Conquering Human Bias with AI: Which Idea Reigns Supreme?

Imagine entering the conference room one morning. On your left is a product manager, on your right, a UX designer. At the far end of the conference table is an executive. You’re pitching a new way to test your website designs, based on a trend of peer companies using a similar testing method. The executive, now irritated, claims that this trend is a fad, and promptly ends the meeting. Months later you find that the trend has indeed taken off, and as a result your company website design is less than spectacular compared to those of your peer companies.

These kind of scenarios occur all too often. And it seems that when making decisions about anything–particularly something as detailed as web page design, the decision falls to either the highest paid person’s opinion–in this case the executive or the VP, someone with particular experience on the topic–say a web designer who insists that “we did it this way at my last company”, or a UX designer or marketing manager who based on their own personal preference and artistic “eye”, decides that the website should have certain aesthetic and functional features. While all of these people may have valid ideas on how to improve the website, all of their ideas are more or less biased and based on personal preference or past experiences, rather than scientifically backed data. Hence, when opinion meets opinion in a boardroom conference, it is more than likely that human politics will take over and the person who is “highest on the corporate food chain” will get his or her idea heard first–whether or not that idea will actually improve the conversions or the user experience on the website. With that said, how should companies proceed to make wiser website decisions in this modern, data-infused world?

Data-Informed Decisions

With the widespread adoption of technology in our modern lives, it becomes difficult to gauge all of the nuances of every choice one can make. Luckily, with the mass adoption of technology has come a plethora of data available for analysis.

Artificial Intelligence has revolutionized the way we approach this unprecedented scale of data. With it have come new tools of analysis. For instance, how should you change the design of your website to optimize your conversion rate? We all know that a friendlier layout will attract users and make it easier for them to do what you want. But which idea reigns supreme?

The Importance of User Experience Design

It is useful to consider the basic tenets of user experience design. If you want to make your users do what you want them to do, you need to guide them there with first impressions. A well-designed, clean homepage sets the right tone from the moment they load the page, but the real battle is fought along the full funnel. This is the series of websites a user will have to click through to ultimately convert. For example, if you sell flowers, a user will visit your homepage, then your products page, their cart, and finally the checkout page. Your goal is to ensure that this process is as clear and inviting as possible.

The principle of cognitive user bias says that users expect to see specific navigational elements arranged in certain positions on a website. For example, the navigation bar is typically at the top of the page, and the cart button is usually at the top-right of the page. Disrupting this principle might subconsciously make a user uncomfortable and perhaps not convert. This is not to say that website designers should not be creative, but that the decisions made need to be tested and trustworthy. So how does one decide among website designs to find the one most likely to entice users to convert?

The Evolution of Website Testing

Traditional methods of determining the optimal website layout have consisted of approaches like A/B Testing. This method, while useful, ultimately leads to a roll out of slow, incremental updates to your website. In the meantime, the marketplace environment has evolved, and you may be left behind in the last era of design. In response, tools such as Sentient Ascend have emerged. Sentient Ascend leverages the power of Artificial Intelligence to decide among multiple website designs at once, instead of the A/B Testing limit of two. In doing so, Ascend allows you to quickly optimize your website for conversion.

Companies are using AI-powered solutions for Website Testing

So how well does an Artificial Intelligence-powered approach to website design testing improve your conversion rate? Companies Clicksco and TWEC decided to let the data inform their decision-making around website optimization rather than the usual process of human bias and company politics. Both companies switched from an A/B testing solution to AI-powered Sentient Ascend, where they were able to test a number of different ideas all at once and arrive to a solution that saw the greatest increase in their website conversions.

Online entertainment retail website Trans World Entertainment Corporation (TWEC) saw a boost in its website conversion rate after switching from a traditional A/B testing solution to Sentient Ascend. After a 6-week experiment that tested 77,760 possible website designs, Sentient Ascent found two ideal website designs that dramatically improved conversion rates by 18% and 14.6% respectively.

Clicksco, a global marketing technology firm that offers data, advertisement, and publisher solutions to a number of  different client websites had similar positive results with Ascend. With Ascend, Clicksco tested many different website designs among many of their web properties. On average, they found a 1.75% conversion increase across all the tests they performed, with the conversion rates of some websites jumping by 17.9%.

By optimizing the conversion rates of their websites quickly and efficiently, both TWEC and Clicksco were able to gain a real competitive edge and see an increase in online revenue. Both companies were clever to make websites decisions based on data rather than human bias. Their results were both positive and lucrative.

AI will be the decision maker of the future

So how should decisions be made? Through a combination of human-machine interaction. Humans come up with great ideas, but our biases get in the way of making the ultimate decision. Artificial Intelligence, on the other hand, can determine the best decision by sifting through unbelievably large sets of data, making decisions more precise than humans can. As a result, AI can help us conquer our implicit biases and help us find out which idea truly reigns supreme–not based on opinion, but based on data.