Data transparency

Energy

In 2014 we focused on obtaining energy-use data from all our offices and data centers around the world. Our office managers’ commitment to sustainability allowed us to get electricity usage data from all nine of our offices, even offices that share spaces and meters with other tenants. In some cases we had to extrapolate our usage based on the data obtained.

In offices where we share meters or fuel supplies with other companies, we estimated our usage based on a percentage of the space we occupy. In coworking spaces where we do not occupy a defined space, we estimated our usage based on our percentage of the space’s total headcount. For those offices where a fuel other than electricity is used for heat or hot water generation, it was more challenging to obtain the data we needed. Wherever possible, we used available information and where data was unavailable, it was noted in our availability matrix below.

We used a variety of emissions factors to translate our energy use into greenhouse gas emissions. For electricity in our U.S. offices and data centers, we used the Greenhouse Gas Protocol Initiative worksheet, Emissions From Purchased Electricity version 4.5 (revised May 1, 2014), which employs the U.S. Environmental Protection Agency’s eGrid emissions factors for 2010. In cases where we had access to electricity supplier emissions factors, these factors were used. For our Melbourne office, we used the figure for CO2e emissions that the utility company provided each month with the bill. In Toronto, we used the emissions factor published in Toronto’s 2012 Greenhouse Gas and Air Quality Pollutant Emissions Inventory. Our London office is able to purchase renewable energy from a supplier through the grid, so our emissions were zero in London. In Berlin and Seattle, our electricity comes from the standard grid mix, which is mostly renewable. In Berlin it is 95.9% renewable with the remaining 4.1% from natural gas. We calculated our emissions based on 4.1% of of our total kWh from natural gas using the Greenhouse Gas Protocol Initiative worksheet, Emissions From Purchased Electricity. For our Seattle-based data center, we assumed zero greenhouse gas emissions based on the fact that the fuel mix is approximately 94% renewable, and the utility purchases offsets the remaining portion that is not renewable. Emissions from our content distribution network, for the most part, are calculated directly by the provider, and we do not have access to a full estimate for kWh used. For natural gas or distillate fuel oil no. 2 combustion, we used emissions factors from the U.S. Environmental Protection Agency.

Estimating greenhouse gas emissions proved to be challenging in a few cases. After months of collecting kWh usage for electricity in our Dublin office, we found that our electricity and heat are actually generated on-site by the Diageo brewery. We were not able to secure the details of that process, including the type of fuel used, so our 2014 greenhouse gas footprint does not include the Dublin office. We expect this to change in 2015. Additionally, while we have a kWh usage estimate in Berlin, we were unable to get exact emissions factors for the district steam system in Berlin. Instead, we used generic emissions factors for district heating from combined heat and power systems fueled by natural gas in the 2006 Revised Intergovernmental Panel on Climate Change Guidelines for National Greenhouse Gas Inventories Reference Manual.

Renewable energy

To calculate the percentage of renewable electricity used by our offices and data centers, we combined the kWh used in offices where we purposefully purchased renewable energy from suppliers through our utility companies, and, in cases where we were not able to purchase a renewable supply, the percentage of renewable energy in the standard grid mix. Our London office purchased 100% of its energy as renewable, and our Melbourne office purchased 22% of its energy as renewable. The energy supplied by the grid to our Berlin office is 95.9% renewable, and the grid-supplied energy in our Seattle data center is considered 100% renewable. When we purposefully purchased just a portion of renewable energy for an office, we assumed that the remainder is from non-renewable sources. We did not have access to the kWh used by our content distribution networks.

The percentage of renewable energy in standard grid-supplied electricity was calculated for our remaining locations by taking fuel source public disclosure statements from each utility company that we use. These statements were found either on the utility’s website or on state-based public information websites. We used the most recent figures available, but the disclosure dates ranged from 2012 to 2014. In one case, for a facility we no longer use but used for a partial year in 2014, we calculated the percentage of renewable energy in the grid mix from 2010 eGrid data from the U.S. Environmental Protection Agency.

We only calculated our percentage of renewable energy for our electricity usage. We did not calculate it for other sources of energy, such as the heating fuels used in some of our offices. Some of these sources are renewable, such as the biogas we use for heat and hot water in London, and others are not, like the natural gas that heats our Hudson building.

Water

Reliable water usage data has been extremely difficult to obtain across all of our offices. The only office for which we have exact 2014 data is our London office. In our Brooklyn office we have 2012 and partial 2013 data for the whole build- ing. To estimate the portion of our water use in Brooklyn, we compared total usage with the percentage of the building that we occupy. We would have been more confident deriving an estimate based on a percentage of headcount, but we did not have access to building occupancy figures.

To extrapolate water usage to all of our offices, we came up with a per-person estimate in London and Brooklyn based on the headcount over the period of time for which we had data. We used the London estimate to extrapolate usage for our Australian and European offices, and used the Brooklyn estimate for our North American offices.

Our 2014 water usage figures do not include estimates for work-based water used at home or in coworking spaces used by remote workers.

Waste

In 2014 we were able to collect almost a full year of waste data from four of our offices—Berlin, Brooklyn, Hudson, and London—which account for approximately 81% of our employee population.

In our Berlin and London offices, waste is weighed and recorded manually on a daily or weekly basis depending on the waste stream. In Brooklyn and Hudson we are using an automated waste measurement prototype devised by our Office Hackers team, in which scales connected to our network send daily measurement data directly to our collection portal. We have already successfully implemented this new system in Dublin and Toronto, and we expect to roll it out to our San Francisco office later in 2015.

