Signal Strength - Linking Speed Bumps

Posted by Eric Ward on August 19, 2010 to

One of the many search algorithm signals you'll hear discussed is link speed. By link speed we are referring to the speed with which your site attracts or obtains links. One widely held belief is if you get too many links too quickly it looks unnatural, and this sends a bad signal to the bots. For the sake of this column let's agree that signals do exist, and link speed is one them.

The next logical questions would be how many is too many and how fast is too fast? Three years ago this month I wrote a column for SearchEngineLand titled Aggressively Seeking Links: How Much Is Too Much?

Here we are a thousand days later, and you still find confusion among the SEO/SEM community as to what constitutes too many or too fast.

Here's my reasoning as to why there cannot be a concrete answer to this link speed question, and an example.

First, there are plenty of perfectly legitimate reasons a web site could attract a surge of a few hundred or even a few thousand new links, literally overnight or within a few days (more on that in a moment). Second, Twitter and blogs make link propagation fast and easy, both real and spammy. Third, for both a brand new web site or an old, link spikes can be totally legitimate, or completely manipulated. (note: some link builders make a distinction between a brand new domain with zero inbound links, versus an old site with an existing historical "link growth profile" that if it suddenly spiked, would raise a flag). But this distinction isn't so cut and dried. Just because a site is old doesn't mean it can't have a natural link spike. If I'm right, the bots have to be able to draw a distinction between the speed at which links are being attracted and the quality and intent of the sites providing those links. All links given are not equal, especially those that come like a hailstorm.

Examples? Four months ago, you'd never heard of the Deepwater Horizon oil rig, and why would you have? There are thousands of oil rigs around the world. They don't have web sites. But when the Deepwater Horizon rig blew up, you heard about it, and web sites were launched about it, both brand new like http://www.deepwaterhorizonresponse.com, and also brand new sections at existing (and very old) sites, like http://www.bp.com/claims. In fact, the BP claims section has over 20,000 links, and that URL now forwards to another URL which is also attracting links.

Neither of these URLs existed three months ago, so nobody could link to them.

I wouldn't be surprised if BP's various "claims" URLs have more total inbound links than any other interior content URL on the BP.com site, all of which have come within the past few months. Unnatural? Nope, not at all. Very real, very sad, and very natural. And I'm sure as soon as BP can, it will kill off those pages, rendering those thousands of links dead. It's ironic that the kind of news that can cause a natural link spike engenders the kind of links you might wish you didn't have.

Back on point, a new site as well as an old site with new content can easily and suddenly generate mass links, perfectly naturally.

This doesn't mean you can go buy or build links with reckless abandon, though. The key is the word “naturally.” As I have written before, every web site has a certain linking potential, yet some linking related occurrences are out of our control, like a site devoted to an oil spill finding itself in the linking spotlight due to a horrible accident. Any engine looking to penalize a site just because it has a huge number of links, or a sudden surge in new links, would have to have the ability to algorithmically recognize when those surges were natural, or manipulated. My belief is that isn’t as hard as it sounds. It’s one thing to attract links from a few thousand news, oil industry and environmental sites. It’s quite another to have thousands of links from sites that have absolutely nothing to do with each other.

I can tell you that spotting manipulated linking patterns isn’t as hard as some people think it is. Link Insight helps me see them almost immediately, and can help you do the same. This also must mean if the engines want to look for a suspicious linking pattern, they can find it. Not every time, but you’d be amazed what you notice when you look across a list with 25,000 URLs/links in it. Without even having to look at the sites themselves, I can often spot manipulated links. From the URLs alone you can spot link spam.

So the real answer to the question is that it’s not about how many or how fast. It’s all about how natural.

If it can be algorithmically trusted as being natural, there is no such thing as too many or too fast.