Can Data Analytics help reduce user interaction cost?
Zero interaction cost is perceived as the holy grail of usability as a field. To minimize interaction cost, designers strive to reduce the reading, scrolling, clicking, typing, information comprehension time and even the technical nuisances of page loads.
Unfortunately, all these things are very hard to completely avoid and that is why zero interaction cost is rarely attainable. Even after minimizing all the hurdles between a user and their goal in our systems, there is still the chance that the user might get distracted or get his/her attention switched to another part of the UI — maybe a popup window or a different page altogether; eventually increasing the interaction cost.
But why is interaction cost so important? Why do designers strive to attain as little interaction cost as possible, even if it means monitoring and taking into account how users interact with every single detail of a user interface?
The concept of interaction cost was actually introduced back in the early days of Human Computer Interaction (HCI) to evaluate the usability of software systems. All usability heuristics (rules for interaction design) minimize the interaction cost for the user, and this makes interaction cost inversely proportional to usability. Minimizing it gives a better chance of success in usability.
A common behavior that not only results in higher interaction costs but also increases bounce rates is pogo-sticking — A sequence that starts off with the user at the search engine results page, or a website’s landing page, then follows the user deep inside as site’s hierarchy and finally concludes with the user back to where they started — dissatisfied and most likely frustrated too. More times than I can count, irregardless of having experience designing and developing applications for the worldwide web, this is surprisingly something that has happened to me.
Through analytics, two metrics can be used to determine whether our systems are suffering from repeated user actions, such as pogo-sticking. One way is by breaking down an application into pages and comparing:
- The number of visits per page (how many times a particular page is visited)
- The number of visitors per page (how many people visited a particular page)
If a page is visited, say 5000 times by 1000 people, in a timeframe short enough to agree that each user had been on that particular website for just one instantaneous session, then on average each person viewed the same page 5 times. Keeping usability in mind, we can safely assume that the additional 4 views per person were frustrating.
To get a better picture of how users interact with a page, and with the pages it links to, besides just narrowing data to the session level and counting clicks, we have to exclude paths where the user visited several pages before returning to the page in focus, or when the user spent substantial time engaging with the destination pages.
Identifying the specific pages where this comparison metric is indicating a high number of repeated user actions is the first step. This would require interaction designers to prioritize which pages to tackle first — based on the overall traffic of the website, since conducting navigation path analysis is not a cakewalk.
Of course, just this ratio of page visits to page visitors is not enough to conclude that our designs are bad and something has to be done about them immediately before any more harm is done. Interestingly, a similar behavior that resembles that of pogo-sticking, could be an indication of a positive user experience. Consider an example where a user on a blog page clicks on 5 of the article headlines, each time being directed to the article content. The user reads the article, goes back and clicks another, since they found the stories interesting. In the same manner, inversely, a different user could click the same 5 headlines, bouncing back and forth between those articles, disappointed, since they found nothing interesting enough to read. Quantitative data from analytics has its limits and cannot explain each and every behavior of a user. Solely, it cannot answer why users choose to do what they do.
Qualitative research is key. Performing user experience research and conducting usability testing are necessary in order to fully determine user behavior. Aside from this, quantitative data-driven decisions are very attractive, and getting attention from business executives and those who hold the purse-strings can be much easier when you’ve got quantifiable data to support your findings and recommendations.
Incorporation of data into the design process is only a complement, and a powerful one if done properly, for the existing qualitative research methods available. Gathering the data from analytics through triangulation, then proceeding to test and verify our findings is a certain way of attaining accurate information when conducting usability research. It maps out a whole new, exciting, complex and multi-perspective view of user experience design that we can learn to appreciate and apply.
References and more:
Read more about pogo-sticking: here
Learn how to apply custom sequence segments on Google Analytics: here
