Content Consumption: Grouping content to analyse site activity

I read with some interest an article by Neil Mason on 'Content Consumption'. He highlights two methods of analyse: firstly to use the folder structure of your site, i.e. using '/' within the URL to group content together, and secondly to use tagging to provide this grouping.

Whilst I agree with his conclusions as to the positives and negatives to each approach, he is also missing an alternative, which to to group content dynamically.

For example, WebAbacus uses transforms to create new fields of content in real-time. This allows the user to create a new field containing the content category, sub-category, or full bread-crumb, of the relevant piece of content. You can then report on this transformed field to show the content group rather than the actual URL or Page Title. You can even split the content groups, allowing the end user to drill-down into each category to see underlying content sub-groups or the actual URL/Page Title.

This approach does not require the site to have a rigid content structure or the need to update page tagging, although both methods are still possible.

The main advantage of this approach is that it is applied in real-time, meaning that you can create your content groupings AFTER you've collected the data. In effect, you can start collecting your data on the 1st of the month, and then create your content groupings (a matter of a few minutes work) on the 15th, and still see how this looks for data collected on the 1st-15th.

Additionally, you can make changes to the content groupings, and because they are applied in real-time, the changes will be immediately reflected in your reports.

Finally, different users may wish to have different content groupings. A dynamic approach is ideal here, as it is the end user that controls the content structure, all without any involvement of their technical teams or changes to the page tagging.

The 'Data Group' transform is just one example of the many transforms within WebAbacus. It is extremely useful for understanding content consumption, but also for how users move around your site.

An example of this in use, can be seen via our 3D Traffic Visualisation tool, shown below, where each content group is shown as a circle/node:

Couple of interesting articles related to analytical techniques

Omniture: Visitor Engagement is just a fad!

http://blog.webanalyticsdemystified.com/weblog/2008/07/omniture-visitor-engagement-is-just-a-fad.html

Eric Peterson’s blog countering Omniture’s statement about visitor profiling being a fad. Personally, I agree with Eric on this one. Visitor profiling is a great technique if applied appropriately. It allows you to build a picture of your customers over time, and thus enabling marketers to tailor their communications. Of course this is only relevant if you have the ability to then communicate with each relevant customer segment. The counter to this is that the web/email have allowed for massive economies of scale. By introducing visitor profiling it allows you to create customer segments, which can be targeted with more relevant messages which should, in turn, improve conversion. The real question as the moment is whether it actually has a positive ROI. Creating new creatives/communications for each customer segment is time consuming even if the actual deployment can be automated.

Consumer Attitude towards Behavioral Targeting

http://webanalysis.blogspot.com/2008/07/consumer-attitude-towards-behavioral.html

A brief blog on people’s attitudes towards behaviorally targeted ads. In many ways no real surprise: People want less intrusive and more relevant ads (if any), but do not like the idea of having their behaviour tracked to facilitate this. Personally, I like the idea of being able to rate ads (bit like the face book model), whereby your response will effect what you see. For example, a user could click on an ad and request not to be shown it again, but could be offered the choice of alternative ads or areas of interest. In many ways this would be a win/win situation, as the advertisers would know that people had seen the original ad (even though they chose not to see it again) along with their thoughts on the ad. The users on the other hand get to choose what ads they see in the future, which could even be linked across sites via the Open ID technology (bit like msn passport). The difference here is that the user would be in control, and would be able to manage their own exposure to ads. Advertisers could also benefit as popular ads may be spread by word of mouth, in a fashion similar to that seen on youTube.

So what do you think? Is visitor profiling doomed, or a waste of time? Should behavioural targeting continue or is privacy more important? Is it different to the offline world of store loyalty cards or using a bank card?

Let me know your thoughts?

And we're back...

Before I kick off, an apology. Despite the initial post proclaiming the virtue of blogging and my intention to start, I'm been a tad lack. It's been a year since I was due to start, and this is the first post!

Well hopefully that is about to change.

Firstly, we have just released version 6.0 of WebAbacus (details can be found here: http://www.webabacus.com/version6.php), which has kept us quite busy over the last few months.

Secondly, I attended the eMetrics summit in London a few weeks ago and decided I'd use the opportunity to test out twitter (You can follow us here: http://twitter.com/WebAbacus).

The experience was extremely interesting and has encouraged me to look again at blogging. If you're new to blogging then I'd highly recommend this post from Avinash Kaushik on the benefits of blogging.

However, for me the true reason is that I have really bored my work colleagues and so need a new outlet!

Anyway back to eMetrics...

A couple of things jumped out to me during the two days. Firstly, was the apparent shift towards an increased need to statistical information, rather than simple numbers. Secondly, was the continued need for consistency and stability.

I spoken with many people frustrated at the continued hype rather than actual substance. At the end of the day, web analytics is rarely a persons only function within an organisation - it's a means to an end. There is a constant debate in the industry as to whether web analytics is hard. Personally, I'd say that it is easy with the high tool and the right people, and a nightmare if either one of these are not present.

Tools need to be simple to use but not simplistic! Far to often this mantra turns into a series of interface gimics, rather than something that is actually usable.

Anyway, I hope to make this a regular thing and I hope to make it two way. So...

What do you think? Is web analytics easy?

What insight do you need from your tool?