Creating greener websites with Wagtail
Team members
Jaden Bruno
Noah Levine
Edward Quezada
Peter Ming
Alex Solodko
Faculty mentor
Markus Pelger
Sponsor organization
Wagtail is an open source content management system that allows users to build, edit, and structure websites.
Project description
Wagtail came to us because they seek to compete against major competitors, such as WordPress, by establishing and marketing a more sustainable platform. Specifically, they want us to determine and quantify the main drivers of carbon emissions for websites and then extrapolate the potential economic and environmental effects upon implementation of the Wagtail system. Based on these findings, we will make suggestions on what Wagtail's next steps should be.
Techniques and methods used
We utilized a multitude of techniques during our project journey. Through our research, we aimed to determine which features of a website have the largest impact on predicting emissions per page load. To do this, we first used CMS Detector to determine if a website is using WordPress as its primary CMS. We then aimed to build a representative sample of the internet, by researching 100 websites spanning 21 major industries.
We utilized Ecograder, a tool that uses web crawling technology to evaluate the environmental sustainability of a website. Ecograder assigns an overall sustainability score out of 100 and reports the emissions per page load. Ecograder also provides a score for the three major functions of website emissions: Page Weight, UX Design, and Green Hosting. These scores are further broken down into 20 subcategories, which also have a score out of 100.
To predict emissions for websites and to test which variables are statistically significant in predicting Ecograder scores, we ran a multiple regression analysis. The variables we used to predict the Ecograder score were the Page Weight score, Greenhosting score, and UX design score.
In order to assess model fit and performance, we looked at both the mean squared error (MSE) and the R² values. The mean squared error indicates the average squared difference between the observed actual outcomes and the outcomes predicted by the model. R² is a statistic that measures the proportion of the variance in the dependent variable that can be predicted from the independent variables. An R² of 0.797 indicates that approximately 79.7% of the variance in the Ecograder score can be explained by the model, indicating a strong fit.
Then, we ran univariate regressions on each of the features to see which ones had the largest R² relative to the Ecograder Score. We further examined the coefficients to better understand the correlations between the Ecograder score and each feature.
Solutions and deliverables
Results
Quantifying Improvement in Ecograder Scores
Using our prior regression model and pertinent descriptive statistics, we quantified the potential improvements in Ecograder score if green hosting is implemented and scores for the most significant features of UX Design and Page Weight are optimized.
Next, these values were inserted into our original regression function for predicting the Ecograder score. Our model indicates that if the average WordPress website implements green hosting and maximizes scores for the four most significant features between UX Design and Page Weight, the Ecograder score of the site will rise on average by 29.25 points.
Quantifying Reduction in Emissions
The potential improvement in Ecograder scores across average WordPress websites can be extrapolated to quantify the total reduction in emissions across all of their sites on the web.
This value was then multiplied by the number of WordPress sites currently in operation to extrapolate it on the global scale. We found that 368.4 million tons of CO2 can be reduced by maximizing the scores in the statistically significant categories of Page Weight, UX Design, and Green Hosting. This figure accounts for a 27.1% decrease in total internet emissions and a 1.0% decrease in global emissions.
Societal Cost of Emissions
The societal cost of emissions is defined as the estimated cost of damage as a result of each additional ton of carbon released into the atmosphere. The Biden Administration currently estimates this social cost to be $51 per additional ton of CO2. Using this value and our estimated decrease in total emissions, we found that implementing our proposed changes results in $18.8 billion of societal benefit. This figure indicates that there are major financial incentives for pursuing more efficient web practices.
Recommendations
Moving forward, there are three steps that Wagtail should perform to capitalize on our research. First, a detailed cost analysis for the implementation of green hosting and improvement of features such as UX design and page weight should be performed. This will provide a clearer image of the feasibility of implementing these proposed changes across all Wagtail sites.
Second, further research should be conducted on potential advancements in the computing sphere. New developments in computer processing abilities or new storage technologies can exponentially improve computing and must be considered in future analysis.
Third, there must be pressure applied at a government level to introduce subsidies that incentivize corporations to transition to green computing. These new policies can ease the initial cost burden that continues to be a barrier for several groups.
To gain a competitive advantage in the industry, Wagtail should carefully consider implementing the changes outlined above. By making improvements in the five significant subfeatures we identified in our regression analysis, including properly sized images, deleting unused code, using efficient cache policy, page rendering, and renewable servers. Wagtail can gain its desired competitive advantage over WordPress.
Along with this competitive advantage, improving these subfeatures will hopefully have a ripple effect on the CMS market and lead to widespread adoption of these practices across the industry. As accessibility to the web continues to expand, it becomes increasingly important for sustainable practices to become common among content management systems. Ultimately, environmental concerns in this sphere will continue to expand in the near future and must be addressed as soon as possible.