Thursday, 15 March 2012

Google (and Bing) Love Anchor Text Link Spam

While many bloggers and the media are calling Google's search results out lately, most of Photo Credit: Elsie Esq.the focus has been on the somewhat low-quality pages that show up for informational long-tail searches. My concern for Google's search results is different, however. As I touched upon in the last newsletter, I'm tired of Google (and Bing) returning sites that use anchor text link spam to get on the first page of results.

For those who don't know what anchor text is, here's a quick explanation:

Anchor text is the words in the clickable part of any link. For instance, when someone links to my site, they typically use either my name or my company name in the anchor text, which looks like this:

Jill Whalen is an SEO consultant.


Visit High Rankings for SEO consulting.

But those links are nojavascript:t all that valuable in SEO terms. What would be better (for search engines) is for the links to be more like these:

Jill Whalen is an SEO consultant.

Visit High Rankings for SEO consulting.

...with the keywords that people might type into Google (or Bing) as the clickable anchor text link.

Search engines assign a lot of weight to the words that are in that clickable link. It does make sense because you're telling both people and search engines what they'll get when they click the link.

The problem is that it's not a natural way for people to link unless they know a bit about how search engines work. It's more natural to link using the company name, even through links that just say "click here" or "more information."

Part of what I do as an SEO consultant is to train clients to think like a search engine. I teach clients to link more descriptively on their own sites via "internal" links as well as linking to other "external" sites. But to get honest-to-goodness natural links – that is, links from others just because they really like you or your company – it's unlikely that the link will have the best anchor text for search engines. And yet, natural links are exactly what Google claims to value. It's what their PageRank algorithm was originally based on.

But today, natural links and true citations are nearly useless in helping search engines show the best sites for the search query at hand.

For instance, this past Sunday I was quoted in a Washington Post story aboutGoogle's less-than-stellar search results.

I spoke with the reporter for quite some time and also emailed him numerous examples of how some companies easily manipulate Google. He was kind enough to mention me and my company (on page 2) in the article, which was great -- but there was no link. I don't know if it is the Washington Post's policy not to link, or if they just don't think about it, or if they have never been taught to link. It seems to me that a mention in the Washington Post in this context provides me and my company some credibility, because the WP is a mainstream news outlet. Yet any credibility I may have gained with the people reading the article is completely lost on Google because there's not only no descriptive anchor text link, there's no link at all!

Instead, the links that Google (and Bing) end up valuing the most are those where people control the anchor text. Unfortunately, when anchor text can be controlled, it often means that the link:

  • was purchased
  • was traded for
  • was added to a site that is controlled by the owner
  • was added to a site in a network controlled by multiple stakeholders (like a link farm)
  • was added into an article or bio of an article and posted to an article directory or similar
  • was used in a blog comment
  • was part of a forum signature
  • was obtained naturally, but the linker had to be asked to edit the anchor text.
Only the last one of those is truly in line with what search engines want to value, and even that one is sketchy because it becomes somewhat unnatural by virtue of requesting different anchor text.

May the Biggest Spammer Win

Most reasonable people would agree that it doesn't make sense that the companies who own or take part in a network of interlinked websites should rank higher than those who don't. And why should the websites that have people "writing" boatloads of blog comments outrank their competitors who have no desire to spam others' blogs? If you're commenting because you'll possibly get some link value rather than because you feel the need to add to the conversation, it adds unnecessary clutter and should be considered spam by the search engines (in my opinion). I wouldn't be surprised if 90% of blog and forum comments fall into that category, as do most articles submitted to article directories.

Here's the Rub

All of the above types of links still count very highly in Google (and Bing). While links and their anchor text are by no means the only ranking factor for how sites show up in the search engines, they are a very large one at the moment. And surprisingly, neither the relevance nor the quality of those links appears to play as big a role as search engines would like you to believe.

You can take any product search query (both highly competitive and somewhat competitive) and review the backlinks of the sites that show up in Google's (or Bing's) top 10 to 20 results and see what I'm talking about.

Random Example

Let's look at the search query "baby furniture," which I just randomly thought of as I was writing this. I'm not going to call out any of the sites by name, and your results may differ slightly from mine, but you should get the picture.

The first site to show up in Google is a big brand, which makes sense. In fact, I wasn't even going to check the backlinks because I figured they likely deserve to be there based on their brand. But then I noticed it's just a random catalog page from their site. So I looked at their backlinks, and sure enough, there are 357 links pointing to that one page, most from completely irrelevant sites. Some are even hacked sites and porn sites. But they've got keyword-rich anchors that Google (and Bing) love. Many of the links are in blog comments and others in "partner site" areas (paid links). Wonderful.

Let's check the #2 site that shows up in Google (which for me was the #1 site in Bing). Looks like there's a whole network of interlinked baby-related sites that use keyword-rich anchor text links to get all the various sites to rank well for those words. While it's possible that all those sites have different owners and they really just want to recommend (using juicy anchor text) all those other baby product sites, it sure smells fishy to me! At least in this case they seem to be on relevant sites, unlike the big-brand one above.

The #3 site that I see on Google is also a big brand, and it looks as if many of their links are purchased from mommy blogger-type sites, based on my random clickthroughs of their backlinks. While they at least seem relevant, most are anything but natural. I would consider them akin to ads, aka paid links. Nothing wrong with them purchasing ads on relevant sites, but it's Google's job not to count paid links, and yet they do.

The #4 site is owned by a big brand, but is separate from the brand's main website. It looks as if this one may not be spamming...yay! They seem to do well based on links from their parent company site and actual recommendations from other sites. I am basing that assertion on the fact that the links are mainly the website name, not a keyword phrase. So Google may have gotten that one right! (That site doesn't show up in Bing's top 10 for me, however.)

Spammers Rule - Google (and Bing) Drool

I think I'll end my backlink checking here because it upsets me to see how easy it is for link spammers to get pages ranked for highly competitive phrases. Remember, this was just ONE random phrase I checked. I have no clients in the baby furniture space or anything like that. You can type in any type of product search for yourself and see similar results. Seriously, I'd be shocked if you could find a Page 1 general product SERP in Google (or Bing) where most of the sites WEREN'T link spamming their way there.

The conundrum for Google (and Bing), as I mentioned in my Google Sucks article, is not that the pages or websites that show up in the results for these searches are necessarily bad or irrelevant. Those top 4 results for baby furniture all seem like good choices at first glance -- which is what makes it so sickening that someone felt the need to link spam on their behalf. On the other hand, can we fault them for using techniques that work?

My Question Is, Why?

Why would Google (and Bing) allow companies who spam to show up anywhere in their search results, never mind in the top slots? I can spot the spam quickly and easily in just a few minutes; surely with Google's fancy tools they could do the same. Are there really not enough sites that don't link spam that are worthy?

I'm not saying that the search engines should penalize the sites in question. That would leave things open to rogue competitors who might spam on their behalf. But why doesn't Google stop counting the spammy links? And why not stop counting anchor text so heavily, since it's nearly always contrived? I have a feeling that Matt Cutts from Google may tell me that they aren't counting those links already, but I just don't believe it. In most of the results I looked at, there were not enough other factors to explain the Page 1 rankings.

I'll leave you with one additional thought...

If Google doesn't want to (or doesn't know how to) not count spammy links, perhaps this is an opportunity for Bing to set itself apart and become the better search engine. I can't help but think that completely discounting unnatural links and anchor text could only be a good thing for the search engine that eventually implements it.

1 comment: