Retailers can’t always rely on customer feedback to measure the success of their ecommerce platforms. Small technical issues, which often go unnoticed by the business itself, can significantly hinder the customer experience.
To combat these concerns, companies are turning to emerging technologies that analyze customer interactions and identify pain points.
According to a May 2019 study conducted by Isobar, 45% of CMOs worldwide planned to invest in data and analytics in the next year for customer experience purposes, while 40% of respondents said the same about digital marketing software.
More than a quarter (29%) of respondents said they use artificial intelligence (AI) to help streamline customer experiences, per Isobar. An additional 46% anticipate using AI in the future.
Similarly, 36% of respondents said they’re currently using machine learning, with an additional 37% who anticipate using it in the future.
To better understand how these technologies are being leveraged, we spoke with Mario Ciabarra, founder and CEO of digital intelligence platform Quantum Metric.
What are some common flaws in the ecommerce customer experience?
The biggest flaws that go unnoticed are the ones that are micro-affecting conversion rates. It might only be about a 2% impact on your business—and your business typically fluctuates 5% to 10% on any given day—so these microbugs on your site can be hard to find.
What are some examples of microbugs?
When cookies are too long or there are too many of them, we see a drop in conversion. If you log in and the cookie that represents your login gets evicted—some browsers haphazardly evict a cookie when there’s too many to fit—that user would no longer be logged in, or their cart would be cleared.
That’s a really frustrating feeling, and this was happening with major retailers worldwide. We found about 17% of our customer base had that same problem and had no idea why it was happening.
What other kinds of bugs are common?
We saw CSS bugs multiple times across different retailers. For example, if a customer tried to pick a color of an article [of clothing] and buy it on a retailer’s website, it worked great on desktop. However, on mobile, the swatches of color weren’t available. So you’d click ‘add to cart,’ and it never added the item. That was a multimillion-dollar issue of people falling out of the funnel on mobile because there was no way to buy that product.
Can this affect payments?
About 25% of the issues that we have found across retailers we work with have been payment integration. Sometimes PayPal can be plug and play, and it just works; but in certain cases it doesn’t, and it’s frustrating.
How has machine learning helped identify these issues?
Machine intelligence is sort of like having an army of analysts going through the problems and asking, ‘Is this a real problem or not?’ We’ve also built a data science platform because we know that [data analytics] won’t solve every problem, but it’s an easy way to start showing off the value of discovery and identifying and quantifying where issues are. That’s the easiest place to start. It’s come about because we simply didn’t have access to this power of database query in the past.
Looking forward, how big a part do you think automation and machine learning can play in this space?
When you think about a retailer in-store, it can actually have someone watch a customer try to reach items on the top shelf and struggle and then move those hot items from the top to the bottom shelf. But with 5 million people on your website, you can’t have one-to-one combat.
Getting that automated with machine learning is what I’m excited about. It’s going to continue to transform. When [ecommerce sites] have a limited amount of resources, using computers will amplify how to deliver the best digital experiences with ease.