malcolm coles

If you know anything about SEO, there's a fascinating Q&A with the Google search quality team over at Digital Inspiration. Here's some analysis of what they had to say, which includes:

• 

  Digital Inspiration has interviewed five of the Google search quality team

  If you don't nofollow affiliate links, your search engine rankings will suffer.

• Links in copy ARE worth more than other ones.
• Fresh content isn't a ranking factor.
• Rel=canonical is suggested for cross-domain redirects - 301 isn't mentioned.
• Google doesn't seem to like guest blogging.
• Linking out both benefits you and doesn't benefit you at the same time. (Hey, don't shoot the messenger).

Of course, just because they say it, doesn't mean it's true. But here goes.

Finally a statement on affiliate links and nofollow ...

... just not the one anyone wanted.

Q10 is about adding affiliate links to blog posts. You might have an Amazon account, for instance, and decide to include a link to buy something that you've reviewed - knowing that you get commission for any sales that result from the traffic.

To many people, this doesn't count as a paid link. You haven't been asked to put it there, you've chosen to put it there yourself. So should you nofollow the link to stay within Google's guidelines or not?

What Google now says

Google has faffed around for some time about affiliate links, and has appeared reluctant, in the past, to say they definitely count, in its eyes, as paid links.

So the first part of the answer seems gracious enough - it's fine in Google's eyes to include links like this. Jeez, thanks for the permission. Sarcasm mode over, the second half of the answer looks a bit more troubling:

While it is legitimate for a webmaster to monetize great content, in order to perform well in Google’s search results it is important to take technical steps in order to prevent unnatural passing of PageRank through paid links, e.g. by either using the “nofollow” attribute or by creating a robots.txt file.

What it's said in the past

This seems to go further than in Google's official statement on paid links, which doesn't really address the affiliate issue. In particular, its page on affiliate programs doesn't mention nofollowing the links. (If you'd like to have these points reinterpreted in a more ranty way, read SEOmofo's comment half way down here).

It also goes further than Matt Cutts did in a recent interview where he suggested that Google just ignored affiliate links:

Typically, we want to handle those sorts of links appropriately. A lot of the time, that means that the link is essentially driving people for money, so we usually would not count those as an endorsement.

Google now seems to be saying that "in order to perform well" you should nofollow the link, and not just rely on Google working it out and ignoring it. The threat of what happens if you don't nofollow it is clear ...

Are links in copy worth more?

Things like block level analysis have been around so long, I'm surprised anyone is surprised that Google would treat links in copy as more important that footer links or auto-generated related content lists.

Anyway, Q13 is this:

"Are all links on a page treated the same or does the order of links matter. For instance, will Google flow more juice to the links that are in the first paragraph of the story than the ones that are in the page footer?"

The answer is this:

Our link analysis is getting much more sophisticated than the original PageRank used to be. To answer your question, we may treat links across different areas in a different way, as some areas of a page might not be as relevant to the content of the page as others.

Reasonable surfer patent

Which would not be inconsistent with the reasonable surfer patent that was recently granted to Google (which they may or may not be using in that form these days - but the direction of travel seems clear.)

Fresh content doesn't matter

I never really got this "fresh content is important for SEO" stuff people trot out. The suggestion seems to have taken hold that fresh content, in and of itself, helps the rest of your pages rank better.

Matt Cutts was interviewed in March 2009 and was reported as saying:

Matt said that fresh content isn’t part of the algorithm, but it almost always gets you more links than if your site is stale.

The answer to Q20 in this latest Q&A suggests that taking a long period off publishing content will make no difference to your existing rankings.

Compare that with the people prattling on in these pages. If you want your site to feel that it's up to date or you want to target more keywords or earn more links, then create fresh content.

But fresh content itself isn't going to help your old content (actually, it might harm it if your site architecture is poor by pushing old content ever more links away from the home page ...)

Changing your domain name

Question 3 was about handling moving your website. The  answer was:

"One easy way to handle duplicate content across different websites is to use the rel=canonical link element. Other possibilities are included in our blog post about handling legitimate cross-domain duplicate content."

It's interesting that they choose to give rel=canonical as the key method they quote - and not a 301 redirect.

Does this confirm Rob Kerry's observations at SMX London about the death of the external 301? To quote the Seo Insight round up:

In January Google clamped down on smart affiliates using cross-site 301s on affiliate links and stopped them passing value ... the key message is, if you’re moving sites, don’t rely on 301s to save all your link juice – build up the link profile to the new domain with fresh linkbuilding.

