Back in June 2009, I somewhat rashly suggested that newspapers should turn off their RSS feeds as no one was using them (I then recanted slightly). Fortunately this means I have the data for back then on the number of people using Google Reader to subscribe to the three most popular RSS feeds for each UK newspaper.

So here's a table that compares the number of subscribers in Google Reader to each paper's most popular RSS feed 18 months ago and today - the overall fall is 68%.

The papers with large numbers of subscribers have seen significant drop offs. A couple have seen an increase (but from a very low base).

Some caveats:

• The search function in Google Reader can be a bit flaky sometimes, but I think I've found the top feeds in each case (although the feed with the most subscribers isn't always the same now as it was then).
• The number of people using Google Reader will have changed in the intervening 18 months - but I'm not aware of any figures which show that its market share has drastically fallen.
• The Times's numbers have obviously been affected by it going behind a paywall.

If you want to explore the data from June 2009, feel free. If you have any time, maybe you could work out the top three feeds now and compare them?

The debate was started by this post on whether RSS is dying (with a lively debate on Hacker News). TechCrunch reached the same conclusion separately. The truth is probably closest to this thoughtful follow up which distinguishes RSS as a mainstream browser-based user-facing service from a behind-the-scenes format. < Update I think that. I'm not saying RSS is dead. Don't flame me!

There's no right score - it depends on the audience, of course. These are ordered by basic score - the order's fairly surprising (though that may say more about Google's scores ...)

Sun

Overall reading level according to Google: Basic

Sun

FT

Overall reading level according to Google: Intermediate

FT

Guardian

Overall reading level according to Google: Intermediate

Guardian

Daily Mail

Overall reading level according to Google: Intermediate

Daily Mail

Mirror

Overall reading level according to Google: Basic

Mirror

Times

Overall reading level according to Google: Intermediate

The Times

Express

Overall reading level according to Google: Basic (which makes little sense given the graph ...)

Express

Telegraph

Overall reading level according to Google: Intermediate

Telegraph

Independent

Overall reading level according to Google: Intermediate

Independent

So if you ever want to know what a London SEO consultant does. It's this.

• July to August growth: 17%
• August to September growth: 17%
• September to October growth: 13.1%.
• October to November growth: 8.3%
• November to December growth: 6.6%

The detail

These accounts had 1,801,044 followers on November 2nd (ignoring one FT account that has been shut). On December 2nd they had 1,919,770 followers in total.

Of the 118,726 increase, 76,812 or 65% was for the @guardiantech account (which benefits from being on Twitter's suggested user list).

As ever, the full spreadsheet is here or you can see the iframe below.

• July to August growth: 17%
• August to September growth: 17%
• September to October growth: 13.1%.
• October to November growth: 8.3%

The detail

These 130 accounts had 1,801,811 followers on November 2nd, up by 137,568 from 1,664,243 on October 1. Of that increase, 95,007 (or 69%) was for the @guardiantech account (which benefits from being on Twitter's suggested user list).

(NB the Telegraph seems to have deleted its badly spelled @TelegraphScienc account, so I've restated October's figures to be for 130 accounts, rather than the 131 I used to track).

The biggest mover was @MirrorFootball, up 11 places to 81st (from 455 to 809 followers), suggesting the Mirror is finally making some use of Twitter (most of its other accounts are near the bottom - and only appear to have moved up a place due to the demise of the Telegraph's Science account).

As ever, the full spreadsheet is here or you can see the iframe below.

The rate of growth has slowed, however. This is a monthly increase of 13.1%, compared with 17% from August 1 to September 1, and also from July 1 to August 1.

What's more, 151,555 of the increase (or 78% of the total) is down to just one account - that of @guardiantech (which owes its popularity to its place on the Twitter Suggested User List). Indeed, of the 131 accounts I'm tracking, 51 have fewer followers than me! You can see the full table here, or below.

You can see the September figures in the table below (or here).

I have more Twitter statistics here.

There are 10,914 Guardian URLs bookmarked, with the Times in 2nd (3,944) and the Independent 3rd (3,196).

Quarkbase doesn't reveal how the figure is calculated. It must use the Delicious API somehow... Click the papers' name to see more Quarkbase data... And if you're reading via RSS, click here to see the video.

There are more than 100,000 citations on Wikipedia that reference the Guardian - more than twice as much as any other UK national newspaper. The FT is way down the list - barely higher than the Sun.

Quarkbase uses a Yahoo search (site:wikipedia.org linkdomain:NEWSPAPER.co.uk) to count the links. Click on the paper's name to see the Quarkbase data.

in terms of daily page views per user, the Sun (4 pages), Guardian (3.1) and Telegraph (2.9) are above average.  Visitors to the Mail site look at just 2.4 pages a day - so while the Mail may have come top in the July ABCe figures, maybe its large number of overseas visitors aren't staying to look round the site.

