This is Beantin's Typepad Profile.
Join Typepad and start following Beantin's activity
Join Now!
Already a member? Sign In
Recent Activity
Good tip! thanks for that.
Good point Adam. This is very true - as someone who (also) does independent audits and analysis of websites, I've seen the advantage of this at first hand on numerous occasions. (And I'll let you off for the blatent marketing this time..!)
Thanks for that Jesper. You're quite right, I'd missed a comma when updating the "Size" event. The other two events (for Width and Height) were correct. The data GA collects and uses for the "Browser Size" overlay isn't available as a dimension in reports. By storing the viewport metrics in events you can do some additional analysis and report building, even export the data - you're not restricted to the browser size visualisation in the in-page analytics. Have a look at for an example.
A different link in the same email would have the same source (unless you have added specific and different campaign variables to the links). Every click will be counted as the same visit, unless there is more than 30 mins between clicks (or for situations such as the user deletes their cookies between clicks, or changes device -first click on mobile, second on desktop).
Thanks Alex. As I wrote towards the end of the post - Done correctly, in an appropriate context, a carousel is not a bad design pattern at all - but as with everything web, you need to test. It'd be great, if you ever get the chance to test some variations based on what you've suggested, if you could share the results.
Thanks for the comment Fredrik.. The only way to learn what works for your website is to test and analyse. my reference to top-down or bottom-up driven attention in point two is refering to information processing. Our web browsing is generally brain-driven (task, goal, experienced-based) rather than a reaction to a(n unexpected) visual stimuli. Completely agree with you about excluding clutter and daring to focus. I'd be inclinded to agree with you regarding Kappahl if they'd used a single giant image (rather than a slider containing an additional three), and if they'd reduced the height of it just enough to give laptop visitors some scent of content further down.
Thanks for the comment Paul... When I did some testing with the iPad later in the day (you can see a photo from that in the article) I started to "learn" a bit more about positioning and angles - still quite tough. It would be really interesting to do some field testing to see how regular people fair. By all means go ahead and send me some of your AR work. Would be interesting to see and try. /James
The "problem" I refer to above only happens if a visitor *re-visits* your site within 30 minutes. In most situations your 20 people visiting site B via site A will be counted as referrals from site A (this may not be true depending on your report period and how many of those 20 make repeat visits via other sources during the report period)
It's a real shame that some countries have chosen to go for the strict interpretation. It's not a good thing for the internet. I like the fact that here in Sweden they've been pragmatic and sensible - understanding that this directive was suppose to get the abusers rather than impose a usability nightmare and in part destroy potential growth and productivity offered by the internet.
Thanks for pointing out the problem with the link Kristin - an extra " had crept in. It's definitely tricky - another aspect that I haven't covered here (10000 words was long enough!) is the situation with the distributed parts of an organisation's digital presence - a Facebook page, a YouTube channel, etc. Not black and white where a company stands there either. Take the example of this blog. It's hosted by Tumblr, but it's my domain. I can't stop Tumblr creating some cookies. Is there any difference if i'd have kept Awkward.
Hi Mark, thanks for the comment. The IAB Sweden recommendation I've talked about above has been developed in co-operation with the regulatory body here in Sweden, PTS - and they are, at the moment, satisfied with that recommendation. Each country has their own legislation (the Field Fisher Waterhouse report I've linked to above is an especially useful overview of that) and to that extent, I like the way you've integrated geo-targeting into your Cookie Control tool. A suggestion for you is that you offer the possibility to change the language/text of all the elements of Cookie Control (and not just the boiler plate). The Swedish regulation (and others) require the information about the data stored needs to clear and easy to understand - depending on your website, it may be difficult to be sufficiently clear and understandable without using the native language of the visitor.
Hi Colin, thanks for the comment! I had spotted that they'd added opt_noninteraction to the specification for track event; but I haven't got round to updating this blog post yet - but I will! thanks.
Thanks for the link Nikke - I'd missed that on Twitter. I'll have a longer look later, but on first glance, yes, some of what you are seeing could well be down to this bug... some of the other things (time on site and average page views per visitor) are down to mathematical consequences of the changes to visits.
I have an android handset - no problem at all listening to the streams from the mobile version of their site. Even if you couldn't, there would still be the advantage of reading the content about the programmes - after all, this was an advert before they had been broadcast.
Indeed. It's actually quite impressive for them to have so few. But as I pointed out, you have to properly hunt to find the correct page - which means you need to be really commited to liking the page. They aren't providing a sufficient incentment to build up that commitment.
Not as far as their attribution model is concerned, no. It's still a last-touch model.
Any further interaction after the initial page view(such as additional page views, virtual page views and events) make Google Analytics treat the visit as not having bounced. GA can't tell the difference between an automatically fired event and one caused directly by further user interaction. So, if you visit a page on a site with this code, your bounce rate will give you the percentage of people who didn't stay on your site long enough for the page to load.
Almost a blog post in itself Jesper! This wasn't their first attempt at marketing the page. The first thing I spotted was an A4 "poster" sat on their customer service desk, complete with leaflets to pick up and take home. Saved one here: It was actually better than "the thumb"! It at least explained how you could find their page. But still... I think it's a classic example of "old-school marketing" and "ease of web" (combined with lack of governance from ICA themselves). Although I'd buy your explanation too! I do feel a tiny bit bad about venting like this - as it's almost certainly an individual (Lena Jonasson, Information manager) who is responsible - but the point wasn't to make her (personally) feel bad, but to highlight a bigger issue.
Thanks for the feedback Jakob! Yes, Aftonbladet have had a play with QR Codes.. You're right about some publications not having all articles available (freely) online. As much as I'd like to, I decided not to solve the whole revenue model problem for the newspaper industry in this little blog post That said, you could use this concept for customer magazines, free papers such as Metro, flyers...
Thanks Per! I've been using GA on some sites pretty much since Google bought Urchin and launch the free version - that original account/profile set is really muddled. Nothing you can do about it (without losing historical data). One of my favourite practical improvements is being able to create Google Accounts for any email address. This has really helped tidy up the admin side of things - and combined with the access delegation feature, the days of sharing login details should be well gone!
Yes, you're quite right. Events in Google Analytics affect the bounce rate. They basically are classed as "doing something", so you didn't Bounce. Which is fair enough - but in this case i'm using an automatic event, which means no-one bounces (according to Google Analytics). You can read more here: It can be avoided by using a seperate profile for the browser viewport events, but unfortunately there doesn't appear to be a wordpress plugin that supports multiple profiles - which means you'd have to do some manual work to your template. I wanted to keep the plugin as easy to install and set up as possible without any odd/awkward additional requirements. What I will do though is update the description to warn people about the affect on the bounce rate.
In a manner of speaking... Everyone got a chance to watch the recording and explain what they were doing and thinking. But, didn't record it this time though - bit difficult to combine recording with some live analysis for the students! In a more "controlled" test, of course James, I would be doing RTA as standard practice
Great. Let me know when you do; always interested!
Hi Martin, Yes Google have altered their algorithm for calculating the published date of a page. Unfortunately I haven't had the time to investigate and test it. I suggest that you create some tests of your own and experiment. I'd be really interested in knowing your findings if you do... /James.
Zoom events aren't something that desktop or mobile browsers generate - yet. Which makes tracking such things quite difficult - especially across a wide range of browers. It's an interesting metric though. I'll add it to my list of things to investigate when I have chance.