Guest blogging: can't you just answer a straight question?

Q11 seems sensible enough. "I've got a new blog but noone knows it's there. So I've guest-blogged  - do the links back to my site count in Google's eyes?"

I mean, "yes" or "no" would seem good answers here. Instead we have this:

In general I would recommend putting that work into your own site, instead of creating content for other people’s sites.

It’s much better to create great content for your blog and to let other sites refer visitors to your site on their own.

That's not the question. The question is "Is Google fine with guest blogging and do links ‘earned’ from writing guest blogs matter?" Are they trying to give the impression that their algorithm looks for guest posts and discounts them - or can't they just say "yes" or "no"?

How does linking out benefit you

The question is:

I know that inbound links will help my site’s ranking in Google search results but is that true for outbound links as well?

The answer is:

No, they don’t contribute directly towards your site’s rankings ... On the other hand, being selective and preferring quality sites to link to might help in how Google perceives your site.

First up, see the yes / no thing above. Surely these two halves of the answer contradict themselves - and it's not a long answer. Or does being perceived better not help your site's rankings directly? Why would it be worth being perceived better then?

Anyway, secondly, Matt Cutts has previously said that:

In the same way that Google trusts sites less when they link to spammy sites or bad neighborhoods, parts of our system encourage links to good sites.

So, I think that, overall, we can take it that linking out is a good idea in general. Sadly, not enough of one to encourage it to happen more. Personally, I'm with Aaron Wall:

[Google] could undo the years of FUD that destroyed the link graph by stating the importance of outbound links, and then putting a bit of weight on it.

Paid link penalties

I've already pointed out the answer to Q9 says that Google appears to have taken action over a UK newspaper group - presumably the Express - offering advertorials with links for SEO benefits. Thanks to Jon Hudghton (and on Twitter) for spotting the interview.

Google says it has "taken action" and no longer trusts links from a major UK newspaper group - apparently referring to the Daily Express website.

Express website

In April, an Express Group email was made public offering links in online "SEO editorials" from £1,000 - with the aim of improving search engine rankings.

Such paid links are against Google's "terms of service" - and the Express (and OK) site initially suffered a page rank penalty in Google's toolbar (ie not much of a punishment at all).

"We saw that"

But Google has now gone further, according to an interview with Google's search quality team at Digital Inspiration. Question 9 asks about newspapers selling links on their sites to companies wanting to rank better in search engines. In reply, Google's Matt Cutts says:

If you’re talking about the recent incident in the UK, we saw that. Google’s quality guidelines are clear on this point: paid links shouldn’t pass PageRank.

Whether the paid links are in an “advertorial” or somewhere else on the page, that would violate our quality guidelines and Google would take action on those violations, both so that the link buyers wouldn’t benefit and so that the link sellers wouldn’t be trusted in the future by Google.

The only incident I know of recently in the UK is the Express case - so Matt Cutts appears to be saying the Express site (and presumably OK site) are no longer trusted by Google's algorithm.

Whatever your views on outing SEO tactics, this would be a fairly severe reaction to not using nofollow on paid links if carried forward. Thanks to Jon Hudghton (his blog's here) for spotting the info.

This is how Bing describes xRank:

"xRank keeps track of notable people and puts them in order for you. We count Bing web searches for movie stars, musicians, and other famous people.

Then we compile our findings into an insightful ranking formula that tells you who the world is searching for most. The result is a cultural snapshot of who's hot and who's not!"

Xrank's top celebrities today

I can reveal that it is, in fact, a load of old rubbish. As you can see from this screenshot, according to Bing, of the top 10 celebrities being searched for today:

• 90s band Ash is number one - maybe Girl from Mars is being rereleased? (Hint: no).
• 90s Manchester band James is number 2.
• Eric Clapton's 60s band Cream is at number 4.
• Prince is at number 5.
• Queen is at number 6.
• And 70s New Romantic band Japan is the 9th most popular celebrity search term.

What a load of garbage.

Ash: ah hah

Ash is first because a volcano is erupting in Iceland, as xRank itself demonstrates. This is the screenshot of its Ash page, showing Ash related stories and headlines - all volcano related.

