Reimagining election monitoring
The challenge
Meta needed to replace a widely-used election transparency tool ahead of a major election while maintaining researcher trust and supporting diverse monitoring needs.
Meta previously provided election data through a tool called CrowdTangle. With the upcoming deprecation of that product, we needed to find a new way to display this information. Time pressure was high due to the 2024 European Union parliamentary elections, which were coming up within a matter of weeks. The solution also needed to be scalable for other elections across the world.
Previous product
While researchers had built workflows around CrowdTangle for years, the dashboard designs in this product presented pain points that made usage difficult, even for experienced users.
Researcher pain points
Researchers found it difficult to understand which keywords and candidates are included in CrowdTangle’s election dashboards and wanted clarity on what was included. The lack of transparency was frequently mentioned as a source of frustration among researchers.
Researchers were often confused about which feed units contain content from the same candidates and/or keywords. This made comparison between data categories difficult as it was unclear if the same candidates and/or keywords were included in the Facebook and Instagram posts feeds.
Researchers wanted to be able to unselect some keywords and candidates from the dashboard. The election dashboards weren’t customizable, so researchers had to search through data unrelated to their areas of interest.
Design goals
Build trust through transparency: Provide clarity on which keywords and candidates are included in dashboard displays.
Make complex data understandable: Organize content in logical groupings with a clear hierarchy to show trends and reduce user overwhelm.
Give researchers more control: Balance power and simplicity by allowing users the flexibility to customize their dashboards by selecting which data is included.
Create a scalable solution: Develop a reusable dashboard framework that is able to support future elections, new regions, and additional datasets.
Political parties and organizations dashboard
Building trust through transparency
The challenge wasn't simply building new dashboards. The challenge was helping users trust a new system.
Researchers needed confidence that they could understand what data was included, understand what wasn't included, customize views, monitor trends, and conduct repeatable analyses.
We began by creating a landing page with an overview of the available European Union parliamentary election dashboards. These included country-by-country breakdowns as well as a dashboard showing all political parties and organizations to help researchers track overall trends across the EU.
European Union election dashboard landing page
Keyword dashboard example
Making complex data understandable
Candidate dashboard example
Researchers weren't just consuming content. They were interpreting large volumes of information. The dashboard structure needed to support analysis rather than merely display data.
To accomplish this we first looked at how researchers were grouping content in order to analyze it. Based on researcher behavior, we set up candidate and keyword dashboards for each country. This structure provided easy access to the data researchers were most interested in examining. Breaking the data into multiple dashboards also served to reduce overwhelm while still providing a great deal of information.
We also provided trend charts for each dashboard, broken down by platform. This allowed researchers to quickly compare the candidates and keywords being posted about most frequently on Facebook and Instagram.
Starting with the high-level data in the trend charts and then listing all of the related posts helped researchers to get an overview before diving into the details.
Keyword selection
Giving researchers more control
Candidate selection
Rather than treating dashboard configuration as a hidden system detail, we surfaced the scope of each dashboard directly within the experience. This helped researchers understand what they were analyzing and increased confidence in the results.
We listed the number of candidates or keywords included in each dashboard. Researchers then had the option to edit the included data depending on what was relevant to them.
By providing lists of relevant candidates and keywords for each election, we made the dashboards simple to start using. By allowing customization, we helped researchers tailor the dashboard to specific investigations rather than forcing a fixed view.
Creating a scalable solution
Instead of just designing a one-off experience for the European Union parliamentary elections, we created a dashboard framework that could scale across elections, regions, and future datasets. While the specific details will vary, the model of setting up a basic dashboard structure with options to adjust based on each researcher’s needs can easily translate to other use cases.
We later built on this concept to add an option for researchers to create fully customized dashboards, which displayed post feeds and trend charts based on selected keywords or content producers.
Results
We shipped an MVP for dashboards in 10 days, providing researchers with access to key data in the weeks leading up to the European Union parliamentary elections.
The launch of dashboards in Meta Content Library led to a quadrupling of researcher engagement with the product.