Lies, Damned Lies...

Whither the page view? (3)

	A little late to the punch with this one, but some interesting numbers were released by Comscore last week which showed that MySpace has beaten Yahoo! in web traffic for the first time. The story behind the story is that Comscore's measure is based upon page views, and the feeling is that Yahoo! has suffered because of its implementation of Ajax across its network.

In related news, a new homepage for Microsoft.com was launched last week which uses more Web 2.0-style technologies (including Ajax) to deliver a smoother user experience. The folks over at MS.com have been working with Webtrends to ensure that the new page has been instrumented correctly. You can read more about this on Eric Butler's blog.

(via E-consultancy)

Indexed

 Felicitously combining two of my favourite things - charts and jokes - Jessica Hagy's indexed delivers a daily visualization that helps you make sense of this confusing world:

Needless to say, my birthday's in September. I suggest you add her to your feed list now.

Hitwise + Alexa = Compete.com

Just as I am starting to despair about what to blog about today, I stumble upon compete.com - a new competitive traffic comparison site offering a hybrid Alexa/Hitwise approach which suddenly seems to have popped up all over the web (I've just seen two ads for it in the space of five minutes).

Reading the compete.com blog, I discover that the company's actually been going for about five years, offering competitive intelligence services on a basis similar to Hitwise to corporate clients. They've just decided to open up their database for public browsing. You can compare the traffic (monthly unique users, rank, page views per visit, visit duration) of up to three sites which are charted competently if unglamorously in their nice, oh-so-web 2.0 interface (is there a mandatory course that website designers have to take these days, entitled "how to make your website look like every other new site on the planet"?)

They commit the unpardonable sin of not providing a mechanism for inserting a particular chart into your blog, so the above shot is a screen capture (courtesy of OneNote 2007's screen clipping feature).

Historically, Compete Inc's approach has been to use ISP data to capture traffic info, but they are in the process of launching a toolbar to gather user info, Alexa-style. The toolbar itself seems to be another monetization strategy in its own right - they plan to use the data they capture from toolbar users for targeted marketing (of course). But really, does the world need another toolbar? How often, as an end user, do you need to check the ranking of a site you're visiting?

The toolbar's only other visible benefit seems to be that it tries to warn users if it thinks the site they are using might be a phishing site (based upon its traffic history info). This is valiant, but since this functionality is in the newly-released IE7, not exactly essential. They will also face a data de-duplicating challenge when trying to knit the toolbar-gathered data together with the ISP data that is their bread and butter.

So will this fly? It's nice to see a competitor to Alexa, and the blog seems thoughtful, though it's the usual run of "How is Zune shaping up against the iPod?" posts that you would expect from a site of this nature. They need to increase their base size and get some international coverage (daily stats would be nice, too) in order to get the scale they need to succeed, though. But thanks for giving me something to write about today.

Google for all your content needs?

My colleague Mark has an interesting post on his blog about a hint from Google that they could re-purpose the AdSense engine to deliver personalized content in general, rather than just advertising content. They'd need to build an explicit preference engine into the system, but if that thought's occurred to me, then you can bet it's occurred to the smarts at the Googleplex. Scary!

Swivel - the YouTube of data?

Should have blogged about this last week, but other demands on my time prevailed.

There's an article on TechCrunch (brought to my attention by my colleague Justin) about the launch of Swivel, whose founders Dmitry Dimov and Brian Mulloy describe as the "YouTube of data". What they mean by this is that they've created a place where users can upload interesting data sets and then plot them against other data sets from other users to look for correlations, such as the interesting one below:

Unfortunately I don't have much particularly interesting data to upload (and the data that I do have that is interesting is confidential), so I wasn't able to try this with some of my own data. Apparently when the site launches, you will be able to upload data and keep it private - though I don't know how many people will be happy to trust their precious data to a relatively unknown third party (not to mention the legal aspects).

If Swivel can overcome this obstacle, however (and they need to - charging for private data is their main revenue source, apparently), then they could be onto something. They're building out significant data center capability to perform correlations behind the scenes and suggest data sets that you might want to compare. But it will be interesting to see whether the correlations they come up with are anything more than just of the 'happy coincidence' variety (for example, the rising plot of oil prices in the chart above could appear to correlate nicely with the usage of World of Warcraft, if you're careful to pick the right range, etc). So perhaps Swivel should have a little tutorial on how correlation does not imply causation on their home page.

The site's other challenge is the cleanliness of the data - even when trying to compare data that was date-based, the site choked several times (doubtless these are problems that the team is working out), but there is a larger issue of 'standardization' of axes or segments. Date is (relatively) easy - you can make some assumptions about the date range that a particular data point relates to - but other ranges/segments are harder, such as:

• Country (problems with old vs new names, regions, etc)
• Age (lots of data is grouped into age ranges, e.g. 16-24, 25-34, but these are not consistent)
• Income (same problem as above, plus currency fluctuations thrown into the mix)

And that's just the axes/segments for humans - other entities like companies have their own characteristics which are not measured in a standard way, especially not internationally.

It'll be interesting to come back to Swivel in a few months when there's some more data in there (and when they have their private data service up and running). I wish them well.

Whither the page view? (2)

I learn via my friends at E-consultancy that ABC Electronic in the UK - who provide traffic auditing services for online publishers - has just ratified a change in its policy on the mandatory measures that must be included in an ABCE audit return. Starting Jan 2007, it will no longer be mandatory to report page views - the new mandatory metric will be unique users. Here's the press release.

