According to the Web Analytics Association, web analytics is: "The measurement, collection, analysis and reporting of internet data for the purpose of understanding and optimising web usage."
That provides an intellectual understanding of web analytics, but does not portray what it encompasses on a day-to-day basis. There are processes, tools and hypotheses to consider when crafting an effective analytics approach.
In the end, web analytics must answer basic marketing questions, such as change in sales and long-term brand-building. It is easy to forget that the web, like television and outdoor media, is another key marketing channel.
In the twenty-first century, the digital landscape plays an important part in marketing, but it is also becoming very confusing. All around the world, we see the trend of increasing numbers of people going online and interacting at a deeper level. Add mobile to the mix and it is even easier to get bogged down by technology and numbers.
So it is useful to take a step back and understand the practical side of web analytics before this incoming tidal wave hits us.
In the beginning, log files recorded hits
It is said that the web -- and, by extension, the entire digital channel -- is the most measurable medium due to its inherent strengths. And marketers like numbers. Even today, we still talk about circulation and ratings.
The first web metric that people think of is the "hit". It refers to each and every request to the web server for content. Marketers liked hits, because it provided a nice round number that represented reach; the larger the number, the louder they could boast about their websites' usage.
However, hits have become meaningless in today's multimedia web. If a single page contains three images, that's four hits for that page, but really only one "page view". And then, marketers started to think in terms of individual people, rather than just the amount of content that was served up. That brought about greater awareness of the "visit" metric.
But the original hypothesis remained: the larger my site traffic, the louder I could boast about it. It paints a nice picture of how well and often your website is being visited.
Unfortunately, very little useful information can be gleaned from site traffic. For example, though a television show could have 100,000 viewers, what marketers really want to know is whether those viewers watched the advertisements. However, the web is an interactive medium. And since it is supposed to be the most measurable channel, marketers demanded more insights.
This request -- understanding site activity rather than mere reach -- pushed web analytics forward into a stage that made it more useful and relevant.
Every conversion begins with a single click
Anyone who has used a website generally moves from page to page. To most of us, we just think of it as consuming a series of content in a seemingly random manner.
But to the web analyst, those actions have produced a journey, or "clickstream". Your interaction within that website was like a driving a car from place to place. You use the route that makes the most sense to you.
To refine the idea, some websites make use of "ideal journeys", or clickstreams that are key to measuring the sites' effectiveness. Marketers want visitors to use such journeys, so they want to ensure that there are as few drop-outs at each stage of those paths. A common example is a registration process.
However, there has generally been decreasing relevance of clickstreams, let alone ideal journeys. For example, increasing end user usage of search engines has resulted in visitors finding your content without following the prescribed journeys. Does that mean your journeys aren't working?
Savvy marketers also recognise that a website does not exist in isolation. Visitors arrive through advertisements or remembering the addresses from a billboard. These "sources of traffic" also affect usage of the journey. Going back to the driving analogy, there may be a single highway, but there are multiple roads leading into it.
Thus, it is more relevant to identify areas of the website that serve two purposes: they are relevant to visitors, and they meet the site's objectives. Ideally, these two purposes should be measured with the same events. Such an analysis moves reports away from clickstream and towards "conversions".
Think of conversions as "landmarks". A registration process may involve four pages, but what you want to know is the share of visits between the first and last pages. If it doesn't meet your target, then you can analyse the steps in between to find out what's wrong.
And now, you should have a nagging little question at the back of your head: "How do I identify the 'landmarks' in my website?" This will be answered in Part 2. We're just getting started, so hang in there!
Balasingam-Chow Yu Hui is a web analyst at XM Asia Pacific.