Stickiness of UK newspaper sites

• Better than average figures are in bold.
• The bounce rate is the percentage of visits that consisted of just one page (so a low number is good).
• These figures are 3-month averages. These change on a daily basis at Alexa - so they may have altered slightly by the time you check. Click the papers' names to see the current data.
• The overall average at the bottom is a simple average - it has not been weighted by traffic.

Page views vs bounce rate

The table is ranked by daily page views per user. The bounce rate is another measure of stickiness. It doesn't exactly correlate with page views, as papers may have differing proportions of loyal, engaged users who visit lots of pages. The more pages that these users visit, the better the page view figure - but they won't affect the bounce rate.

The Telegraph has a worse bounce rate than the sites near it in the table, perhaps because the great success with its Digg tool doesn't always lead to multi-page visits?

Using Alexa data

There are issues with using Alexa data like this as it underrepresents UK users, who may have differing usage patterns to other visitors. However, as it seems to underrepresent them more or less equally, the rankings should be OK even if the absolute figures are all out by the same margin.

Oh, and all the papers are doing better than me! Visitors to this blog look at 1.5 pages a day and I have a 76% bounce rate (which gets a bit worse when I publish things like Tweets people pray their bosses doesn't see).

• Overall, Google drives the most US traffic - 19.1% of the total traffic to the 9 UK national newspaper sites.
• News aggregator Drudgereport.com is the 2nd highest source of US traffic, accounting for 12.7% of the total.
• Next is Yahoo at 5.2%, with Facebook 4th at 1.6% (see methodology at the end).

Here are the US traffic numbers for each site, with the top 5 referring sites. Update: The Times says is doesn't recognise the OptMD figure and has given me its internal data, which is below.

FT: 5,960,589 unique USA visitors in June

• drudgereport.com 15.53%
• yahoo.com 11.05%
• google.com 8.37%
• wsj.com 2.29%
• bloomberg.com 2.23%

Daily Mail: 5,199,078

• google.com 22.44%
• drudgereport.com 15.94%
• yahoo.com 4.70%
• facebook.com 3.14%
• digg.com 1.95%

Daily Telegraph: 4,087,769

• google.com 19.00%
• digg.com 11.48%
• drudgereport.com 9.13%
• yahoo.com 3.42%
• facebook.com 1.88%

Guardian: 3,676,498

• google.com 25.11%
• drudgereport.com 5.73%
• yahoo.com 3.23%
• facebook.com 3.13%
• sofacinema.co.uk 1.74%

I'm not quite sure why sofacinema is sending so many americans to the guardian (the guardian owns it - but it's a UK service I thought)

Times: 2,805,815

• google.com 21.02%
• drudgereport.com 13.59%
• optmd.com 3.08%
• yahoo.com 2.94%
• evri.com 2.26%

OPTMD is a pop-under advertising seller - so the Times is paying for its 3rd most popular source of traffic.

The Times says it doesn't recognise the OptMD number and sent me this tweet: "Internal figs for Times Online, last 30 days, US visits only: Google.com 26% Drudge 16% Foxnews 3% Digg 3% Yahoo 2%"

Evri.com is a search engine with a widget - the Times has it on its site.

The Sun: 2,419,319

• drudgereport.com 19.48%
• google.com 19.11%
• adsnewsint.com 5.71%
• yahoo.com 2.95%
• facebook.com 2.17%

Adsnewsint.com looks like a News International ad network - again, the Sun appears to be buying traffic off other NI companies.

Independent: 1,317,298

• google.com 22.80%
• drudgereport.com 11.66%
• websitehome.co.uk 2.80%
• yahoo.com 2.66%
• reddit.com 1.88%

Again, a little odd that a UK site building site is sending so much US traffic ...

The Mirror: 748,098

• google.com 36.36%
• accelerator-media.com 5.21%
• yahoo.com 3.65%
• google.co.uk 3.18%
• bbc.co.uk 2.58%

Express: 63,216

• google.com 11.23%
• accelerator-media.com 10.34%
• bbc.co.uk 6.09%
• sublimemedia.net 5.53%
• dailyexpress.co.uk 3.70%

Even with two advertising networks in there, the Express still doesn't get much traffic.

Methodology

The aggregated %s at the top (ie amount of traffic from Google, Facebook etc) will understate any sites that don't appear in the top 5 for some papers.

This mainly applies to Facebook, which isn't a top-5 referrer for the FT. However, this will only add a couple of % at best to its overall figure (IF facebook were at #6 for FT, it could account for, at most, 2.22% to the FT's traffic, which in turn is about a fifth of the total, so the Facebook figure might be understated by 0.5%.)

Drudgereport, Yahoo and Google are in the top 5 for the main papers, so the aggregate figures for these sites are accurate.

While the Compete data includes referring sites and search, it doesn't seem to include direct traffic, so bookmarks etc are presumably not counted in its totals.