Merit helps you discover and read high-quality content on the internet.
Tired of reading low-quality content on the internet? Merit uses algorithms to analyze the quality of writing on websites. Merit flags the highest-ranked websites in your search results so you can save time and access better information when browsing the web.
Merit allows you to experiment with what you consider high quality content. Below are the factors —like Reading Time and Reading Score—which you can adjust according to your own preference. These factors are indicative of low quality or high quality websites.
- Reading Time is how long it takes to read an article, as measured by its length. We theorized that longer articles mean the author has put more effort into them, and higher effort is linked with quality. In real-world testing, a higher Reading Time was a decent predictor of quality, so we've turned Reading Time on by default.
- Reading Score is how sophisticated the writing on a website is. We theorized that more sophisticated writing is associated with writers that have more expertise and are better thinkers. In real-world testing, a higher Reading Score was a decent predictor of quality, so we've turned Reading Score on by default.
- Unique Links is a measure of how many unique links are in an article. We theorized that more links mean an article has more sources, which indicates a better-researched article. In real-world testing, a higher number of Unique Links was somewhat predictive of quality, so we've turned Unique Links off by default.
- Images is a measure of how many pictures are in an article. We theorized that more images means an article was produced with higher effort, which is linked with quality. In real-world testing, Images were an uncertain predictor of quality, so we've turned Images off by default.
- H1 and H2 count are measures of how many headings are in an article. We theorized that more headings means an article is better organized. In real-world testing, header counts were an uncertain predictor of quality, so we've turned header counts off by default.