Carbon Footprint for Google Ads

Carbon Footprint for Google Ads encompasses emissions arising from the following activities:

• Scope 1
  • Fossil fuels combusted on-site, such as diesel for backup power, natural gas for heating, and fuels used in fleet vehicles.
  • Fugitive emissions from data center HVAC system coolants.
• Scope 2
  • Google advertising products electricity use, including that from Google-owned compute and networking equipment and ancillary electricity services such as cooling and lighting, whether inside a Google-owned data center or a facility owned by others (location-based and market-based calculations). This also includes electricity use from Google owned & operated servers operating outside of the data center. Note that outside of the data centers, equipment is located in a variety of location types (many of which are not owned by Google) so the industry average PUE will be used to estimate operational efficiencies of ancillary electricity services.
• Scope 3
  • Upstream lifecycle (embodied) emissions of data center equipment.
  • Upstream lifecycle (embodied) emissions of data center buildings.
  • Business travel and commuting associated with employees who work at Google data centers.
  • Generating electricity that is subsequently lost during transmission and distribution.
  • Extraction, production, and transportation of fuels used to generate grid electricity.

Carbon Footprint for Google Ads excludes emissions arising from the following activities:

• Emissions from Google networking equipment deployed outside data centers (except Google’s content delivery network servers which are included).
• Downstream end-of-life emissions of data center equipment and buildings.
• Embodied emissions associated with grid electricity generation facilities and equipment.

The customer-specific Carbon Footprint for Google Ads report (the ads carbon reporting), is calculated automatically. This section describes how these calculations are made.

Key concepts

• Google advertising products use a shared computing platform. These compute resources -- processing power, memory, storage, networking, etc. -- are shared across many advertising customers.
• Google is organized around units of functionality called internal services. An internal service is a particular software functionality that's run on Google's data center machines. Google advertising products use many of these internal services.
• Electricity use is one of the largest sources of greenhouse gas emissions powering advertising. Data centers consolidate compute resources into shared buildings. These buildings consume electricity to run the computing equipment and additional power for lights, cooling, power systems, and other ancillary needs.
• Electricity is provided by a wide variety of generation plants operating on individual grids all over the world.The greenhouse gases arising from electricity generation vary with the generation fuel (e.g., natural gas, coal, wind, sun, water) among other factors. Each grid's generation sources differ, and within a grid, the sources will differ over the course of a day.
• Disaggregating electricity use for advertising and its resulting carbon footprint to specific customers presents a technical conundrum. Determining a customer's footprint is very complex due to the layers of shared resources called upon to serve customer compute needs. Developing new allocation methodologies and assumptions (as discussed in detail below) enables customer footprint reports to be presented which are appropriate and representative of each customer's computing use and product choices.

Calculation summary

Carbon Footprint for Google Ads first calculates energy use as a function of compute usage and data center resource requirements. Then, Carbon Footprint for Google Ads calculates location-based and market-based carbon emissions from electricity use in addition to proportional allocations of emissions arising from non-electricity sources.

The market-based emissions metrics match Google's clean electricity purchases to relevant data center loads to establish regional market-based electricity emission factors wherever Google purchases clean energy. In the market-based emission reports, the regional market-based emission factors replace the location-based emission factors.

Energy use

Energy use and allocation to internal services for data center machines

To allocate the total machine energy usage to internal services, Google separately evaluates the energy used when running a workload ("dynamic power") against the energy used when machines are idle ("idle power"). Each machine's hourly dynamic power is allocated to the internal services it supported that hour, based on relative internal service CPU usage. Machine idle power is allocated to each internal service based on its resource allocation (CPU, RAM, SDD, HDD) in the data center.

Overhead energy use–power systems, cooling and lights–is allocated hourly to every machine within the data center, and its users, based on the machine's total energy use that hour.

Google's shared infrastructure services track the usage of other internal services that call them. This enables the shared infrastructure services' energy usage to be reallocated to those internal services based on their relative usage. For some internal services that don't have sufficient usage data, Google uses internal costs to re-allocate the shared infrastructure's energy consumption.

When these calculations and allocations are complete, we have hourly power use allocated to each internal service in each data center.

Energy use and allocation to internal services for Google’s Content Delivery Network

Energy usage of machines outside of the datacenter, although monitored, isn’t as comprehensive as machines within the data center. A conservative approach has therefore been taken to estimate power usage, by calculating the maximum power consumption per machine type on a global basis, and making the assumption that each machine of this type then uses this amount of power at all times.

Unlike in Google-owned data centers where actual data center Power Usage Effectiveness (PUE) is used as an input into calculating overhead energy use, an industry average PUE is used as the majority of these locations are not owned by Google.

When these calculations and allocations are complete, we have hourly power use allocated to each internal service in each Content Delivery Network location.

Greenhouse gas emissions

Greenhouse gas emissions from electricity: location-based calculations

Google calculates location-based greenhouse gas emissions on an hourly basis by multiplying location-specific energy use by a grid electricity carbon emission intensity factor. This reflects the actual mix of power sources (fossil fuels, renewables, etc.) feeding into the grid in the location of energy consumption. Notably, location-based Scope 2 emissions don't account for energy purchasing choices or contracts – like energy attribute certificates (EACs) or power purchase agreements (PPAs).

