In today’s post we take a look at how the Bing ranking algorithm considers content quality for ranking web documents. There are often many web documents on the web that are relevant for any given user query and an effective ranking algorithm should consider not only relevance but also the authority, utility and presentation of the content in order to provide the best search results. My colleague Michael Basilyan will provide some insights on how Bing algorithms incorporate content quality in order to improve the ranking of our results.
Dr. Jan Pedersen, Chief Scientist, Bing and Information Platform R&D
Every day millions of people come to Bing to find information that matters to them. Consider the query breast cancer symptoms. In earlier generations of search, this query would be simply treated as three keywords to match against the web. Behind that query, however, is a real user with a real information need. The returned results can have life-altering ramifications. Much of the content matching this query on the web is of low quality: written hastily by non-medical experts and incomplete. Queries like these motivate us to go beyond keyword matching. We want to help our users find content that is authoritative, useful, well written and well presented.
In today’s post we’re going to take a look at how the Bing ranking algorithm considers Content Quality. We will shed some light on recent updates to the algorithm and Bing’s perspectives on Content Quality: how we define it, how we measure it, and how it impacts the ranking of our results.
Three Pillars of Content Quality
When we discuss Content Quality, we are referring to three aspects of a website or page that we call the three pillars of Content Quality:
When we evaluate the page with regard to authority, we’re trying to answer the question, “Can we trust this content, its author, or the website?” In addition to understanding the web anchor graph, a variety of factors are used to establish and determine the authority of a page. These include signals from social networks, cited sources, name recognition and the author’s identity.
Authority is treated differently in different query segments. For example, for health topics, our algorithms will prefer documents written by professionals from well-known sources.
When considering the utility of the page, our models try to predict whether the content is sufficiently useful for the topic it is trying to address. Does the page provide ample supporting information? Is it at the appropriate level of depth for the intended audience? We prefer pages with relevant supporting multimedia content: instructional videos, images, graphs, etc.
Another important criterion in evaluating utility is gauging the effort and level of expertise required to generate the content. Websites serving unique content are preferred to those recycling existing data or widely available materials. A great example of this are real estate listing sites. These sites generally syndicate information available elsewhere (via MLS or government sources). However, even these kinds of sites can move up in the ranking results if they set themselves apart with unique value that others in that category may not have, such as school information or nearby transportation options.
A well-presented page will have an easy-to-read, accessible design, and will make its primary content easy to find. In contrast, poorly presented websites require the user to wade through introductory or unrelated material to access meaningful content.
Appropriate usage and presentation of advertising is an important aspect of Content Quality. Bing will promote and support websites and webmasters that provide ads relevant to the content of their website and place ads so that they do not interfere with the user experience. Pages with well-designed layouts will be preferred to pages that hide content behind ads, fail to clearly delineate ads from the main content, or feature ads that are easily confused with navigational elements.
Digging into some examples
The three pillars of Content Quality are central to the user experience and to finding information on the web. In fact, they are designed to support the well-being of the web and the interests of our users and webmasters.
Let’s consider the following page as an example where we see some Content Quality issues:
From an authority perspective, this article is problematic because it is written by an unidentified author (“admin”). Navigating around the website yields little information about the source of the content or its authorship. The page also fares poorly from a presentation perspective: four or more ads precede the main content and take up a lot of area above the fold. Furthermore, the ads are not always clearly delineated from the main content and are easily confused with navigational elements or hyperlinks.
For an alternate example, consider this page:
Navigating around the website, yields little information about the source of the content or its authorship. The page also ranks unfavorably in terms of presentation, containing four or more advertising sections that precede the main content and take up a lot of area above the fold. Furthermore, the ads are not delineated from the main content and are easily confused with navigational elements. In terms of utility, this page seems to be poorly written and has no supporting content or multimedia. Overall, this page has relatively poor Content Quality. Our ranking algorithms are designed to prefer better designed and more authoritative content.
Content Quality is a Primary Factor in Ranking
In Bing, the relevance of a result is a function of:
- Topical relevance to the query (“Does it address the query?”)
- Content Quality (as measured by the three pillars described above), and
- Context (“Is the query about a recent topic?”, “What’s the user’s physical location?” etc…)
Much of our work is spent in improving our ranking engine’s performance along these three axes while determining the optimal tradeoffs. When we prepare a change to the ranking algorithm, we always systematically evaluate its performance against these axes and validate it with real users on Bing.com to ensure increased user engagement and satisfaction.
Content Quality is one of the primary dimensions along which we optimize our ranking models. This is good news for web masters because the three pillars of Content Quality are largely in their control. And the great news for users? Content Quality means we are serving more authoritative, useful, and better designed content.
On behalf of the Bing Content Quality Team —
Michael Basilyan, Senior Program Manager