This is due to the way that we identify content on your pages. We identify content using a machine learning model which disregards template elements such as headers and footers. For more information on this, see "Readability: How does Siteimprove identify content on my page?".
If our content detection algorithm struggles to identify content, for example due to HTML tagging of the page, then some content can go unevaluated.
Generally speaking, pages with little content result in a higher readability score (i.e. a higher level of difficulty) if they contain either long or polysyllabic words.
Pages should always be reviewed manually to ensure that they live up to your organization's standards. For this purpose, we suggest you can start by reviewing the pages with the highest Readability scores.