Ash casts a cloud over Bing's xRank results

James: welcome to the rubbish data

James is second because people search for people called James - not because of a renaissance of interest in the band.

James is a popular search term. But not because of the band.

Cream of the crop. Not

Cream is 4th, but mostly because of food searches

Ice cream anyone?

And the rest

I could go on, but it won't surprise you to learn that Prince of Persia is behind Prince being in there, lots of people search for the Queen as in the monarch, and Japan is a popular country.

Yes, Bing seems to be unable to apply any sorts of filters to search terms when it comes up with its top 10.

Still, as searching Bing for xRank doesn't even return the xRank page first, maybe it's hoping it will quietly be forgotten. If you are interested in data around topical searches, I did a post comparing Google's various methods. There are some more in this roundup of keyword research tips from SMX on SEO Insight.

Mobile versions of websites - what a train crash they often are. As I write this, at 10.30pm on Monday night, neither the BBC nor Guardian mobile websites are mentioning that Gordon Brown has promised to resign ... a story that their web news pages are unsurprising leading with - and have been doing so for several hours.

I'm sure it's hard to implement a mobile version of a website - with ITV's mobile site particularly demonstrating why they're often rubbish.

But I expected rather more of the BBC and the Guardian. In fact, if you've been using their mobile sites for the last few days, you'd have been under the impression that there has been little news about the election to report...

The Guardian

Here's the Guardian's mobile version tonight - no mention of Gordon Brown quitting.

Guardian: no sign of resigning PMs

And here's its web version - leading with the news.

Guardian web: Brown's resigned

The BBC

Here's the BBC mobile site on Sunday morning - literally no mention of the election whatsoever. It's been like this for days...

BBC mobile: any election news?

And here's the web version of its news pages at the same time. Oh, there's an election.

BBC web: there's an election!

And tonight, again, there is no mention of the election and no mention of Brown's resignation on the mobile version of the BBC's news pages:

Brown has resigned, I'm sure ...

Back to the drawing board, please.

Here is some information from the Electoral Commission website that relates to the scandal of people not being able to vote. It covers the 10pm issue.

According to this popwerpoint document from the Electoral Commission, entitled "Polling Station Staff Briefing Session - European Parliamentary election - 4 June 2009":

Electoral Commission rules on 10pm

Close of Poll

Must close at 10 pm

Make sure the doors are closed

Anyone who has been issued with a ballot paper by 10 pm must be allowed to vote

However cannot issue after 10pm, even if elector was in a queue at 10pm

Here's another training quiz for polling station staff (word document):

Q.12 Just after 10pm, a person hands in a completed postal vote. What do you do with it?

Accept the envelope and hand it in at the count with the others

Refuse to accept the envelope

Send the person to the elections office

A.12: If a person hands in a postal vote just after 10pm they should be told that it cannot be accepted as the deadline for receipt of postal votes has passed.

This document, the current handbook for polling station staff (PDF), says:

No one may be issued with a ballot paper after 10pm even if they are inside the polling station and waiting to receive their ballot paper.

Warning from Electoral Commission in "Issues for consideration in connection with a potential combined UK Parliamentary general and local government election in England on 6 May 2010":

9.1 The issues affecting polling stations are highlighted below:-

• Staffing – polling station staff will need to deal with the marking of the register, the CNL and the issue of two ballot papers – consider the implications for staff numbers and the training that will be necessary.

• Voting booths – consider if there is a need to increase the number of booths due to potentially higher turnout and voters taking longer to mark two ballot papers than one.

This training document has a module about "Explain how to deal with a queue of voters as only those issued with ballot papers by 10pm can vote." No details gives.

I'm not quite sure why the Sun runs Sun Vote as it just ignores it (EG when its readers aren't that fussed over a hung parliament).

But its own on-site poll from last night about who won the debate (and unlike other sites, these aren't easy polls to vote in - you have to go through a lengthy sign-up process)  reveals that Clegg won. Asked "Who won the 3rd debate", Sun readers voted:

• Nick Clegg 50.9%
• David Cameron 29.6%
• Gordon Brown 18.6%

Not sure how this squares with Clegg being toast as its main story says ...

Sun Vote results

Clegg: wins 3rd debate

Main story

Clegg: toast

What a surprise. Not.

Google appears to be trialling some new related-search links over at Google.com - a list of "brands for ..." when you search for products. (Update: official confirmation).

