Introduction
Impactly is a calculator that assesses the environmental footprint of an email campaign.
The aim is to directly visualize the most effective levers according to the various parameters of a campaign.
The environments taken into account are the Numberly MarTech Platform (NMP) and Adobe Campaign.
Based on the European standard Product Environmental Footprint (PEF) and using the Life Cycle Assessment (LCA) methodology described in ISO 14040-44, Impactly guarantees its compatibility with existing measurement systems, its comprehensibility by third-party experts, and the reliability of the results. Nevertheless, as digital services rely on shared infrastructures, calculations remain estimates, as they depend on sometimes complex allocation rules.
The sources are taken from the NégaOctet database (a reference on digital) and which the ADEME also uses for its Empreinte database.
Perimeter of calculation
Working time
Time spent on conception, design, integration, targeting and project management can be included in the calculation. The unit is the full-time equivalent day (FTE/day). Hardware typology is also taken into account according to 3 levels: Marketing, Marketing Tech and IT (developer).
Data Centers
We distinguish 2 factors for the Data Center part:
- The power consumption of servers, switches and routers, as well as the consumption linked to cooling.
- The manufacturing footprint of these devices in proportion to the time they are used to deliver the service.
The network
The local loop (the last mile) is considered, again in terms of manufacturing and power consumption, with a distinction made between the mobile (3G/4G/5G) and fixed (fiber/copper - xDSL) networks.
National and transatlantic fibers (backbones) are not taken into account due to their significant mutualization and virtually zero impact in the calculation.
User devices
Once again, the impact of manufacturing and the power consumption of devices (internet box, computer, screen, smartphone or tablet) are taken into account, in proportion to reading time.
Settings
On loading the application, a default parameter setting is applied. Data specific to the campaign to be measured must be entered.
Volumes
First, the number of mailings and the number of opens. It's important to understand that only openers download the images contained in the e-mail. By default, the rate is 30% of opens.
Time spent
For the agency's conception, design, integration, targeting and project management. Time spent by the client is accounted for in its own balance sheet and should not be included here to avoid double-counting. The unit is the full-time equivalent day (FTE/day).
The reference year
The year of routing determines the emission factor linked to electricity production, which varies from year to year and from country to country. For current-year routing, year n-1 must be used for projection.
Email weight
A distinction is made between HTML code, which will be sent to all recipients, and assets (images, webfonts, css stylesheets), which will only be downloaded by openers. The unit is the Mega Byte (MB). We consider a weight of 80 Kb (0.08 Mb) or less for HTML to be a correct value, and 2 Mb or less for all assets (images and webfonts).
Reading time
Reading time affects the calculation of consumption on recipients' devices. It varies according to the nature of the message and reading interest (and therefore the quality of targeting). We observe a reading time of 10 seconds (high estimate) for commercial communications if they correspond to recipients' centers of interest. For newsletters of a more informative nature, the average reading time can rise to 30 seconds.
Destination country
The carbon emission factor of electricity production is different for each country, depending on its means of production (nuclear, gas-fired, coal-fired, hydroelectric, wind, photovoltaic, etc.) If the same campaign has recipients in several countries, they should be added to the selector. If there is only one country, those selected by default should be deleted.
Routing in periods of lower emissions
A lever for reduction may be to route your campaign at the time of day when the kWh of electricity has the lowest emission factor. RTE's eCO2mix site shows that this emission factor can vary from simple to double during the day. The switch is used to simulate routing during a period when the mix is at its lowest (50% of the annual average, i.e. 15 gCO2e/kWh instead of 32.4 gCO2e/kWh). This parameter is only valid for France. It is therefore taken into account if the routing tool is the NMP because the infrastructures are in France (Adobe's AWS data centers are in Ireland for Europe) and if France is one of the destination countries.
BtoB or BtoC campaign
The type of recipient influences the average rate of mobile equipment and the type of network used.
A BtoB target, by nature less mobile, will by default have an equipment rate of 35% smartphone, 30% PC and 35% laptop, 35% mobile network.
For BtoC: 65% smartphone, 25% laptop and 10% PC, 65% mobile network.
These rates can be modified in the "Distribution setup" parameters.
The mobile network, and 5G in particular, is much more energy-consuming.
Activatable levers
The whole point of Impactly is to see the distribution of the footprint according to the specific parameters of a campaign.
Firstly, narrow the target if possible, excluding inactives for example, or adding more restrictive targeting criteria.
For low-volume campaigns, the time spent on design will generally be more present because it's independent of the volume of the base. In this case, the use of templates to reduce the time spent is an interesting avenue.
For low-volume targets, the time spent on design will generally be greater because it's independent of the volume of the base.
For targets in France, the routing time can have a significant impact (generally between 11am and 3pm. A predictive algorithm is currently being developed to refine the recommendation.
For campaigns in France, the time of routing can have a significant impact.
For high-volume campaigns, reducing the weight of HTML and images is a good lever.
Interpretation of results
It's easy to understand now that we can't simply answer the question "is my campaign weight right?"
On the other hand, we can apply best practices for targeting and optimizing code and images, and derive kpi if necessary.
HTML code: up to 80 Kb OK between 80 and 120 Kb Optimizable and above 120 Ko: Too heavy
Assets (img + fonts):up to 1 MB OK between 1 and 3 MB Optimizable and above 3 MB: Too heavy
Some orders of magnitude:
- 13 kgCO2e = 100 km by car (on average)
- 2 tCO2e = 1 round trip Paris-New York by plane
- 10 tCO2e = Average emissions of a French person per year (between 8 and 12 according to studies)
Conclusion
To respect the Paris Agreements (COP21) and stay below 2° of global warming by 2100, the average annual footprint per person will have to be 2 tons of CO2e in 2050, 5 times less than today.