Friday, December 26, 2008

When Circulation Figures Don't Tell the Whole Story

This PRESS GAZETTE story about web traffic on newspaper websites is interesting, if only to highlight the fact that, so far as I can see, they are still counting their numbers using the "unique visitors" system, as opposed to "absolute unique visitors". This is what caused so much trouble earlier in the year between one blogger on the left and one or two other bloggers, including myself. He accused us of manipulating our readership figures to make it look as if we were getting more readers than we actually were.

I was quoting 350,000 monthly readers (unique visitors), as opposed to the absolute unique visitor figure of around 70,000. The 70,000 figures is the number of individual (or more accurately individual computers) which log onto my site during the course of one month. A better explanation can be found HERE.

According to official figures this week, The Guardian website gets 26 million unique visitors each month. To the uninitiated, that would look like 26 million individuals. But it isn't. At least I believe it isn't. If I visit the Guardian's website thirty times in a month I count as 30 of the 26 million.

So we are still not comparing like with like.

Much of the Guardian's online readership is foreign. This table of UK unique users is interesting...

UK unique users

Guardian.co.uk: 9,548,426 (36.7%)
Telegraph.co.uk: 7,394,866 (32.2%)
Times Online: 7,231,485 (33.4%)
Sun: 6,936,051 (42.2%)
Mail Online: 6,441,804 (30.9%)
Mirror Group: 3,372,104 (57%)

This suggests that The Guardian probably has an Absolute Unique Visitor figure of around 2 million and the Mirror only around 600,000. I base that on my own sites proportions of absolutes to uniques. That may of course be entirely wrong. But if it isn't, it means that individual blogs are now getting as much as 15% of the traffic of a whole national newspaper website like the Mirror's. Not bad for a one man band like Guido or me, is it?

19 comments:

RobertD said...

The difference is probably greater than you credit yourself with. Major bloggers like you and Guido are the destination sites where we go to seek ideas. The MSM blogs are where you refer us to for more background. With a few exceptions I would not start out a session at an MSM site.

Peter Mc said...

Re your Hammers tweet. Yes, winning 5-1 and getting out of the bottom 3...

Backbench Warrior said...

Yes, the fact that you link us to them but the newspapers don't link to you is a big factor worth considering.

cabalamat said...

Instead if saysing "unique visitors" which clewarly implies number of separate people, would it not be better to say "pageviews" if that is what is intended?

E.g. say "the Guardian got 26 M pageviews in Nov 08, of which 9.5 M were from the UK."

That way everyone knows what is meant.

Unsworth said...

But it's only of any importance to The Guardian, I guess. Simply because they feel the need to justify their existence (shareholders, advertisers, proprietors, bonuses etc).

In your case it's a matter of credibility (largely) whereas in theirs it's a matter of money and keeping their jobs.

PaulD said...

So now we know - this is why the Guardian keep Polly Toynbee. There must be thousands of people who visit the Grauniad website just to see what this infuriating woman has to say this time.

Same goes for Monbiot and a few other resident loonies. They are circus clowns.

The only surprise in your league table was the Mirror stats. But I suppose they do show some decent pictures.

Admin said...

As a fan of your blog and a Pompey fan I have decidedly mixed feelings today :P

Ilja Nieuwland said...

The problem with these statistics is that whichever criterion you apply, they're still unreliable. Individual IPs include computers in internet caf├ęs, on university campuses, etc, which may be accessed by more than one individual. Conversely, if your ISP uses dynamic IP addressing your one computer can show up as two or more separate IPs.
We (large science site) stick to 'unique visitors' because it shows individual moments of attention for a site, but also because most other institutions do. And despite all the caveats it's still the least muddied way of comparing visits.

As to influence of your and other blog, I think RobertD's points are well made.

Rossa said...

Well you have 2 readers in this house using the same PC. We don't read the blogs together so do we count as 1 unique visitor or 2? And I don't even know if my PC uses dynamic addressing so we could be more than 2....schizophrenic or what!!

Matt Wardman said...