So search for digital cameras at google.com (NB not everyone sees this, so it must be some sort of experiment) and it shows, immediately under the search box, "brands for Canon Olympus Sony Nikon Kodak" - where each is a link (leading to a search result for your original search term plus the brand name). Like this screenshot.

New "brands for" links

US now - UK in the future?

Click these screenshots to see them full size. Or squint and you'll see these extra links above the news, which is above the shopping results - all of which are pushing the natural results way down the page.

How the US results look when you search for digital cameras

Here's another example, also featuring a local search box, too, for mattresses.

Mattresses: brand list, local search box, the lot ...

And here's one for cheap laptops - Google associates the brands Dell, Acer, HP, ASUS and Apple with this search term. It's editorialising now! (And Apple, really?!?)

Brands for cheap laptops: Dell Acer HP ASUS Apple

Oooh and Katie Clarke pointed out an interesting one: car (as with all these, click to see it bigger). That gives smart as the first option. Lucky them!

Cars: smart comes first

UK now

The current UK results

The shape of things to come? (You might like my previous post on the new look google local results and the implications for natural search).

The Salisbury Journal has revealed that its "dog bumps nose" story (no, I'm not making this up) has been read by 130,000 people - making it the paper's most read article ever. You'll realise how extraordinary this is when you read the story ...

Original story

Here's what it said ...

POLICE in Ringwood are investigating the circumstances surrounding an injury to a dog’s nose.

A woman was walking her dog and puppy, when her puppy ran off.

Her dog ran after the puppy but collided with a male dog walker and his dog.

The woman's dog received an injury to his nose.

I'm not joking. That really was the original story. All of it.

Follow up story

Wait, there's more - a follow up.

MORE than 130,000 people logged on to view our top story today.

Even Mock the Week star Dara O’Briain tweeted about it.

What’s that, I hear you cry?

Was Cheryl Cole kissing David Beckham in the cathedral?

Nope.

Our most viewed story ever involved a dog, with an injured nose.

At only four sentences long, it probably breaks most records for page impressions per word anywhere on the web.

So, how does it smell? One commentator asked.

The original story now has 53 comments, blaming NuLiebor and immigrants for the problem. They're joking. I think.

Hat tip to Drew Broomhall.

I got an email yesterday about what the Times will be offering once the paywall comes down (or does it go up - I think of it like a portcullis ...) Here's a screenshot of the main bit, explaining the culture planner and live debates.

Times paywall email

The rest of the email says that

Every day we'll hold live debates and discussions at thetimes.co.uk. Covering everything from sports, to fashion, to business and the latest issues in the news, you'll be able to engage with the biggest names and put experts on the spot if you wish.

On the sundaytimes.co.uk you can plan your perfect night out, with our unique interactive culture planner. Book tickets, remotely record TV shows, and stay informed with the help of our critics.

These are just two of the many new features you'll find on our new websites. Soon you'll be able to do more than just read the news - you can debate it, interrogate it, watch it, shape it and be part of it 24 hours a day.

Interesting

According to the Sun, a hung Parliament would be a disaster for Britain, while David Cameron "left his opponents reeling" in the second election debate.

The Sun's online polls tell a different story. (And, unlike the Daily Mail's, these are not easy votes to polljack. You must go through a proper sign-up process with email verification and provide personal demographic information).

Sun readers split over hung Parliament

Sun readers split over hung parliament

In a Sun poll that just closed this morning (suspiciously soon after the yes vote changed from just under 50% (last night) to just over 50%), they asked: "Would a hung Parliament be bad for Britain?".

Just 50.6% say yes. 46.2% say no - with the rest as don't knows. So fairly evenly split on the "disaster".

Sun readers: Clegg won 2nd debate

Sun readers: Clegg not Cameron won second leaders debate

When it comes to the second general election debate, the Sun readers, unlike Sun HQ, thinks Clegg won it. This is the vote as I type this:

• Nick Clegg 46.3%
• David Cameron 36.6%
• Gordon Brown 14.2%
• There was no clear winner 1.6%
• I didn't watch the debate 1.4%

A crushing Clegg win then ...

Hope for us all yet

Maybe the Sun's attempts to bully the British public (and other newspapers) into backing Cameron aren't working, just as they misjudged the mood over Brown's handwritten letter to a dead soldier's mother. Let's hope so.

