Google has for a long while indexed PDF files and their contents. However, it seems to be less well known that Google not only follows links from within PDF files too, but that these links both flow PageRank and pass anchor text benefit. Various commentators I read differed on whether the links are devalued relative to an equivalent HTML link, but some benefit is certainly attainable.

(The subject came up because a certain client’s press release handling was recently centralised with its US arm, meaning that their new press releases were being hosted off-site and in PDF. This suggested an opportunity for happy international vampirism via links in the PDFs. Sadly, I was thwarted in this.)

The major downside to using PDFs seems to be that they acquire incoming links much less readily than the same content in HTML - Rand of SEOMoz claims that people are less than half as likely to link to a PDF, presumably because they are still seen as inaccessible. Hmm. At least one of our ebook creators has been paying attention to PDF linking anyway; the recent iCrossing “How to Start Blogging” ebook contains a single highly targeted link back to icrossing.co.uk. Limiting the number of links is useful in a PDF as nofollow cannot be used to selectively direct link benefit - recent news about nofollow notwithstanding!

I was also interested to read that (allegedly) no duplicate content penalty is assessed against HTML that duplicates the content of an available PDF; perhaps we should take the time to offer any new iCrossing ebook in both formats?

Post from: Connect

PDF Links

The issue of quality score has been receiving even more press than usual of late.  Unsurprisingly the consensus is that click through rate (CTR) is king.  We all know the ins and outs of maximising your CTR to ensure the core generics and brand terms get maximum traffic for minimum cost, but how do we get the mid to long tail performing well?

No Sandbox for the Long Tail

When a keyword with low traffic volumes is launched it takes far longer for you to prove its worth to Google.  We often see high volume terms appearing surprisingly high in the rankings when they are newly launched, before bedding down to a normal level.  This is the paid search version of the ’sandbox’ where Google can quickly assess whether your ad is any good and assign a true quality score pretty quickly.  For the less popular terms though Google doesn’t rely on this method, rather it takes the word history from everyone else’s campaigns.  In this way if you have an obscure term albeit relevant to your account, if everyone else has used it poorly in their campaigns your quality score is likely to be low.  Unfortunately the only way to combat this is to raise the bids… 

Hang on, That Sounds Like Yahoo! 2004

Though this feels like a bit of a step back, the key thing to remember here is prevention is better than a cure.  If you launch your campaign with the bids as you’d like them in month 3, the chances are that your mid and long tail terms won’t perform as expected.  Over time you’ll tweak ads, raise bids and probably see minimal improvement as they get trapped in a cycle of lowering CTRs and quality scores. 

Back Yourself to be Bold

A much better approach is to go in with quite high bids on these terms.  The risk of blowing your budget is minimal as the point of these terms is that they have low search volume, you should be monitoring a new account regularly anyway so nothing will get out of control.  Once the quality and relevance of your terms has been established the bids will lower naturally and you can bring them down to the desired level with ease.

Land on Your Feet

It’s not so much a case of hitting the ground running with these, but ensuring you don’t twist your ankle before you set off.

Post from: Connect

Launching the Long Tail

An interesting new study by the guys over at Catalyst Group shows that although users like Microsoft’s new search engine Bing, they are unlikely to switch.  From a usability focus group of 12 (expected to capture 80-90% of issues), after using both engines 8 said they preferred Google while 4 said they liked Bing.  That being despite the fact that 82% preferred Bings design, 64% preferred Bings organisation of features and another 64% preferred Bings refinement and filtering options.  With regards to relevance of results, the majority of users thought both engines preformed equally well.  I think  this goes to show how entrenched Google has become in our thinking when it comes to search.  Even though Microsoft won 3 out of the 4 tests and tied the last, two thirds of users were unprepared to switch.  After all, we no longer search for something, we Google it!

What I found particularly insightful, was in seeking out participants for this study the researchers were unable to find anyone who hadn’t heard of Google, but none of the participants had heard of Bing.  Granted it’s a new engine, but it just goes to show the marketing mountain Microsoft has to climb if it wants to even get close to knocking Google off the top spot (although I’m sure having $100 million to throw at the problem won’t hurt!)

When it came to paid ads (searching for ‘digital cameras’) it seems that again Bing came out top with users spending 150% more time looking at the ad space at the top of the page (4.9s vs 2.0s), possibly due to the refinement options available at the top of the page.  With Google fairing a little better when it came to ‘Hotels’ search (2.7s vs 1.7s).

(Click image for larger version)

One of the more innovative features found in Bing is its ‘quick preview’ option that lets you get an additional snippet of information on a site without having to visit it.  However this feature was only discovered by 1 of the 12 users (and only by accident).  When told about this feature, most users felt it wasn’t very useful and didn’t place much value in it, citing the information was not always representative of the site and often simply copied the information already presented in the search results.

