An introductory model on site navigation, time on site and pages per visit
Data. We started recording it when man painted the first rock. It’s a representation of the state of human existence as it relates to his social and physical world, a representation of the mind. Paintings on rocks became the data of future generations study. Now we are in the era where we produce data on a scale which has started to reflect who we are in our entirety.
“From the dawn of civilization until 2003, humankind generated five exabytes of data. Now we produce five exabytes every two days… and the pace is accelerating.”-Eric Schmidt, Executive Chairman, Google
The array of the type of data we produce ranges from social commentary, medical history, genome sequencing, weather systems, to economic indicators and so on. Stored in electronic formats this data becomes the mines from which the skilled can extract and connect information about the human condition and the world. The universe itself is a set of data and information produced from these encodings. The whole universe is a state of computation, its digital physics, a real theory: here’s the link.
Data is reality and it is our story. It reveals secrets about ourselves that we would never have seen if we had not captured it in our electronic systems. It is the key to our future and one of the most important resources of the civilisation we are currently in. We can prove economic theory, social phenomena and even understand biological aspects of ourselves as a collective. It is in data that we find the objective truth, it reflects how thing really are and not how they are perceived to be. Data, in this view, is the ultimate truth!
The new buzz word today is smart data. Basically, modern clients want their data to have PhD’s. This article is not a how-to guide for transforming your large excel files and data dumps into magical unicorns. What this article will try do is show how simple metrics can make data really smart. We will be looking at two interlinked metrics: time on site and pages per visit. These well-known metrics have a vast story to tell when matched with site navigation.
Under our model they requires web properties to be studied individually. One would begin by looking at the architecture and the purpose or end goal of the site in question.
Figure 1
Let’s contextualise. In figure 1, there are three pages before the page on which our desired conversion occurs. Each page has a certain quality and purpose. If it is desirable that users should navigate through these steps or that the users’ information needs are supported through such a structure, then we can assess our question of time and page per visit.
Figure 1 is a way you try to channel users and, it is therefore based in design. If A is the homepage or a portal page, B is the information/ content heavy page, C the sales page (marketing) and D the page which a user takes a desirable action, then each page will have a desired time frame. The less time the user spends on A the better, B should be contextualised by an average reading speed (e.g. 240 words per minute) and C is dependent on how you try sell the users to complete the desired action on page D. This will give you an average time of X. Now we must create a ratio that we will judge what is really happening on the site.
Set X to time on site (through the A-D channel).
Set Y to number of pages in the user’s journey.
Set M as the benchmark, the golden ratio for this specific journey.
Hence:
Say m= Y/X
And Y = 4 pages and X= 8 minutes
Then m = 0.5
In the perfect world our results would look like figure 2
Figure 2
This is usually never the case. When the majority of users fall away from the mean of 0.5 we must study the pages individually (under the context of the whole path). One would then optimise each page individually to create a match with our ratio.
BUT
This is where it becomes interesting and you will be faced with multiple questions. We will not answer these questions here, instead, but create them.
Questions:
1. Which pages don’t reflect the correct time for its context within the path?
2. Why do they differ?
3. Is there an issue on this page or is the page redundant?
4. Is there an additional page in the flow or do people navigate from A to C and skip B (if possible).
5. If answer to 4 is yes, why is this case and what does it tell us about the user? – (you can readjust the ratio to account for the additional page if it is found to support the end goal on D).
Additional questions beyond m.
1. Are users reaching C?
2. Are they reaching D on another path flow?
3. What paths dominate the flows and is this dominance desirable in context of our primary objectives?
Once these questions have been answered we can find new insights. Perhaps users are more interested in other actions on site. This may necessitate up a marketing drive or new area for your site. On the other hand, this may hurt your end goal. You will have to investigate how to optimise your page towards the golden ratio for the desired action on step D.
Conversely, if users take other paths toward D then what does this say about the user? For example, are they skipping B, maybe this indicates they don’t need information to support their decisions. Then this tells us the properties of the psychological state of a user on your site and can give insight into the strength of your brand. If users spend a longer time on B then this will tell a different story. One could go deeper into the study and access where users drop off in this funnel.
By creating a ratio we are able to work towards a benchmark. One is also able to obtain various insights into the user on a page flow bases.