This is interesting, particularly in the light of my previous post about what might replace the page view. It represents a big change for ABCE; though I think they're pretty pleased because their job just got easier (a big part of what they did was trying to work out whether people were faking their data to game their page view numbers).

Whilst I think this is the right step for ABCE to have taken, there's undoubtedly more work to be done, because advertisers need information about 'opportunities to see' as well as just total audience. You might have a huge audience, but if they only see one ad each visit, you're not going to have much luck getting a message to them (except as part of some wider cross-site effort).

Webtrends gobbles up Clickshift

A couple of days late on this one, but today's (well, Monday's, technically) market consolidation news is that Webtrends has acquired Clickshift, purveyor of paid search bid management software (as a service, natch).

This acquisition nicely fleshes out Webtrends' solution set, allowing them to compete across the board with Omniture and WebSideStory who both have bid management tools built into their analytics solutions. And now they can become a partner of our adCenter product, too.

You can trust a Scouser...

... to find something to moan about in almost any circumstance.  Apologies to any friends from the North (of England, that is) who may be reading this; if you're offended, I'm afraid I now live too far away to come and do a Boris.

Whither the page view?

There's an amusing graphic on Steve Rubel's Micro Persuasion blog lamenting the demise of the page view (1994 - 2010, says Steve). Steve highlights the fact that, as web applications move beyond the traditional 'page by page' model, utilizing technology like Ajax and Flash, the page view - which, together with the click, is the cornerstone of how all online media is sold - is on its last legs.

The reason for this, if you haven't already read this on a hundred other blogs, is that these new (actually, not new at all, but let's gloss over that) technologies allow content to be refreshed in one part of the screen without the whole page (and its attendant ads) being refreshed. Steve describes this problem as online advertising's "dirty little secret". I don't agree with that - sure, there are a number of competing initiatives to solve the problem, such as the Web Analytics Association's standards committee, and the IAB's version of the same thing - but plenty of noise is being made about the issue.

Steve rightly identifies that the issue is more a business issue than a technology one, though there is a bit of the "we're all doomed!" in his tone. In my opinion, this is a market efficiency issue, and those tend to sort themselves out (with the odd casualty here and there) by themselves without too much bother. But the post got me thinking: what alternatives to the page view exist? Here's a not-at-all-definitive list from the recesses of my own brain:

1. Click events. Unless you're into radically rethinking application design, user interaction will still be facilitated by clicking the mouse. Not all mouse clicks are created equal, of course; so a site couldn't just publish its click numbers - that would be a bit like the old hits metric, since it would be so easy to game the system by creating a click-heavy site (lots of drop-down boxes, radio buttons etc). So you'd need some way of identifying which clicks returned content, and which didn't, which would lead you to a...
2. Content events. Although page views are being broken down into smaller pieces, there is still some intuitive mileage in the idea of the 'content event'; a package of activity where the user requests some new content, and the content is displayed to them. This is a bit like a mini page-view. Content retrieval is Ajax apps is usually done by asking for a piece of XML from the server, and then rendering it (using JavaScript) in the browser. However, the thing that makes Ajax interesting is the fact that content can be retrieved from the server in the background (don't forget, the "A" is for "Asynchronous"), whilst the user's doing something else - for example, the next e-mail in the list is retrieved whilst you're reading the current one. So you couldn't just count the number of XML 'pages' retrieved from the server, because any app could game this by pre-fetching.
3. Time on site. For relatively static parts of an Ajax app interface, the amount of time the user spends interacting with the site can be relevant, because ads could be coded to auto-rotate every 30 seconds or so. There are several challenges with this, however: firstly, it's actually quite hard to determine how long a user spends on a site because when they leave, they just disappear - their final trackable action is the last thing they did before they left, which could be many minutes before they actually left the site. Paradoxically, the serving of timed auto-rotate ads could help here, because if you know that you served 10 ads on a 30-second rotation to a user, they were at the site for at least 9 minutes 30 seconds, and for no longer than 10 minutes. The second challenge is that ads in static parts of a site design tend to be ghettoized by users - that is, they quickly learn where the ads are and ignore them. This is not a new problem - it's why the banner has suffered as an online ad format. Finally, auto-rotating ads are much less well-suited to contextual advertising, since the ad rotation can't take the currently displayed content into account.
4. 'Ad refresh events'. This is a sort of combination of the above three measures - a media owner designs their app and builds in some technology for automatically inserting and rotating ads on some basis, linked to clicks/content events. So to take the example of a mail or feedreader app, display ads might refresh once every three content events, whilst a smaller contextual placement might refresh every time the 'content pane' (however defined) updates. It doesn't actually matter how the media owner does the refreshing, or on what basis - merely that they stick to what they've said they will do, and that this can be measured. Those two last things are, of course, the sticking point - how to make sure the owner of an app behaves honestly? Plus, the media owner will have to be able to publish information about refresh mechanisms for the various types of placement on offer; I guess you'd see something like:

Placement	Ad display events
Home page top banner	1,534,346
Home page contextual	6,324,235
Login page display	3,232,453
App interface banner	768,255
App content page contextual	2,850,235

Now that I've written those out, it's clearer to me what a can of worms this is. It's an irony that, having been the most accurately measurable marketing medium of all time, advances in technology mean that online is likely to get less, not more, measurable in the future.

Have I missed any options? Let me know in the comments box if I have.