What all of the above shows is that although Bing is far better than its predecessor, with so little to choose between Google and Bings search results and the fact that Google is now such a household name it is unlikely that people will switch in the droves that Microsoft would likely hope for.  Unless Bing has a few extra tricks up its sleeve don’t expect Google to be dethroned anytime soon.

Catalyst Group Bing V. Google Usability Study -

Post from: Connect

Users prefer Bing but unwilling to switch

The SEO industry has been awash with lively discussion since SMX Advanced in Seattle a couple of weeks ago.  At the event, Matt Cutts (head of Google’s Webspam team) let slip with a few comments that suggested that PageRank sculpting no longer worked.  The exact nature of his comments have subsequently been torn apart by shrewd SEOs, and so early this week, Matt decided to clear up the situation with this blog post.

PageRank sculpting is the practice of manipulating the flow of PageRank throughout a website using the NoFollow tag, so that the most important pages get the most PageRank.  The benefits of this technique have long been debated, with some webmasters apparently seeing significant improvements in terms of visibility and traffic following the implementation of PageRank sculpting, while others seeing no change at all.

Matt Cutts yesterday dispelled the myths by clarifying exactly what happens to PageRank that meets a NoFollowed link - rather than being channeled elsewhere within the website, it simply evaporates.  To add insult to injury, this “evaporation” has been happening for over a year, leaving many SEOs with egg on their face; plenty still thought this was a valid and useful amendment to website architecture.

Matt Cutts has been issuing veiled references to the fact that NoFollowing probably isn’t the best use of a webmaster’s time for a while.  Here he responds to Randip Dhesi - one of iCrossing’s Natural Search Analysts - who asks “what are your views on PageRank sculpting”.

Matt responds by saying that you’re better off ensuring your website has top quality content and great links.  Nofollows should be used to keep search engines away from certain types of page - a login page being a good example.  He also suggests that PageRank sculpting should be as something you do without the NoFollow tag - i.e. by reviewing your internal linking structure, and making sure you’re linking to the correct page from the right place.

Potentially this could really change the SEO landscape – lots of people are talking about what this means in terms of linking to other websites.  For example, the Wikipedia NoFollows links to external sources. While this is the correct use of the NoFollow tag in this case, it now means that the Wikipedia is losing significant amounts of PageRank. It’s unlikely that the Wikipedia will stop linking out, but some webmasters will be much more precious when it comes to their PageRank.

This has an even bigger impact when you consider blog commenting. Commenters are traditionally allowed to leave a link to their own website when adding value to someone elses blog post. If you’re the owner of a blog, it now means you’ll be losing PageRank everytime somebody leaves a comment. Precious bloggers may prevent visitors from leaving links, which in turn may reduce the number of people prepared to comment. It’s a vicious circle that could erode the virtues of the blogosphere.

For more information on this, it’s worth following the comments over at Matt Cutts’ blog, and opinions on this over at SEOMoz.

Post from: Connect

Page Rank Sculpting Debunked

Diamond:

Coming in from outside the top 30 and ranking in position 9 on Wednesday. Bell and Elephant also coming in from outside the top 30 and ranking in position 9 and 10 on Thursday:

Now these results are not campaign changing. Even getting to bottom few listings of the first page wont drive that much traffic - though I guess more traffic than any previous position may have been getting. However, today we are seeing a new entry at position 1 for the first time in a very long time with Direct Line:

I wonder how much traffic/money MoneySupermarket and Confused.com are losing for every hour that they are not in their regular positions of 1 and 2. An important thing to bear in mind is these results are very temporary. They can change from day to day or even when you refresh the SERP. You need to wait for consistent rankings before you react to any changes, other wise you will never be able to form a useful strategy.

One thing that is for certain Google is trying out some new algorithm tweaks and rolling back to previous versions.

Could this be preparing the way for the Vince Update that was released in the USA earlier this year.

Could it be more fallout following Google’s announcement regarding PR sculpting - although reportedly that update happened nearly a year ago and so why are we only seeing the effects now?

Would be great to hear your thoughts on what’s a foot with Google and anywhere else we are seeing a good old fashioned Google Dance!

Post from: Connect

Car insurance SERPS - Google keeping us on our toes!

Yes, as a journalist, a failure to disclose a conflict of interest was not just wrong from a legal standpoint but against the very essence of journalism, which, even in the business press, we liked to think was to unearth and fully reveal ‘the truth’.

And, yes, bloggers, in contrast, come from all walks of life and, in many cases, are driven by an imperative far removed from such pompous ideals. Most often, the imperative is commercial, whether to build a profile, increase a site’s SEO ranking, or, even - as a ‘media owner’ and affiliate - drive traffic and sales to advertisers.

