Are We Getting Mixed SEO Ranking Signals?
Speculation time! In this week’s post I review scenarios where I believe that Google’s alleged ranking signals might not be what we accept or even conform to guidelines published in Webmaster Central help files. Here are some of the factors that I find ambiguous or unclear, along with some thoughts on why conventional wisdom might […]
Speculation time! In this week’s post I review scenarios where I believe that Google’s alleged ranking signals might not be what we accept or even conform to guidelines published in Webmaster Central help files. Here are some of the factors that I find ambiguous or unclear, along with some thoughts on why conventional wisdom might be wrong.
All indexed pages pass PageRank—or do they? Consider an e-commerce category page that has links to 100 products on the same site. The e-commerce category page has a PageRank of 2, or lower. The page has no text to speak of, except for a title tag. It has no external links pointing directly to it. Consider the possibility that Google will not pass PageRank (or the Bing equivalent thereof) because of the perceived low quality of the category page. From a strict technical perspective, this page should be passing PageRank to the product pages, but if it is a low quality page, why bother? If the page suddenly spouts some genuinely useful text, perhaps the search engine will look at the page differently.
NoFollow links are not used as a signal—or are they? The definition of NoFollow suggests that a link that uses this attribute would not pass PageRank to the page that it links to, or that Google isn’t paying any attention to links that are NoFollowed (in either a positive or negative sense). However, this interpretation would ignore some powerful signals in those links, and I believe there may be more nuanced interpretations of how Google interprets NoFollowed links.
One simple way to use this is to see who is buying ads. You would need to tune this a bit, because forum and blog comments are not ads, but some basic filtering will get us a long way here, as detecting forums and blogs is generally pretty easy. Note that buying ads is not a bad thing, but it could be data that the search engines want to have and use in some fashion.
You could also potentially use NoFollowed links to see who is doing a lot of anchor text rich commenting in forums and blogs. Could be a spam signal, yes?
Many government web sites have a policy of NoFollowing all of their external links. But government sites don’t link (NoFollowed or not) unless they are in fact saying that a site provides something of value. In other words, a link from a government site by default is a more important than average vote! Think they doesn’t count at all? I don’t believe it.
And then there is the dreaded Wikipedia, with its standard NoFollow policy. Yet getting a profile page for yourself or for your company is not easy. Being featured in Wikipedia is certainly a relevance signal, isn’t it?
Web site references which are not links pass no value—or do they? This one is particularly easy to debunk. We already have confirmation that Google counts web references as positive signals for local search ranking purposes. It is also generally accepted that data consistency across all the references to your site on the web is a critical signal for local search too. So not only does Google notice the references to your site, but they are looking at address and phone number info as well. Clearly, being used as a signal in local, but what about web search?
There are tons of media sites that refer to businesses all the time without linking to them. Yet a lot of media mentions using your brand name (even if not linked) would suggest that your web site deserves some visibility in search engines too. In addition, search engines monitor social media sites for breaking news. Many of these sites have raging discussions taking place on events as they are happening. For example, a high volume of activity on Twitter could be a strong indicator of an important event, which the search engines can use to surface breaking news in their search results.
302 redirects do not pass PageRank—or do they? Largely true, but is it true 100% of the time? The definition of a 302 redirect says that it is “temporary,” but what if a 302 is in place for over one year. Would you still consider it temporary? Or, if you are examining a large, prominent web site that has moved content a number of times, and always uses 302 redirects, but never removes them? If you were Google, you might start to pass some of that link value through, particularly if in parallel some of the sites that linked to the old page update their links to point to the new one.
Search engines follow a limited number of links per page—or don’t they? Also largely true, but not always. Major factors in deciding how many links Google looks at on a given web page are authority and trust of the overall site, authority and trust of a page, perceived value of a page and perceived function of the page (e.g. is it an HTML site map page?). So while you may have heard that you should limit the number of links to 100 on a page, sites that have enough positive indicators from these signals can definitely handle more, and conversely, poorer quality sites can probably handle less.
Ultimately, context drives a lot of these exceptions. What it comes down to is that there are probably hundreds (or more) of these situational types of things that search engines take into consideration. Make sure to question your assumptions when you are making SEO decisions. Learn to analyze whether or not there is potentially a signal available to the search engine that differs from the expected norm, and be aware of that possibility. If you know of other such scenarios please mention them in the comments section below.
Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.