The hourly grid carbon emission intensity data used as input into Carbon Footprint for Google Ads encompasses only the emissions associated with electricity generation; it does not include other lifecycle stages. The hourly emission factor data is provided by Electricity Maps. Where Electricity Maps data is unavailable, Google uses country-specific annual average carbon emission intensity factors published by the International Energy Agency.

To calculate emissions, Google multiplies the hourly energy use for each internal service at each location by the appropriate carbon emission intensity factor for that hour and location to determine the internal service's location-based electricity carbon footprint per hour and location.

Greenhouse gas emissions from electricity: market-based calculations

The market-based electricity footprint is estimated by matching Google's clean electricity purchases to relevant data-center loads according to GHGP standards.

Google calculates its market-based emissions on an annual basis, taking into account the actual generation of our clean electricity contract facilities and the electricity used at each site. Google estimates market-based Scope 2 emissions using publicly-available annual grid emissions factors from government sources, including those published by the International Energy Agency and U.S. Environmental Protection Agency.

In each region where we purchase clean electricity, a market-based annual emissions total is calculated for Google's data centers. The location-based electricity emissions are scaled down using the ratio of the previous year's renewable electricity percentage in the region. This scaling factor is multiplied by the granular location-based emissions calculations in the appropriate region to create monthly market-based emission reports broken down by customer.

The market-based scaling factor is updated once per year, as it relies on the market-based emissions calculations for Google as a whole. For this reason, the market-based emission reports don't provide a dynamic view of Google's electricity purchases and renewable generation at any given time; rather, they are representative of our renewable energy activity from the previous year.

The GHGP Scope 2 guidance limits zero-emission claims from purchased clean energy to those instances where the clean energy is generated and then used in the same geographic region and within a reasonable timespan.

Note that the emission factors differ between location-based and market-based emissions data.

• Hourly greenhouse gas emission factors are used to calculate location-based emissions. That's because the electricity generators supplying electricity to the grid are constantly changing; an hourly greenhouse gas emission factor takes into account the mix of generation sources in use hour-by-hour. When matched with hourly electricity load data, this calculation method produces an emissions figure that is sensitive to the relationship between electricity demand on the grid and the resources called to supply it. Note that this is then aggregated to daily emissions for apportionment purposes to produce the internal service's location-based electricity carbon footprint per day and location.
• Annual greenhouse gas emission factors are used to calculate market-based emissions. That's because these are the emissions factors that are most often used in other Scopes of corporate greenhouse gas accounting. These emissions factors combined with the impact of Google's carbon-free electricity purchases makes it a suitable data source for including in reported Scope 3 emissions inventories.

Electricity footprint

Electricity footprint allocation to Google’s advertising services

A huge variety of services run within the Google boundary, the first step is to identify the internal services which are connected to advertising and then quantify the emissions from each service.

The emissions per internal service are aggregated to a daily basis and these emissions are present for each location where servers are located.

Electricity footprint allocation to customers

Completing the above steps results in a total carbon footprint for Google’s advertising services, specifically providing emissions for each internal service in each location split on a daily basis. The next step is to allocate the footprint of these services to specific customers, aggregated into meaningful units (in this case, account level data supplied on a monthly granularity basis).

This is performed using activity level data including logs. This ensures that for a specific account, the relevant portion of emissions is only included if work on an advertising internal process for that customer has occurred in a specific data center on that day.

Non-electricity emission sources

Non-electricity emission sources

While emissions from electricity production represent the majority of Google advertising’s carbon emissions, other emission sources do contribute to the total.

Carbon Footprint for Google Ads uses data streams for these non-electricity sources that originate with Google company-wide emissions inventory. As such, emissions from non-electricity sources are calculated and added to the advertising footprint on a less dynamic and less granular basis than emissions from electricity. While we measure electricity use and associated location-based emissions on an hourly basis, emissions from other sources are established on either a monthly or annual basis, and are not available with any geographic specificity. Note that the Google company-wide data for embodied emissions of data center equipment and data center facilities has not been assured.

To allocate company-wide emissions from non-electricity sources to the customer-specific breakdowns in Carbon Footprint reporting, we establish an allocation factor – a ratio of the customer's advertising electricity use to the total Google electricity use – and multiply that factor by the emissions from each source, determined as described here.