And, true, most often - since this is not formal ‘publishing’ - there are no clear boundaries between commercial and editorial, which means that the editorial output can be indistinguishable from a blog’s commercial needs.

But, if anyone was in any doubt, all editorial, save that produced by the BBC and Channel 4 under their public broadcasting mandate, is driven by a commercial imperative.

And one of the impacts of digital media on traditional media businesses has been the steady erosion of the once sturdy Chinese walls between editorial and commercial. Trust me, the pressure to produce advertiser-friendly content has never been higher.

Also, web 2.0 platforms and social media in general, function entirely on the principle that, not only will the cream rise to the top so that ‘real journalism’ becomes most popular (and therefore more commercially successful), but that the implications of being found out in a piece of jiggery pokery are so serious as to far outweigh the benefits (consider Habitat’s recent Twitter-fail).

So, I get paid to blog for iCrossing. I am also - on a consultancy basis - its head of marketing. That might potentially bring up conflicts of interests if a reader were expecting thoroughly independent comment about digital marketing to be appearing on a digital marketing agency’s own website. As it is, I often make reference to this arrangement, just in case.

And, although what the FTC is after is much more understandable - click-trading masquerading as independent review and fraudulent misrepresentation (’this company is great’) for commercial gain (’I own a stake in it’) - it should have more confidence in the web-enabled democratic process.

Post from: Connect

The FTC should trust in web-ocracy

Since its launch around 3 weeks ago there has been a lot of industry buzz and discussion about Bing’s pros and cons, whether the results are any good, and generally if it is actually any better than MSN’s previous incarnations. Read Jonny’s to Bing or not to Bing post for a more general overview. The consensus seems to be it is better, not just in appearance but also usability and effectiveness.

From a paid search perspective however we are really not that bothered by all this - we simply want to know if it gives ROI!

With this in mind I’ve taken conversion and click data from the past 3 weeks and compared it to the previous three weeks on MSN/Bing - and some trends have shown up already.

As impressions and spend will vary according to budget allocation and other factors, I’m more interested in the cost-per-click, cost-per-order and sale/lead conversion rate.

Retail Clients

Looking at iCrossing’s retail clients first the results, of course, vary depending on the account. One client has seen the CPC more than double whilst another’s has fallen by 33%.

Overall the CPC has increased by 55% which has, in turn, led to a 40% increase in CPO overall - bad news for Bing so far.

The good news is that conversion is up by 11% with one client experiencing a 62% increase in the conversion rate.

Travel & Leisure

Our travel and leisure clients have not seen such pronounced changes overall. The biggest increase in CPC was only 14% with one client experiencing a 66% drop in CPC.

It evened out across all clients with CPC working out under 1% different. The same is true for CPO and conversion rate with both almost unchanged.

Other verticals such as B2B have seen CPC up by 12% but conversion up by 31% and CPO down by 14% - better for Bing without doubt. So it seems, in its first 3 weeks at least, Bing has had more effect on retail than other verticals with traffic and conversions. costing more.

Across all accounts the CPC is up by 17% - not good until we consider that the conversion rate is up by 18% and CPO has fallen by 1%.

To conclude, it seems Bing is a good improvement on Live Search and its predecessors in many ways. It seems a better option for PPC too; you’ll probably pay more but will get more for your money.

It remains to be seen if Bing can overcome MSN’s previous shortcomings, principally its unpredictability from one month to the next, and low traffic of course. but overall it seems they’re off to a good start.

Don’t open the champagne yet though. Microsoft - put it on ice for 6 months, we’ll give you a pat on the back for the timebeing

Post from: Connect

Bing paid search: Microsoft - put the champagne on ice

Unfortunately there is no magic formula to writing highly effective paid search but here at iCrossing our approach is led by a simple four-step process which we’ve seen get great results time and again: Test, Analyse, Refine and Repeat.

Test

Develop a hypothesis and implement a change.  Keep your test simple and manageable. Ensure you capture enough data to make a decision.

Analyse

What was the impact of the test?  What was affected? Was the result positive or negative?  Can you call the test a success?

Refine

What learnings from the test can be implemented on a wider scale? What changes were not successful and need rolling back?

Repeat

Keep up the momentum!  Most testing programmes fail because people don’t keep at them long enough. 

The key to success is iteration and tenacity - you may not see stellar results from your first test but persevere and you will be rewarded.  The testing process is simple but highly effective:  choose a high traffic adgroup, write a new creative and run it against your existing copy (make sure that you tick the ‘rotate’ campaign setting in Adwords to ensure a more even split of impression).  Analyse the results and keep the winner.  Keep testing and beat your last set of results, extending your tests to cover more adgroups and campaigns.

To help get you started with your own testing programme for paid search ad copy I’ve sketched out 6 different ideas, which if you work through methodically (not all are going to yield improvements so make sure you don’t skip the ‘refine’ step) I guarantee will improve your conversion rates!  

