Recently I’ve bought a laptop, a new PC (custom built), and a new TV, and like many customers I like to read impartial expert reviews of products before I buy them. For cheaper products I don’t mind using the customer reviews on places like Amazon, but when I’m spending upwards of £100, I usually like to get an expert’s opinion.

For a long time now this has been all but impossible to do, because both Google & Yahoo’s search results have been flooded with crap from sites such as reviewcentre.com. Often these results are simply holding pages “awaiting” a review - an example of this is shown below:

These pages rank because they’re on domains with a high level of authority, and there is often some unique content such as a price comparison feature - of little value to a user looking for reviews.

Then there are the retailers/affiliates who target the keyword [reviews] in order to benefit from the (likely highly converting) search traffic in this area. Often there are no reviews on these sites either.

Increasingly it seems that the only way to get legitimate reviews is to type in the model #, and then some random feature of the product or some superlative such as ‘good’, ‘great’, ’superb’, etc.

Will someone please build a site that aggregates only expert reviews of a range of products? Please?

I was going to blog about this whole thing, but Graywolf has summarized it well.

All I’d add is that Google are playing a dangerous game by pretending that they can handle duplicate content. What Google are effectively saying here is that people with good intent don’t need to worry about duplication, because Google’s algorithm can handle it. Maybe on a macro scale it can, but on a per-website level, and for the average webmaster, it really can’t.

I don’t buy their line that duplicate content penalties are a “myth” (as I’ve seen a direct 3 month duplicate content penalty on a large site), but I do think that there are filters applied that can effectively kill a website’s ranking for competitive terms, which can amount to the same thing.

Although Susan Moskwa mentions briefly in the post the Webmaster Guideline point:

Don’t create multiple pages, subdomains, or domains with substantially duplicate content.

That point is largely ignored for the rest of the post. But it is the most important point - the best way to avoid duplicate penalties, filters, whatever you want to call it, is to create a good information architecture from the outset. This doesn’t just help the search engines (even if Google appear to down play it in this post), it helps your users.

The issue stems from the use of various URL parameters to track a visitor’s session on a website. This is slightly different to the old session ID parameter problem, in that it doesn’t create quite as many duplicate pages; but it is important nonetheless.

I’ll work through a couple of examples of this; firstly I’ll look at Omniture’s SiteCatalyst product, using Channel 4’s website as a guinea pig. SiteCatalyst’s tracking code appears in a URL as intcmp, which stands for ‘Internal Campaign’, and can be identified quickly in the SERPs using a search like this:

As you can see from the snapshot above, there are nearly 10,000 results with this parameter added to the URL. If you visit any of these URLs you can delete the ‘?intcmp=blah’ section and get exactly the same page. Internal campaigns come and go, and as the campaign disappears, the link to the campaign disappears, but leaves behind an extra URL that the content can be accessed from.

A nice example for Channel 4 is this Internal campaign URL:

• http://www.channel4.com/bigbrother/?intcmp=homepage_flash (PageRank 3)

Clearly this link was indexed through a Flash based link on the home page. When this link disappears from the home page, it leaves this URL behind, sapping authority and causing duplicate content from the original page:

• http://www.channel4.com/bigbrother/ (PageRank 6)

Now moving on to the second culprit, WebTrends, there is a very similar issue here. This time the nasty campaign parameter is WT.[something] - this is broken down in the following image:

This time the chosen victim is a site called LuxuryLink - let’s take a look at the SERP:

Again, a couple of examples from this - the tracked version:

• http://www.luxurylink.com/LL/home_win_trip.php?WT.ac=MOENON03 (PageRank 5)

And the untracked:

• http://www.luxurylink.com/LL/home_win_trip.php (PageRank 4)

So this time the tracked URL has actually gathered more authority than the non-tracked version, but still the page’s authority is being split and duplicate content is created again.

How to fix it

Now this is not necessarily the fault of the web analytics companies, and definitely not the webmasters. There are many reasons why URL parameter tracking needs to be employed, especially in large companies where the split between the marketing team and the web development team is too pronounced to be able to implement HTML code changes with ease.

However the ideal solution is generally to use a combination of on-page JavaScript and browser session cookies in order to track visitors to your site. This doesn’t interfere with your URLs - remember, cool URIs don’t change!