• Embodied emissions of data center equipment: This emission source encompasses the activities necessary to extract, refine, and transport materials to equipment manufacturing locations, and the emissions associated with the manufacturing processes. Using lifecycle analysis, Google has established a per-piece embodied emissions footprint for data center equipment. This footprint is then amortized over a 4-year timeframe to create an annual emissions burden for each piece of equipment.
  The total number of machines resident in Google data centers and the summed emissions of all equipment is updated monthly by adding the new machines and dropping those at the 4-year mark.
• Embodied emissions of data center facilities: This emission source encompasses the activities necessary to extract, refine, and transport materials to data center construction locations, and the emissions associated with the construction itself - including site infrastructure such as coolant systems and power systems. Using lifecycle analysis, Google has established a data center construction emissions footprint, which is then scaled up or down based on the size (data capacity) of new data center additions. This scaled footprint is then amortized over a 20-year timeframe (chosen to match our financial accounting standards).
  On a monthly basis, Google adds newly available building capacity to its running calculation of embodied facility emissions.
• Fossil fuels combusted onsite: This emission source encompasses all data center onsite fuels use, for example for backup power, water and space heating, and transportation (fleet vehicles). On an annual basis, Google collects all relevant records and sums its total data center fuels use and calculates the resulting carbon footprint as part of its annual emission reporting process.
  The total data center fuel emissions number is updated annually for Carbon Footprint for Google Ads calculations.
• Data center employee commuting and business travel: This emission source encompasses travel and commuting associated with employees who work at Google data centers. On an annual basis, Google collects travel records and estimates of employee commute modes, in each case creating a worldwide total emissions footprint for the activity. This worldwide Google total is then downscaled to data center employees by multiplying the proportion of data center employees vs. Google's total employee count, to create the data center emissions total.
  The total data center commute and travel emissions number is updated annually for Carbon Footprint for Google Ads calculations.

Electricity use

This section describes Google's method for bottoms-up energy consumption calculations.

First, every machine runs workloads for one or more internal services. Google records the internal services using each machine, on an hourly basis. Similarly, Google also records power use by machine, on an hourly basis.

A machine's power use will be a mix of power used to execute workloads (dynamic power), and power used when the machine is idle (idle power). There are two different methods to allocate these machine-level power uses to the internal service level:

• Each machine's hourly dynamic power is allocated to the internal services it supports during that hour. When a workload is running, the major resource contributor to energy consumption is CPU usage. Google monitors CPU usage inside its data centers per machine and internal service workload. If one internal service is using the machine, it allocates the machine's dynamic energy consumption to that internal service. If a machine supported more than one internal service, Google allocates the dynamic power proportional to the CPU usage of each internal service running on the machine.
• Idle energy consumption is allocated to Google internal services based on each internal service's resource allocation in the data center. An important driver of machines at idle is the desire to have compute resources (CPU, RAM, HDD, SDD) "at the ready" to execute uncertain but potentially large workloads without delay or interruption. Idle power is distributed based on the level of compute resources that has been purchased, whether or not the internal service is using those resources. This allocation results in idle power allocations per internal service, for each data center location.

Data center electric overhead load (power systems, cooling, lights) is then allocated to every machine within the data center. Google measures this load at the building level and estimates it more finely at the sub-building level, using validated algorithms as part of Google's Power Usage Effectiveness monitoring system. The sub-building estimates are allocated across the sub-building sector's deployed machines in the same proportions as the completed dynamic and idle power allocations.

Next, the power required by the shared infrastructure services software layer is allocated based on the usage of those infrastructure services by higher-level internal services. Overhead load for shared infrastructure services is included in their allocations. These allocations remain at the internal service (not machine) level.

For the internal services that don't have sufficient usage data, Google uses back-charged costs between the internal services to reallocate the shared infrastructure's energy consumption.

Greenhouse gas emissions

This section describes Electricity Maps' calculation.

Grid carbon emission factors begin with electricity generation data from balancing authorities. This data provides the intraday energy mix, which is the relative production of electricity by the different electricity plants available on the grid. Electricity Maps then adds the real-time electricity import and export between interconnected grids.

Finally, Electricity Maps uses the Intergovernmental Panel on Climate Change (IPCC) (2014) electricity generation emission factors for each electricity generation source (e.g. coal, natural gas, hydropower, etc.) to create a volume-weighted hourly carbon intensity factor (emissions per megawatt-hour generated) for each electric grid. You can review Electricity Maps' carbon intensity factors here.

Note that Electricity Maps does not provide data for all Google server locations, with particular gaps in Asia. Where such data is unavailable, Google uses country-specific annual average carbon intensity factors published by the International Energy Agency.

Google maps the relevant carbon emission intensity factors to each of its server locations. We then multiply the hourly energy use for each internal service at each location, by the appropriate carbon emission intensity factor for that location to determine the internal service's location-based carbon footprint per hour and location. Each internal service's footprint is summed every 24 hours to create a daily footprint for that internal service in each location.

Allocation to customers

Given that at this stage, the emissions for each internal advertising focused service is known for each location for each day, the next step is to allocate the footprint to individual customer accounts. For each customer, using activity based data such as logs and other data points, it can be established which internal processes within each location performed work on their account and also the proportion of work for that day which was performed by the services. Once the Account ID is provided along with date parameters the following occurs:

1. For each day within the provided date range, for each location and each internal service used, emissions for the Account ID are calculated.
2. This data is then aggregated together to get total emissions for that day across all internal services and locations used to service that Account ID.
3. The data is then added up across the date range to provide the total emissions for that Account ID, broken down into Scope 1, Scope 2 (location based), Scope 2 (market based) and Scope 3.
4. This is then presented to customers broken down on a per month basis.

Note that there is validation performed to ensure the aggregation of all the customer electricity carbon footprints is equal to the total location-based electricity carbon footprint for Google’s advertising efforts.