There was a story yesterday that the Express has been emailing SEOs selling links. I went over to the Express home page. And was fairly shocked to see it has a toolbar pagerank of just 4, which seems incredibly low for a newspaper site (er, it puts it on a par with my homepage toolbar pagerank!).

It definitely used to be higher than this.

This blog from April 2009 had it as PR7 (scroll down to the graph). As does SiteLogr (which collates data on websites). The same is true of Surcentro.

Update The OK website has also been hit. Its toolbar pagerank is now 2 as Andrew points out in the comments.

• 154 days ago it was PR 5 according to realwebsiteworth.com.
• 39 days ago it was PR 3 according to website shadow.
• 7 days ago it was PR 2 according to sitelogr.com.

I originally said it looks like Google has reacted to the revelation by penalising the visible pagerank of the Express. This data for OK suggests its being downgraded for some time - but did take another hit in the last month.

Express: pagerank penalty

Interesting.

Since the first TV debate, Nick Clegg has started to be taken seriously by the newspapers (or been the victim of a series of hatchet jobs, depending on your point of view).

The charts show how the number of stories about Nick Clegg has soared in The Sun, The Daily Mail and The Guardian - even allowing for the fact a general election is on.

Comparing April with January:

• Daily Mail: 37% more stories about David Cameron - more than 4 times as many about Nick Clegg.
• The Sun: 27% more stories about Cameron - more than 11 times as many about Clegg.
• The Guardian: 80% more stories about Cameron - more than 6 times as many about Clegg.

Daily Mail

Clegg bounce in the Daily Mail

Sun

Clegg bounce in The Sun

Guardian

Clegg bounce in The Guardian

Figures calculated using the sites' own search function. Pictures from My David Cameron and Daily Mail. April figures are until today, 22 April.

The News of the World has trumpeted a poll of mums claiming Cameron is "storming" into the lead.

The figures show the opposite of this - read my dissection of the stats over at The Media Blog.

And shame on Netmums for encouraging this.

It's so quiet in the south of England - and probably the rest of it too. So once everyone's home and the ash has cleared up, how about we keep the airports shut?

Join my Facebook group: Let's keep the airports shut - the ash means it's nice and quiet everywhere.

The Conservatives and Labour have now launched their manifestos. The Liberal Democrats launch theirs tomorrow.

As with their party leaders' websites, the parties' online approaches to their manifestos leave a lot to be desired - although Labour have done a better job than the Conservatives.

What they should have done

The launch of a manifesto is a tricky thing - you don't want the details to leak out beforehand - but the moment it's published, you want people to find it when they type "party name manifesto" into Google.

Living URL

What you should do is have a "living URL" for your manifesto - one URL (like ...com/manifesto) on which you always publish your current manifesto with links to older ones. That way, your current manifesto will benefit from links to that URL (even from when old manifestos were kept on it).

In the run up to the launch, you spray this URL around liberally, and get your army of bloggers to link to this URL, maybe encouraging them to use handy anchor text like "party name manifesto".

Home page

If you've forgotten to do this, another suggestion is to make sure your home page HTML title contains the word manifesto - that way maybe Google will return your homepage when people search for your manifesto - better than nothing ...

The manifesto itself

It's fine to do the manifesto as a video, podcast and downloadable PDF - but you also need it as a series of webpages with rich interlinking. Some people will want to read it this way - and search engines will definitely want to see it this way.

What they've actually done

Labour party

They've made an effort at a living URL. According to wikinews, the 2005 manifesto was on http://www.labour.org.uk/manifesto.html, a URL which now redirects to http://members.labour.org.uk/ - so at least they've put a redirect in place, even if it's not to their manifesto, which is acutally on http://www2.labour.org.uk/manifesto-splash.

If you search for labour manifesto on google, you see this screenshot.

Labour manifesto search results

Their homepage appears with an HTML title of "The Labour Party Manifesto 2010" - and they've temporarily redirected their homepage to their manifesto homepage, so this all works nicely from a searcher's point of view.

You may notice the second result - the Membersnet 3.2.4 Beta with a URL of http://www.labour.org.uk/manifesto.htm. That's because of old links pointing to that page from when it did contain the manifesto (but which now redirects to the members' site (via a 302 temporary redirect. Sigh).

The manifesto itself is a series of HTML pages which link to each other plus a video.

So not perfect, but at least it's all findable.

Conservative party

They only launched their manifesto today. But they have made a pig's ear of it.

