Trouble with Google’s Snippets

Recently, there has been some fuss about Google’s page snippets showing inaccurate and misleading information. Google’s page snippets help the searcher quickly find relevant search results and are the parts of a search engine result page directly under the title of the page (in larger, blue text) in black. Snippets contain phrases from the indexed page, and the phrases themselves contain text that matches the keywords typed in Google’s search box.

In one incident, the first of its kind, someone typed in “Zwartepoorte,” a name of an auto dealer in Europe, and “failliet,” the Dutch translation of “bankrupt.” One of the results was a page from the site Miljoenhuizen.nl. “Failliet” was found in one section of the page and “Zwartepoorte” was found in another section. Google usually notifies the searcher that the snippet phrases came from non-contiguous parts of the page by adding ellipses between the phrases. This was the case in this incident (the ellipses between “BMW” and “Dit”), shown below. However, a legal battle ensued because the resulting snippet essentially said that the auto dealer was bankrupt, when that was not the case. The web master of Miljoenhuizen.nl ended up having to pay the consequences.

Ellipses are probably used to conserve space so that Google doesn’t have to put each phrase on its own line. They are also used to signal a break in text. The issue is that the ellipses are hardly noticeable nor distinguishable from other ellipses that may be in the original text. If they were a bit more noticeable, the average user might then investigate their meaning. Google could even add a helpful link next to the “Cached” link, such as “Composite Snippet.” It might look like this:

Another issue I see is that the definition of “non-contiguous” is left up to Google’s algorithm and their engineers. I’m sure there are known ways of how to deal with this potentially slippery issue, such as separating the text by some large distance (as it appears to Google’s algorithms, i.e. ignoring any Javascript and styling) or by putting them in separate HTML elements. These are simply ideas off the top of my head and not necessarily realistic ways to approach the issue. Especially if the content on your pages are dynamically created, it may be very difficult to circumvent some inaccurate snippet text being shown on Google’s result page.

However, there is a way to cause Google to stop showing all snippet phrases from a specific page on your site on their search results. You can find more information about this on Google’s blog post, The Robots Exclusion Protocol. Essentially, you would add the following meta tag on any page for which you want Google to hide the snippets:

Beware that, in addition to completely removing the snippet, this renders the cached version of that page (and therefore the Cached link) inaccessible to searchers.