If that is not possible for whatever reason (or if it is and you have URLs with the parameters above indexed in Google), use your robots.txt file, and use it wisely. I’d always recommend using Google Webmaster Central’s robots.txt syntax checking tool before implementing any major changes to your robots.txt file, but these rules should clear out the duplicate URLs:

Omniture:

User-agent: *
Disallow: /*intcmp=

WebTrends:

User-agent: *
Disallow: /*wt.*

Hang on, Bebo - isn’t that a social networking site rather than an airline ticketing company? As outlined by Trend Labs recently in a post entitled Spam Early Spam Often, fake Bebo profiles are being set up and used as a way to lure unsuspecting users into clicking on nefarious links within a profile.

This now appears to be affecting Google’s results too, at least for this query. This profile appears at #3 for the query:

Although these fake profiles have very low quality content and just link to an incredibly spammy domain, they have around 1,000 (spammy) backlinks according to Yahoo, and apparently that’s enough for Google to put them above the fold for a competitive query such as this.

Even though Bebo is a well known social networking site. Not anything to do with airline tickets. Simply being hosted on Bebo is enough to rank for this term.

Bring on the semantic web…

However I think there aren’t enough conspiracy theories about Google’s world domination (rofl) - so here’s another one:

Google releases facial recognition technology and its own browser in the same week in an extraordinary attempt to not only discover everyone’s various IP locations (home, work, mobile - erm Android), track their browsing habits, but also to know everyone’s faces! What a huge amount of granular marketing info!

Scary.

Ubiquity is one of a number of tools currently knocking around that look into changing the way we interface and interact with the web. It aims to break down the barriers to mashups and allow the average user to, say embed maps in an email, by simply selecting some text, pressing CTRL-SPACE, and typing ‘map’:

This kind of tool has implications for all areas of web use and is well worth keeping an eye on. One day soon a tool similar to Ubiquity (if not Ubiquity itself), will emerge as a game-changer, and as web professionals we need to be considering where web use is headed; tools such as this (and the excellent work of Mozilla Labs in general) provide some good clues.

This indicates to me that there is a likely PageRank penalty of 6 months from the date that a site is flagged as badware. Of course presumably sites must rectify the badware problem in order to get their PageRank reinstated.

People are just better together. I am who I am because of everyone.

Bleugh. It’s just a mobile phone network.

Anyway, beyond the sheer gut-wrenching self-satisfied, clever-cleverness of these adverts, the thing that has got a few people’s attention is the addition of a subtle little “search online for ‘i am’” in the bottom left.

What? No URL? But how will we remember to Google it when we get to the office? Well besides the fact that the advertising campaign is instantly forgettable for all but this niggling little thing, I thought “fair enough, they’re trying something new, and it’s to do with search so that must be good”.

So I did remember it and Googled ‘i am’. So apart from the sponsored link at the top, where’s their website? Surely months of planning and meticulous execution have gone into the organic SEO for this campaign? Apparently not. The first time I searched online for ‘i am’, their website was hovering at #70 in Google, just below a site called ‘arseiam.com’ - how ironic

Even now, this site sits at #6 in Google, below the #1 result i-am-bored.com, and is nowhere to be seen on Yahoo or Live search. Astoundingly, Orange did not see fit to do home page takeovers on any of these other sites. Surely if you can’t dominate the natural listings you can buy your way onto the destination pages of other sites that rank better than you?

So here’s a couple of pointers for future offline ’search online for…’ billboard campaigns:

1. Make the advert interesting - make people want to search for something to find out more.
2. Create some kind of mystique around the campaign. Take some pointers from Nine Inch Nails, who did a great online treasure-hunt style campaign last year.
3. Give something away for FREE!! Provide an incentive for people to look for your site.
4. If you’re lucky, people will search for your term on their iPhone or smart phone - help them want to by giving them entertainment (see point 1), a reward (see point 3), or instant gratification (make of that what you will).
5. Choose your key phrase wisely. Don’t allow it to be pushed down from the top. Have meetings about it. Look at the competition for that phrase, and ask yourselves “are we good enough to rank #1 for that in every major search engine?” Don’t let the answer be “no” or “maybe”.
6. Plan and execute a great SEO campaign before you spend any money on offline advertising. Look at dominating the SERPs for your chosen term.
7. See point 1.

According to mad.co.uk, £30 million is going into this campaign. Come on Orange, you can work miracles with that kind of cash can’t you (and your PR7 main site)?

Note: Of course I’m not going to link to this abysmal site, but it’s

• i [dash] am [dash] everyone [dot] co [dot] uk