 mean)

0.306 (EIA 2027 forecast)

0.138 (IEA 2027 forecast)

There are no rows in this table

⁠

Landfill methane emissions generation rate

The impact of placing kitchen scraps into landfill (key metric being kg-methane/short ton-kitchen scraps landfilled) was derived based on values (specific to “food waste”) reported in

EPA WARM 2020 v15 data⁠

Parameter

Value (constant across all scenarios)

Moisture content of kitchen scraps

72.2%

Carbon content of kitchen scraps

49.5%

Kitchen scraps-carbon to methane-carbon conversion yield

42%

Mass methane-C to mass methane

1.34

Landfill gas collection efficiency, national average for kitchen scraps

52%

20-year (biogenic) methane global warming emission factor (GWP20)

80.8

100-year (biogenic) methane global warming emission factor (GWP100)

27.2

Counterfactual fate of kitchen scraps: %-landfilled

80%

Counterfactual fate of kitchen scraps: %-composted

20%

There are no rows in this table

⁠

Kitchen scraps generation rate

The amount of kitchen scraps generated per person and the number of people in a household affect the rate of kitchen scrap generation. If a Mill member lives alone or travels a lot, they could have a lower kitchen scraps generation rate per device per day.

The model currently assumes 1.326 lbs-kitchen scraps/household/day, based on ReFED’s Residential Surplus Food statistic of 30M short tons generated in the U.S. in 2019 (

ReFED⁠

) and using

Census data⁠

suggesting 124.01M households in the U.S.

Parameter

Value (constant across all scenarios)

Kitchen scraps generation rate (lbs-wet scraps/household/day)

1.326

There are no rows in this table

⁠

Counterfactual fate of kitchen scraps

The model currently assumes 80%-to-landfill : 20%-to-commercial-compost, for all counterfactual scenarios. This data was based on findings from Mill’s market research studies, specifically referencing what respondents deemed “most likely to subscribe” currently do with their kitchen scraps.

Changing this landfill : compost split in the counterfactual scenario affects the marginal greenhouse gas emissions avoided by an annual Mill Membership. This is because landfilling results in greater emissions than composting: 1.29 kg-CO2e/lb-kitchen scraps landfilled (GWP20), compared to 0.235 kg-CO2e/lb-kitchen scraps composted (assuming composting emissions account for transport, process, and fugitive emissions) and not counting biogenic CO2 (

EPA WARM Management Practices⁠

San Joaquin Valley Air Pollution Control District⁠

). These estimates are based on calculations from our Scoping LCA.

Mill plans to update this study with findings that compile a full year of operational data. Check back for an update later in 2024.

References

⁠

Argonne National Labs’ GREET 2022⁠

, CARB-published (California Air Resources Board) on-road fuel production pathways were referenced for transport fuel emission factors based on model to assess the lifecycle pathway emissions for fuels and chemicals.

⁠

Balcombe et al, 2018, Methane Emissions: Choosing the Right Climate Metric and Time Horizon⁠

, referenced for rhetoric on the use of 20- and 100-year global warming potentials.

⁠

Ecoinvent 3 LCA Database⁠

, referenced by to quantify the climate impact of Mill kitchen bin bill of materials (BOM) and manufacture.

⁠

EIA Electricity Frequently Asked Questions⁠

, referenced for the 2021 national average electric grid emission factor.

⁠

EPA Emission Factors Hub⁠

, referenced for common emission factors.