If you search for conservative manifesto, you get the Daily Telegraph website followed by two old pages about other manifestos.

Conservative manifesto search results

One reason is that, according to wikinews, the 2005 manifesto was on http://www.conservatives.com/tile.do?def=manifesto.index.page, which they haven't redirected - you just get page not found. The current manifesto is on http://www.conservatives.com/Policy/Manifesto.aspx.

Also, there is no HTML version - just PDFs and MP3s, not ideal from a searcher's or SEO point of view.

Lib Dems

They launch theirs tomorrow. Preparations are not going that well ... On the one hand, search for lib dem manifesto today, and at least their home page is first. Sadly, the word manifesto isn't mentioned on it. There is a list of what they stand for, but I want the actual manifesto - or at least news of when it launches.

Lib dem manifesto search results

And what's all that other stuff underneath?

In particular, they look like they have a subdomain for their manifesto: manifesto.libdems.org.uk.

This informs us that:

Although the next General Election does not have to take place until 2010, we have to be prepared for a snap Election and that is why we need your views now on what our key messages should be.

Guys - IT'S HAPPENING! (Hat tip to Martin Belam for the idea).

I've reviewed the online presence of Gordon Brown, David Cameron and Nick Clegg over at econsultancy. Why not take a look?

Having trouble verifying your Bitly Pro account? So am I. Here's what I've learned.

To set it up, you need to log on and go to http://bitly.pro/settings where you need to verify two things - that your short URL is yours (I'm using mcol.es) and that the domain you want stats for is also yours.

To verify that the short URL is yours, you need to change the A record in your domain registrar's DNS settings. Your registrar can help you with this, hopefully...

To verify the domain you want stats for (ie where the short URL will redirect to), you need to do one of adding some metadata, uploading an HTML file or creating a CNAME record. The first two are easier ...

As things stand, I'm having no luck with either verification, however.

The key things appears to be to wait. The welcome email says this (which I didn't notice at first as I was too excited):

Please keep in mind that your domain verification process may take anywhere from 10 minutes to 24 hours, depending on your DNS settings and the nature of your website.

The help text also says that:

First, note that DNS changes can take up to 24 hours to propagate. Also make sure that you are logging in with your bitly.Pro credentials, if they are not the same as your regular bit.ly login.

So the verification of your short domain will take time. You need to change the A record setting then wait. And I can confirm that, after waiting a day or so, this then verified.

This doesn't explain why I can't verify the long domain, especially as UncleWilco and John Garcia can. Hmm.

If you're testing it after it's been verified, note that the root domain redirects to bit.ly - this isn't a bug, as the help text also explains:

Can I redirect the root domain for my Custom Short Domain to my own site?

Redirecting the root domain (for instance, examp.le, with no hash at the end) to a user defined URL is not a feature available in the free version of bitly.Pro at this time. We wanted to give web users seeing a short URL for the first time a way to figure out that it is powered by bit.ly, having the same spam protections and analytics tools. Redirecting the root domain URL to a user defined location will be a service available in our forthcoming premium bitly.Pro Enterprise service.

My advice? Carry out the verifications steps - and then wait!

Johnston Press is dropping the paywall on its local papers, with the number of subscribers said to be in single figures. People have already started to draw conclusions from this:

• "The conclusion is clear - charging for local news online is something of a no-go." (Paid Content)
• Rupert Murdoch: "he’s on to a surefire loser." (Bitter wallet).

However, my view is that the Johnston Press experiment tells us precisely nothing about anything. The reason? Johnston Press had implemented its paywall in the worst way possible. All you can learn from this is that a paywall that makes no attempt to sell the content won't sell any subscriptions.

Why Johnston Press's paywall was rubbish

Here's an example of its former paywall (ie what you saw when you tried to read an article).

Northumberland Gazette paywall

As I've pointed out before in my paywall review, this was rubbish.

It used jargon, describing itself as "premium content", and failed to explain the benefits of subscription. With this particular paper, they described themselves as "the n/a site". As far as I am aware, this bug was never fixed during the whole four month paywall experiment.

How to implement a paywall properly

I won't rehash my paywall review, which had good and bad examples. But let's see what Rupert Murdoch's been doing since I wrote that.

The Times

Here's the current information about the coming paywall for the Times:

Times paywall info

And the Sunday Times:

Sunday Times paywall info

Altogether more engaging and attempting to sell the content - unlike the Johnston Press example.

