On-site Optimisation: Page Content, Google Panda and Latent Semantic Indexing

Page Content

Page content was never really an issue in the past and as long as you naturally used the keywords for which you wanted to rank (and also semantically related variations of those words), then that would be all that is required. Now that the Panda update is out in the wild, content has become more important for reasons other than relevance.


Google Panda

There are a lot of things to read about Google panda so I won’t waste your money giving you an explanation when others can do it much better than me. There’s a link at the end to a good webinar from a service provider that I use and it’s definitely worth a listen.

Optimising your website for the Panda update can basically be summarised as ensuring the content doesn’t incur a penalty based on a high correlation to websites that have been scored with a low level of visitor satisfaction.

By talking in extremes, it’s quite simple to use common sense to determine how best to comply with Google Panda. If you can imagine the worst possible site for user-experience, then you can form ideas of best-practises based on the opposite of these.

Examples can be minimising the number of adverts and affiliate links, very short content, content that isn’t related to users’ searches and other things like this.

Whether you are using Google Analytics on your site or not, Google can determine various metrics relating to user engagement such as time on site and bounce rate through visitors to the site that have the Google toolbar installed, visitors using the Chrome browser, and also the speed at which a person leaving a SERP to your website returns again to the same search page (indicating lack of satisfaction with the page on your site).

I’m sure it’s quite obvious that by writing good, engaging content should keep Google happy with regard to these things.

Some interesting data that came out in the webinar about Google Panda was that the average page ranking in the first position in Google has over 2000 words of content. I think this is quite a bit higher than many people expected and should be taken into account when planning a new site.

Even if you can’t think of that much content to write about on the homepage, then perhaps you can include excerpts of a number of recent blog posts or something similar to increase the amount of relevant content. This is how I plant to increase the amount of relevant content on my own pages.

Similarly for pages that are blog posts, it may also be worth considering allowing people to leave comments as this will again increase the amount of relevant content on the page.

If you can also couple this with Facebook comments, it can help increase traffic too.

As a side-note, Facebook comments use JavaScript as standard so they’re obviously not “readable” by Google but there is an API that allows their content to be output to the page for indexing by Google and it can be legitimately hidden to users to that they only see the javascript comments.

This is still allowed by Google as you’re not hiding page content from them as such and are still displaying the same version of a page to the users and to their spiders.

I would expect there to be a WordPress plugin out for this soon if there isn’t already so it might be something to consider when working on clients’ sites.

I’ve still not looked into this a huge amount myself yet as we’ve only recently had our first panda penalty but I would imagine that images/videos would also be a good thing in regard to Google’s user-engagement metrics.

One other thing to bear in mind is that, because Google is only running the Panda updates manually at the moment, any changes you make to improve a site that has been penalised in relation to this may not be updated for several weeks and the rank would only improve again when they push out another version of the algorithm.


Latent Semantic Indexing (LSI)

LSI content is basically keyword-rich content that is related to the keywords you are trying to rank for. You may have noticed this in the search engine with phrases that are bolded in the SERPs that you didn’t actually search for.

This happens because Google considers these other bolded phrases to be highly relevant to the keyword you originally searched for.

I personally find that if you write natural content about the main keyword phrase you are trying to rank for, then it’s difficult to do so without also using it’s semantic variations.

Instead of trying to fit the exact target keyword into your content unnaturally, just write naturally and it’s likely that Google already understands the semantic relationship.

The lack of importance in using your exact target phrase within the page content was also highlighted recently in the Panda webinar and I think the average keyword density across thousands of pages ranked in the top 3 positions was as low as 0.5%.

The bottom line is write good, long, engaging content in a way that naturally uses the target keyword and it’s semantically related variations.