If Google starts sentiment analysis of backlinks, SEOs will need this…
As many SEOs now know, Google is planning to unleash a series of updates to its Penguin algorithm-update-thingy.
While Matt Cutts was about as transparent as he could be, without tipping Google’s hand to spammers,[highlight color=”yellow”] Greg Boser told a packed room at Interactivity Digital that he believes sentiment analysis will come in to play when looking at backlinks.[/highlight]
How will Google use sentiment analysis in its next update?
[pullquote]…Google will start using some kind of automated sentiment analysis to determine whether a backlink is not just authoritative, but positive or negative in sentiment.[/pullquote]Well, Boser speculates that Google will start using some kind of automated sentiment analysis to determine whether a backlink is not just authoritative, but positive or negative in sentiment. Why? Well, there are many industries that receive plenty of back links to their web sites, but they are usually from articles or posts that attack their practices or industry. Using sentiment analysis, Google will be able to add extra weight to those links that are neutral or positive, and devalue the negative links.
Of course, we don’t know if Google will start implementing automated sentiment analysis on backlinks–heck, it may already have started–but as an SEO, it would be a good time to start shifting your focus to making sure you acquire links that are positive in sentiment towards your page and desired anchor text.
That’s where Trackur comes in. 😉
All of Trackur’s paid plans include automated sentiment analysis as standard. What sets us apart is exactly how we calculate our analysis of a page’s sentiment. We don’t just look at the overall sentiment of the page–a page could be positive towards one keyword and negative towards another. Instead, we take the keyword being monitored and automatically calculate the sentiment of that keyword on the web page in question.
What does this mean for SEO?
Well, in addition to finding any mentions of your company or product online, you can use Trackur’s automated sentiment analysis for improved back link prospecting. Here’s how:
1. Enter your desired anchor text as your “keyword” in Trackur.
2. Select only to view news, blogs, and perhaps forums.
3. Trackur will return matches to all pages that include your desired anchor text. You can then select to view only those that are positive or neutral.
4. Now you have a list of those bloggers/journalists that are already talking about your keyword and have nothing but positive things to say. You can reach out to them and ask for a link in their existing post, or persuade them to write about your company.
5. You can export the data to CSV so that you can slice and dice it in Excel. For example, you could view only those pages that have a positive sentiment and then sort by those that have the most influence! A double rainbow win!
6. You could even go back to step 2 and select Twitter only so you can find those tweeting about your industry with a positive tone! Or even target links/tweets from a specific country!
[pullquote]Provide us with a URL and your desired anchor text and we’ll analyze that page for sentiment towards that keyword![/pullquote]If you have any kind of developer skills–or already have your own SEO toolset–then you can take this to the extreme with Trackur’s automated sentiment analysis API. You can provide us with a URL and your desired anchor text and we’ll analyze that page for sentiment towards that keyword! A great way to help you–or your users–find quality, influential backlinks…with a positive sentiment!
Whether Google does indeed start implementing automated sentiment analysis in its algorithm, we don’t know. Either way, you’ll be a step ahead of your competition with fresh, positive, influential backlinks! 😉
I have an odd question. Working in web hosting I have many clients whose content is NSFW and it is normally considered “bad” if they link to us. But what if the sentiment/text is “Hey this host is great, we’ve used them 10 years” Does the sentiment nullify the reputation of the source?
Well, that would be a question for Google. 😉 However, I would suspect that if we’re smart enough to look at the sentiment on a sentence level, then Google is too. 😉
Vikki, that’s an interesting question. To answer it fairly, turn it around a little and ask “If someone has a terrible reputation, especially in being deceptive or manipulative, do I want to trust their sentiment anyway?”
In other words, where a site has a reputation problem based on manipulative and deceptive stuff, Google will never give any trust or credence into the sentiment they claim to express. They can’t trust it is any more genuine than their links.
If anything, it seems to me that Google gives more credit to negative content than positive content. Google ranks positive and negative content all the same. Look at how many brands have to fight against the likes of ripoffreport.com and pissedconsumer.com for ranking on their own branded search terms. Another reason I believe Greg’s prediction regarding sentiment of links will never be given much clout in the Google Algo: look at politicians and news. People want to know the good, bad and ugly.
I don’t think we’re talking about the same thing though. Google may indeed rank negative content, but it may decide that a scummy site shouldn’t benefit from a link from a valuable source, if that source is criticizing or attacking that site.
If the “scummy” site gets a link from the “valuable source”, it’s been influential in some way. I.e. It was relevant enough to discuss. Including the link means the site editors intend for their visitors to see it, a critical endorsement. I guess there are differences in (algorithmic) interpretation of popularity, relevance, and influence. TBD.
I think you are on to something Andy… they adjust Local rankings, in part, using sentiment analysis, so it makes sense they would scale that skill-set in some way.
Thanks for the input Andy!
The problem is, how do they know when a comment is positive and negative? They can’t just assume that all positive words mean the review is positive can they. For example if someone wrote “Just had a really good meal at X…NOT!!” Google may see that as positive when it is clearly not. I agree that this is going to happen but it is going to be really interesting/frustrating when it is released because I can imagine them getting it horribly wrong many times!
Amazing post..
I think there is still much to be discussed. The biggest problem is how
this will affect the good sites, even Google to find a common
denominator.