It’s always exciting to bring-in and increase traffic to your website via multiple channel marketing campaigns. However, any issues on your website that could affect digital customer experience and create struggle, will impact more customers as traffic volumes ramp-up. Struggle detection is therefore critical.
A small issue lying deep down in the conversion funnel could cause major damage and result in masses of lost revenue opportunities.
Despite this, your figures from revenue, conversions and/or funnel analysis may well tell you the campaign is successful (unless the traffic crashes the website!).
These high-level numbers have a tendency to be misleading while increasing sales figures often cover-up long-standing problems on your website.
This final point is the most important, because, if you don’t know a problem is occurring, then you don’t know that your conversion rate could be doubled!
Struggle detection – Gain deeper insight
While web analytics tools provide a basic understanding of the traffic on a website, digital customer experience (CX) analytics offers the deep understanding of how well your customers are being served in your digital store.
If you’re interested in your customers’ digital CX and want to know about the conversion opportunities you aren’t yet capitalising on, struggle detection offers everything you need to know.
When splitting the traffic on a website based on where each session drops out, you can expect a relative static distribution, thus a consistent conversion rate. For example, the figure below illustrates one way how website traffic can be segmented on a retail website.
Window shoppers are purely browsing on the site, not adding things to the shopping basket. Shoppers that have added things to the basket, but have not started the check-out process can be defined as a potential purchase. These two segments constitute most of the traffic on a retail website. They are the prospects that could come back later on and move down the conversion funnel. Their ratio varies depending on the type of the business and its position within the market.
The next segment ‘struggle purchase’ is the bottleneck of the website that could potentially prevent traffic from converting, resulting in low ROI from your marketing campaign. This segment may even make all your marketing efforts redundant and actually damage the brand reputation because of the struggle they experience at this stage.
The struggle purchase segment can account for 70% to over 200% of the conversion. Given a 2% conversion rate, this crucial segment is between 1.4% and 4% of the overall traffic. It means that struggle detection is ultimately anomaly detection.
While it is difficult to discover this segment, it is ultimately rewarding once it is accurately identified as it opens a ‘rabbit hole’ and exposes all kinds of unexpected issues on a website.
- You may discover the account login does not work, even after customers’ multiple attempts to reset the password.
- You could uncover a hidden error page that clears the whole shopping basket
- You could find out that the coupons sent out during marketing campaigns do not work because of technical issues.
- The submit payment button, the very last step before conversion, becomes a dead button for certain operating systems and browsers.
It is a combination of some of these struggles that could amount to 70% to 200% to your conversion. As serious as these issues are, they can be easily missed, even when you are equipped with various session replay tools. This is because they only account for 1.4% to 4% of the traffic and your analysts simply do not have enough time to painstakingly replay all user sessions.
Without proprietary machine learning algorithms, surfacing these sessions is just like finding needles in a haystack. Anomaly detection is a challenging problem in the machine learning space and standard analytics technology barely touches the surface of the solution.
At UserReplay, we offer the solution of struggle detection by utilizing a combination of advanced analytics and various machine learning techniques on multiple data sources generated during a user journey.
In the following posts, we will detail information on what we have accomplished in solving this and will present our findings.
Stay tuned folks! ?