I'm not going through this again ;-), but the figures *are* de-duplicated. See the Glossary from the Guardian certificate for December, for example:

http://www.abce.org.uk/ABCE_PDFS/Guardian.co.uk1108w.pdf

Any difference is that the abce measure logfiles, while Google Analytics counts readers with Javascript on in their web browsers. Yes, it is a significant difference and the ABCE gives higher numbers.

Matt

Floating Voter said...

Don't fully understand the reasoning behind this post but

- absolute posts are not a good measure of readership because it assumes that the particular computer reads the Blog once a month. We do come back a few times a day even from the same computer!

- those stats seem to indicate that the announcement of the death of the "dead tree press" is somewhat premature.

iain said...

a propos of nothing I note the BBC always quote weekly audience figures, as in 'Terry Wogan has 7.5 million listeners.'

But of course they mean 'in a week' so if you listen every day you apparently become five people.

A lie - pure and simple.

Chris Paul said...

They will have a far different ratio Iain. For reasons including those RobertD cites.

And the figures Guido tends to cite e.g. comparing his GuF with one small segment of the Guardian output are based on apples with apples comparisons.

And so - thanks Matt - are the ones you are citing.

Page views - cabalamat - are of course yet another different thing altogether. If I look at an average of five pages a visit to Iain's and visit every day my page views will be about 150 times as great each month as my unique visitor value.

The greatest muppet in the blogosphere - Nadine Dorries - counts "hits" which is another thing again, counting all the elements of a page as separate hits every time they are reloaded.

If Iain used that and had say a dozen elements each reload the figure for my hits would be around 1800 times my absolute unique visits per month - which are 1.

Happy Christmas!

Tim said...

If you had included a link to the specific item rather than a general link to my site, newcomers might be more likely to discover that you took the figure Google gave for 'visits' and presented it as the figure for 'unique visitors'. Not a general figure or a similar figure or an approximate figure or a previously-thought-reliable figure, but an exact figure for one thing that you presented as the figure for another thing entirely.

Don't tell people I'm wrong about something and then fail to link to the article you're criticising, Iain. It's poor form for any blogger.

The figure provided by Google Analytics for 'absolute unique visitors' is simply a more reliable version of 'unique visitors' than you will get from some other free counters/analysers. The Press Gazette article you link to gives no indication that newspapers are using a free tracking package other than Google Analytics. In fact, it clearly states that they use ABCe as an independent auditor.

(This, compared to spinmeister Paul 'Guido Fawkes' Staines, who rather hopefully claimed that the Google Analytics statistics he offered amounted to an 'independent audit' by Google.)

If you look under 'unique user' here you'll note that several options are available to newspapers wishing to reliably count unique users/visitors. You may also note that ABCe use log file analysis and not JavaScript-reliant tracking. So there are grounds to suspect some limited variation, but none that I can see that would justify assuming a 5-to-1 variation as you do here.

You certainly have no proof of the following; "If I visit the Guardian's website thirty times in a month I count as 30 of the 26 million."

Finally, you now claim that you have 70-90K unique visitors (i.e. readers) a month, but your profile on the Guardian website suggests otherwise:

"Iain is also Britain's leading political blogger with more than 300,000 regular readers a month." - source)

Is this a result of the Guardian's confusion over traffic figures, or yours?

Clive said...

re Iain at 5:32pm

The radio figures for both the BBC and commercial stations have been produced by RAJAR (Radio Joint Audience Research Limited) since 1992. RAJAR is owned by RadioCentre, which is the trade body representing the commercial stations, and the BBC.

According to their website, weekly reach is defined as:

"The number of people aged 15+ who tune to a radio station within at least 1 quarter-hour period over the course of a week. Respondents are instructed to fill in a quarter-hour only if they have listened to the station for at least 5 minutes within that quarter-hour. Between 24.00-06.00, listening is recorded in half-hour periods."

So I think you'll find that your statement...

"But of course they mean 'in a week' so if you listen every day you apparently become five people.

A lie - pure and simple."