South Coast Today: a paywall in action

And here's a paywall in operation at another one of Rupert Murdoch's papers, South Coast Today. Here's what you see once you've hit the 10-article limit:

South Coast Today paywall

Then you're taken to some screens which explains more. You can choose to register or pay:

South Coast today: register or subscribe

Not sure of the benefits of subscribing over just registering?

Registration vs subscription compared

Interestingly, you can even get the site and paper bundled in one subscription if you live in the right area (note that a subscription to the paper only or the website only cost the same).

Print and web options

Conclusion

These methods of selling people paid subscriptions to websites may not be perfect (I think the South Coast could do a lot better with the box that tells you that you need to subscribe to read more).

But they make a much better fist of it than Johnston Press has.

I wrote in an email at the time Johnston launched its paywall that:

The danger is that a badly executed paywall puts Johnston Press and others off. They could do it SO much better, which would surely improve conversions. All a bad trial tells you is that bad trials don't work ...

And that's my conclusion today. A paywall that doesn't sell the benefits of subscription won't sell anything. And that's all that Johnston Press's experiment has proved.

YouTube has a tool - YouTube Insights for Audience - that lets you check out which sorts of people use YouTube, and what sorts of videos particular subsections watch.

So by checking UK viewers only, we can see their age ranges (40% are aged 18-24, for instance) and gender split (72.7% of UK viewers are male).

Age and gender split of UK's YouTube audience

Or you can compare the age distribution of UK male and UK female users.

Age splits of male and female users

You can also see what sorts of videos different users are interested in and what they search for. UK men want Lady Gaga Telephone. The women want the same, just not as much.

The tool doesn't give any clue where the data is from, and some of it conflicts with YouTube's stated demographic data (EG YouTube's factsheet claims its user base is "evenly divided between males and females"). Thanks to Ian Miller for pointing out the tool.

Adam Sherk found that Twitter drives less than 1% of traffic to the US newspaper and magazine sites he asked (but noted that this still makes it "a top 25 referrer for all the sites and top 10 referrers for most").

Twitter as a referrer

Figures for the UK

Trying my luck somewhat, I asked the three largest UK newspaper sites (Telegraph, Daily Mail and Guardian) what their figures were.

• For the Mail, under 0.5% of their referrer traffic is from Twitter.
• For the Telegraph, 0.5% of global traffic and 0.25% of UK traffic currently comes from Twitter.
• For the Guardian, 0.4% of their page impressions in February came from Twitter.

Interpreting these figures

These figures are slightly lower than in the US (but that is where Twitter began) but will have grown strongly over the last 12 months from data I've seen for another newspaper site. I'm sure, as with the US, this puts Twitter in or near the top 10 referrers.

How many people

To put these into perspective, these newspaper sites get around 2 million unique users a day, so 0.5% is 10,000 visitors a day. While not to be sneezed at, it shows that search, news aggregators and direct traffic (RSS/bookmarks) are still more important than Twitter (whatever those of us on Twitter might think).

Twitter.com vs Twitter clients

But please bear in mind that interpreting Twitter traffic is notoriously hard - you can measure traffic that comes from twitter.com (although not always if it's using a URL shortener that frames the target page). Measuring traffic from other desktop clients (tweetdeck etc) is much harder.

@dirktherabbit pointed out recently that less than 20% of people access Twitter through the web. Danny Sullivan had a good rant that addressed many of the problems with people who release Twitter statistics (with a shorter summary of that here).

Looking at the stats for my own blog (comparing direct traffic (which is how Google Analytics records Twitter client traffic) to twitter.com referrals on days when there is a spike in traffic due to some Twitter-driven lynch mob), between 2 and 4 times as much traffic comes from desktop clients as twitter.com.

Thank yous

Thanks to James Bromley, Julian Sambles and Martin Belam for taking the time and being willing to share this data. If you'd like to follow some of their papers' Twitter feeds, try these links:

• The Daily Mail has a main headline Twitter feed at @mailonline. The easiest way to find its other feeds currently is to see who @DMAILsports is following.
• The Telegraph has a list of all Telegraph Twitter feeds. There is also a Twitter list of all these accounts, plus one of all Telegraph journalists on Twitter.
• The Guardian also has a list of the official Guardian Twitter accounts.

Oh, and you could also check out my previous posts on newspapers on twitter.