In cases where we were missing data from a select week, we estimated it based on the average waste from the four weeks immediately surrounding the missing week. In instances where we were missing data from consecutive weeks in a row, we estimated those weeks based on the average waste from the 12 weeks immediately surrounding the missing weeks. In locations where we collected plastic bags for separate recycling, we combined those figures with the amounts of plastic included in the bottles and cans category, as the plastic bag category is relatively small.

Our 2014 waste figures do not include estimates for work-based waste generated at home or in coworking spaces by remote workers.

Energy, waste and water data availability for 2014

Business travel

To arrive at our emissions calculations for business travel, we worked with an external partner, Closed Loop Advisors. Trip information was pulled from travel expense reimbursements (supplemented with more detailed routing and class of service information where available) from sources such as American Express corporate card reports. 2014 emissions factors came from DEFRA, the United Kingdom’s environmental agency. Different emissions factors were used for short, medium, and long-haul flights, and all of these included an 8% increase to account for indirect flight patterns and circling. A blend of emissions factors was used based on Etsy’s mix of service classes, primarily coach and premium economy, with less than 5% business and first. A radiative forcing factor of 1.9 was applied to account for the net additional non-CO2 warming effects of aircraft emissions (e.g., NOx, soot, contrails, high-altitude emissions), as per DEFRA guidance.

Employee commuting

In 2014 we used an updated survey methodology in an effort to more accurately capture our employees commuting behaviors and the associated estimated carbon emissions. We surveyed 352 employees (out of a total of 580), and extrapolated the results across the entire employee population to arrive at the estimates published in this report. The survey-based methodology used to calculate emissions from commuting are in line with the GHG Protocol: http://www.ghgprotocol.org/files/ghgp/Chapter7.pdf.

In our survey we asked Etsy employees to indicate which modes of transportation they used on each of five workdays. We then summed the total number of modes, and calculated a propor- tion based on the modes they took in a given week. For example, if someone said they took the subway four days, and a cab one day, then the proportion of their time spent respectively is subway: .8, and car (shared or alone): .2. We then multiplied their mileage (calculated based on zip code) by 10 (assuming that they go that distance each way, two times a day), and then multiplied this mileage by the proportion spent traveling their weekly distance to calculate weekly mileage per mode of transportation.

We calculated the distance each employee travels using zip or postal code. The distance used was “as the crow flies,” not driving distance. We used “as the crow flies” distance instead of driving distance due to variations in mode routes: Ferry is direct; the bus is not; the subway is more direct, etc.

The following websites were used to calculate these distances:

• United States & Canada: http://www.zip-codes.com/
• Other markets: http://distancecalculator.globefeed.com/Country_Distance_Calculator.asp

If we couldn’t calculate the distance between two zip codes, we used the midpoint of the reported range of distance.

Marketplace shipping

To analyze and calculate our emissions from shipping items sold on Etsy, we worked with an external partner, Closed Loop Advisors. The analysis includes emissions from transportation (e.g., postal service vehicles and planes) and mail handling fa- cilities (i.e., post offices and sorting facilities) as well as detailed research on the fleets, facilities, and practices of the national postal services of the top six countries where Etsy transactions occur. This research covered over 85% of shipments, and these findings were extrapolated to the rest of the world based on each country’s size and population density. Data sources includ- ed academic studies on mail routing and conversations with postal services contacts, as well as annual reports and sustain- ability reports from the postal services (United States: Annual Report, Sustainability Report; United Kingdom: Annual Report, Sustainability Report; Canada: Annual Report, Sustainability Report; Australia: Annual/Sustainability Report; France: Annual Report, Sustainability Report; and Germany: Annual Report, Sustainability Report).

Transportation emissions

Transportation emissions were calculated for each shipment based on the distance from the buyer to the seller, the average weight of the category of item(s) included in the shipment, and a blend of emission factors for the different types of vehicles in which the shipment would have traveled, depending on its route.

Distances

Distances were based on the coordinates (latitude and longi- tude) of the buyer and seller postcodes for the six countries included in this analysis, with adjustments for routing through shipping facilities, and indirect driving distances and flight routes. For the remaining countries, we used the coordinates of the country’s largest city plus an adjustment for expected average distance within the country based on its area and popu- lation density (taken from the CIA WorldFactbook). First and last mile distances from residences to post offices were estimated in the U.S. by categorizing zip codes as urban, rural, and suburban, and using Google maps to obtain driving distance to the nearest post office for a sample of zip codes in each category (as well as a sample of U.K. postcodes). Average distances for rural and for urban/suburban were then applied to all countries based on their level of urbanization (from World Bank data).

Weights

Weights were based on the average weight of each category of item in a large sample (over 20 million records) of U.S. and Canadian transactions. Average weights by category were calculated for intra-country and cross-border shipments, and applied accordingly to the full universe of transactions.

Emissions factors

Emissions factors were based on 2014 emissions factors from DEFRA, the United Kingdom’s environmental agency. The fleets for each of the six primary Etsy countries were grouped by expected usage for the various legs of the journey: first and last mile between residence and post office, transport between district and regional sorting facilities or between regional sorting facilities. The appropriate emission factors were selected based on the description of the vehicles in the annual report. A blended emission factor for each leg was created based on the expected proportion of mail carried by each type of vehicle on that leg of the journey. For air travel, different factors were used for short, medium, and long-haul flights, and a radiative forcing factor of 1.9 was applied to account for the net additional warming effects of aircraft emissions, as per DEFRA guidance.

Facility emissions

Facility emissions were calculated based on annual report disclosures from the postal services in the top six countries. Facility Emissions were split into emissions attributed to letter handling and parcel handling, and parcel-related emissions were divided by the number of parcels handled to obtain Facility Emissions per parcel. The average Facility Emissions per parcel for the six countries researched was applied to the remaining countries.