... is either a lie itself, a willful misunderstanding or the result of a combination of bigotry against the BBC and an inability to do the most basic of fact checking.

Iain Dale said...

Tim, I have no wish to reopen this again, but your opening para is simply wrong. I did nothing of the sort. I used the figure for Extreme Tracking's "Unique Visitors" - their terminology, not mine - as that is what I understood it to be. I explained it quite clearly at the time. You blew this up to insinuate that I was deliberately inflating my traffic, which I denied then and I deny now.

As to your last paragraph, you seem to think I control what the Guardian writes about me. If only.

I'm sure you have already pointed out the errors of their ways.

Clive, you seem to be very confused. What you have written seems to confirm the point I was making.

Clive said...

Iain, I think you'll find you're the confused person. The important bit is this...

"The number of people aged 15+ who tune to a radio station within at least 1 quarter-hour period over the course of a week."

Number of people. Not number of listening sessions. The "at least" bit should give you a hint. Don't listen to the show, not counted as a listener. Listen for at least 1 15 minute period in the week, counted as a listener.

Now RAJAR does this in a far more methodical manner than the analysis of webserver logs or tracking javascript. It gives real people real surveys to fill in. So there is none of this confusion between "listens" and "listeners" to abuse the internet terminology.

If I fill in the survey (when chosen by the market research company I'm given a weekly listing diary to complete), then if I mark that I listen to Radio 4's Today programme between 8:00 and 8:30am every day, I am still recorded as a single listener in the weekly reach figures. Not as 5 listeners.

Simple really, and not a lie in sight.

Tim said...

Tim, I have no wish to reopen this again,

And yet here we are.

but your opening para is simply wrong.

I assume you mean the part where I link to the relevant/specific post you failed to link to, and say that you took the figure Google gave for 'visits' and presented it as the figure for 'unique visitors'? I based this on data you volunteered, Iain, and there is very little room for confusion. There are no two ways about it; you took the Google statistic for visits and presented it as the figure for 'unique visitors'.

You say:

I did nothing of the sort. I used the figure for Extreme Tracking's "Unique Visitors" - their terminology, not mine - as that is what I understood it to be.

I say:

Quite a coincidence that this 6-figure number for unique visitors from Extreme just happened to be exactly the same as the figure offered by Google Analytics for 'visits'.

You blew this up to insinuate that I was deliberately inflating my traffic, which I denied then and I deny now.

IIRC, your full response to those findings included highlights such as this comment about my suitability as a father and role model, claims that I was a liar, a loon, a professional 'joke', and an idiot, plus repeated assertions that the whole thing was the invention of a delusional obsessive driven by malice, jealously, etc. etc. etc.... so pardon me if I smirk at your concern about my insinuations (that people would have been better placed to judge for themselves had you linked to the relevant/specific post or perhaps even this key follow up).

As to your last paragraph, you seem to think I control what the Guardian writes about me. If only.

I rather do, actually, when the subject of discussion is your author profile at Comment is Free. That the 300K figure is (here we go again) exactly the same as one you have offered in previous versions of your bio furthers my suspicions, as do the passages of text that match the bio on your site word for word in places:

-

"Iain stood as a Conservative candidate at the last election and was Chief of Staff to the Rt Hon David Davis MP in the recent Conservative Party leadership election. He is a former political lobbyist, financial journalist and publisher and was the founder of Politico's Bookstore." (Guardian)

"Iain stood as a Conservative candidate at the last election and was Chief of Staff to the Rt Hon David Davis MP in the recent Conservative Party leadership election. He is a former political lobbyist, financial journalist and publisher and was the founder of Politico's Bookstore." (Iain)

-

I'm sure you have already pointed out the errors of their ways.

Is this assumption of yours the reason why you've failed to alert them to the errant traffic claim until now? If so, then don't delay any further on my account. You can't go around accusing a newspaper of inflating its traffic figures five-fold (accidentally or otherwise) when your bio on that newspaper's website does the same thing (accidentlly or otherwise).

Iain Dale said...

Whatever.