1.  Ditch your brand terms

You’re brand is an asset and I’m not suggesting stopping bidding on your brand.  What I am suggesting however, is that you may be better off without your brand term in your copy for generic keywords.  Why? Well, the characters saved by cutting out the brand term (which users will still be able to see in your display URL anyway) may well mean you can better convey your USPs and get those extra clicks.

2. Test different ‘calls to action’

Do your ads tell users what you would like them to do?  If not, they should!  People are far more likely to buy or apply for your goods or services if you ask them (nicely) to.  Test different ‘calls to action’ and see which garner you the highest response rates.

3. Differentiate your messaging

Paid search SERPs, at their worst, can be a bit bland and repetitive due to limited character limits and a high pressure for relevance.  Use this propensity for homogeny to your advantage and differentiate your message from the 9 other ads on the page which all promise “Cheap car insurance” (or whatever).  Get creative.  You may initially increase CPCs but a higher CTR from standing out from the crown will soon see them drop again.

4.  Use ‘official’, ‘authentic’ or ‘TM’ symbol in your copy

Conveying to users that you’re the official vendor for your product or service can offer a clickthrough and conversion boost.  Don’t be afraid to draw attention to your quality or market-leading position.

5. Include a price point

The extent to which this tactic will work will largely depend on the competitiveness of the price on offer.  If you know you’re price is good, shout about it.  If your price is not so good, this will likely reduce your clickthrough rates although a potential upside may be that it reduces your overall cost per acquisition (as you’re only going to be paying for more qualified clicks).

6. Include a phone number

Get yourself a unique phone number for paid search and include it in your copy.  Even though people are searching online, many will still take the opportunity to speak to a sales person (and thus also an opportunity for your team to help close a sale) if you let them - especially if it’s a high-consideration purchase.    Including a number in copy may well save you a few clicks and enhance conversions.

Happy testing and let me know how you get on!

Post from: Connect

6 great tests guaranteed to help you create higher converting PPC ad copy

Image credit: by nolanisan

Last week Yahoo! announced that their search engine would be supporting the Common Tag – a recently released new tagging format for Web pages, to go alongside their Search Monkey functionality. Publishers will be able to use these new semantic tags to make their content more “discoverable, connected and engaging”.

What is a Common Tag?

Tags have been used for a long time to organise  content online. However, in the absence of a common tagging format for Web content, the benefits of tagging have been limited. Individual things like New York City are often represented by multiple tags (e.g., “nyc”, “new_york_city”, and “newyork”), making it difficult to organize relatedl content; and it isn’t always clear what a particular tag represents - does the tag “jaguar” represent the animal, the car company, or the operating system?

Common Tags are defined using RDFa, a standard format for expressing structured data within HTML.

It is worth noting that Google recently announced using RDFa tags to help structure its data – although at this time is not part of the Common Tag group.

An example of the Common Tag in use:

Just as you can tag content in other documents by making reference to the URL for the content prior to the rel=”ctag:tagged” attribute, you can use the same mechanism for referencing any section of content by referencing the “id” attribute for the DOM element containing the content in the URL.

For instance, if you had a document which contained two paragraphs using the markup:

<p id="first">Star Wars set a new standard for visual effects in space drama. </p>
<p id="second">Luke's interest in his heritage made him vulnerable to the Dark Side.  </p>

To indicate that the second paragraph is about Luke Skywalker and not some other Luke, you could tag the content of the second paragraph with the following markup (which could appear anywhere in the document.)

<div xmlns:ctag="http://commontag.org/ns#"
         about="#second" rel="ctag:tagged"> 
    <span typeof="ctag:Tag"
         rel="ctag:means" resource="http://dbpedia.org/resource/Luke_Skywalker"/>
</div>

Note: By using an absolute URL you could tag the second paragraph from another document.

Find other examples of how the Common Tag can used

Uses for the Common Tag :

// Video content can be tagged for semantic terms, avoiding having to transcript the video and having that content appear on the page

// Image search can become more refined

// Content will not always need to have an instance of every type. E.g. a page about ‘Heathrow Airport’ will not need to contain instances of the terms ‘LHR’ or ‘London Heathrow Airport’ to be able rank for these terms.

Why this may not take off:

// Until there is a CMS that has this type of functionality built in Journalist and Copy Writers may not feel they need to be able to add these tags easily enough to warrant using them. Web Developers may not want to constantly have to change the code of a page template whenever new content is added.

// Would a CMS provider warrant it necessary to build in this functionality if larger sites, e.g. Google, are not using these tags?

Post from: Connect

The Common Tag & Semantic Search

It makes us realise that the jargon and terminology  native to us in digital can easily be mis-understood in the offline world.

And you, would you know what a browser was?

Post from: Connect

What is a Browser?