Ok, so a new round of Times Online website numbers are out, this time courtesy of comScore. As you will have already read, it looks like The Times and its Sunday sister title have seen a 42 per cent fall in unique users between May and July 2010.
However, Beehive City did a bit more digging to get behind the headline data.
Firstly, we should compare July 2010 with July 2009 to paint a clearer picture of what’s going on in Wapping. It is important to compare last month’s figures with a time before the looming spectre of the paywall was a factor that was either (a) attracting intrigued punters to the product, or (b) driving people away in anger at the thought of it.
Year on year, things don’t look quite as bad for Times Online. In July 2009, the website attracted 2.426m unique visitors compared to 1.614 last month, a more modest 33 per cent drop.
That, though, is not the only comparison.
We asked comScore, who told us that in July 2009 the total pages viewed 24 million. The average time on site 6.4 minutes.
In July, the total pages viewed 9 million (a fall of 62.5 per cent), average time on the site 4 minutes (down 37.5 per cent).
An interesting factor that’s been overlooked in the current analysis is the number of pages that people are viewing.
In July 2009 when Times Online was footloose and fancy free, the site attracted an average of 10 page views per visitor. Now in those days, most of the stories contained on Times Online were one page – occasionally they would spill over onto two if they were 700 words plus.
Also, a large number of visitors would come into the site via a Google search for a topic, and land on a story. In fact many never troubled the homepage. So, the stats would suggest at that point an average story read of say eight per visitor, based on the ten page views.
Now, in July 2010, the page view number per visitor has tumbled to six views per page (the actual figure is 5.6). However, even that page view number per visitor is almost certainly flattered.
Why? Well, firstly, you have to log-in to the website now, and our sources suggest that this pop-up screen would count as a page view on the comScore metric. Secondarily, everyone enters through the homepage because none of the content comes up in the search engines.
So even non-payers will rack up two page views per visit, one for the home page and one for the log-in screen. Meanwhile payers are up to two page views – maybe three if the site logs you out and you have to log in again (as has been reported as a consistent bug) – or four if returning to the homepage after logging in also counts as a new page view.
There’s more.
Times Online stories are no longer laid out in the same manner as in the free and easy era. Keen users of the two Times websites will notice that almost every news item is now on two pages, and many are on three or even four.
A quick flick through the other national daily websites and you will see that The Times gives you much less ‘bang’ before you turn to the next page, often you’re clicking page two after less than 300 words.
So to read the same eight stories as the average reader would have done in the free era will generate perhaps twenty page views, or even twenty five. But in truth, the paying customer will almost certainly be a keen reader. Or so you’d hope, reading not just eight stories a month but at least a sixty on average – which is still only two stories a day.
So our paying, one story a day reader, might well generate 180 page views, and certainly 120. All of which would boost the average of six page views per reader up.
Let’s take our average ‘paying I love The Times reader’, generating 120 page views in a month. Then take thirty-nine other non-paying readers who generate an average of three page views a month, because half of them forget you have to pay for The Times or Sunday Times online and visit on two separate occasions.
Together the one payer (120 page views) and the thirty nine non-payers (another 117 pages) produce an average of 5.9 between them – close enough to comScore’s figure of 5.6 rounded up to six.
Now clearly there are lots of ways that you could get to that average, but the essential point is that to produce an average of six views per visitor you will have to assume that the number of paying customers will be relatively low. That’s because the paying customers are generating so many more page views than before – because they have so many pages to click on now.
On this illustration the proportion of paying visitors could be as low as 2.5 per cent, of comScore’s 1.614 million. Which would be 40,000.
That is way, way less than the 1.614 million visitor number (still heralded by people not paying attention as good news). But the 40,000 figure for payers is plausible given that internal rumours suggest that only a few tens of thousands of people have signed up. And, in fact, on the early leaks, 40,000 wouldn’t be too bad a starting point.
The essential point, though, is not to be blinded by the data. ComScore’s 1.614 million figure in July suggests vast numbers of people could have signed up to the Times paywall. But in fact, relatively small numbers of people paying could easily produce the page per viewer average that we see.










You say: “Well, firstly, you have to log-in to the website now, and our sources suggest that this pop-up screen would count as a page view on the comScore metric.”
You don’t have to log-in every time you visit the site (unless you have told your browser to reject cookies, which most people don’t). Log-in is automatic. You can revise your guestimate upwards again.
Small point, if that’s a “recent” pic of Rupert, he’s bought a wig and had a facelift!
The pagination is purely client-side JavaScript, so it wouldn’t count as an extra page view. The full text of the article is within the HTML and just the JavaScript just shows/hides the relevant “page”.
Dan,
Good piece, but a question, which you address only in passing.
You say ‘non-payers could rack up page views’ – if so, this could be the bulk of page views, but could it also contribute towards numbers of absolute unique visitors as well?
Tim Montgomerie at ConservativeHome has really pushed the (£) suffix to indicate that a link goes to a subscription only service. Sadly, it’s not ubiquitous practice yet, so I’ve visited the Times website maybe 30-40 times this months, and then quit as soon as I realise it’s The Times and therefore behind a paywall.
Twitter links, blog links, various other archival searches – do all of these ‘visits’ count, as although the visit is quickly aborted, the user has ‘visited’ the Times website (adding a unique browser and a page view just for the front page).
That would explain the less than fatal drop off in traffic, but the very low page-views-per-visit number, if 80% of ‘visitors’ were seeing the paywall and leaving.
Any clarity greatly appreciated.
Morus
It’s impossible to work out easily. If a three day a week print buyer also buys online (say because they don’t get to the news agent on Sunday) then Times Online is nicely ahead. If a regular five day a week print buyer switches to £2 a week online, then TOL loses out etc etc. You can only know over time, taking print and online subs in aggregate. Until we get an agreed way of measuring online subs, that will take a while – Dan
Thanks Dan – though surely the only way of working out whether the figures are good news or bad news is to relate them in terms of income to what NI would have been making had they not introduced the paywall? Even with a number seemingly as low as 40,000, it might still be an economic success if that subscription income plus whatever rates ads are served to users at (isn’t the theory that with subscriber info to go on, those ads ought to be better targeted and thus NI can charge more for them?) results in a total income greater than the ad income from the free-to-view site.
There may well be another issue here, though, because subscribers to the print edition get automatic access to the website as well (while subscribers to the iPad edition don’t, btw). Some print subscribers will probably not bother using the website but one is presumably fairly safe in assuming that some of those 40,000 are effectively getting their online access free, because it’s bundled with their print subscription.
No easy answer here, but I’ve estimated previously that a print subscriber is worth about 2.5 times an online subscriber in terms of revenue — Dan
Tim/Dan, do you have any figures available that could reveal what this means to the bottom line? The key data is going to be whether 40,000 paying subscribers and 9 million page views/ads served bring in more, less or the same income as 26 million page views/ads served. (Guess that’s impossible to get hard info on unless you’ve got a mole in NewsInt’s ad sales dept.) On the face of it, 40,000 paying customers who previously used to read the site for free sounds pretty good; but only if NI are making more from them than they were making from ads on the free-to-read site.