“A Social Dilemma” Is TUP ecomm a human technology?
AI solutions for eCommerce are a driving force behind conversion rate optimization. With machine learning being a leading method in achieving this goal it leaves to question. Is there a social dilemma in the ecomm world?
When designing an algorithm, you have to determine the goal. At this point, machine learning takes over–analyzing the data, finding the patterns, testing the patterns, and predicting future patterns. The predictions are the likelihood of achieving the predetermined goal.
Basically, machine learning is trial and error at scale.
So, what’s wrong with this? Where are the dangers?
There are two:
1.The goal – making sure it’s transparent and agreed upon.
2.The methods to achieve this goal – an understanding of the consequences of each method.
An easy example of a transparent goal is ecomm. If you go to an ecomm site, you understand that the site is designed to get you to buy something. And, as digital customers, we appreciate this. We like faster page load times, clear paths to desired merchandise, and promotions.
So what methods are leveraged to achieve conversions?
Data collection: IP address (location), source (from where you came), clicks/taps, scrolls, cookies (these can track if you’re a returning user, what other sites you’ve been on, etc.)
Algorithmic predictions: machine learning models take the data collected, compare it to your past data as well as to the data of people who fit your same patterns and change what’s presented to you.
Content and UX/UI changes: “items you might like” and “your personal 10% promo code” are examples of content we’ve grown not only to expect but also to like.
The more subtle changes are perceptual. What colors do we respond best to? What font sizes make it easier to read? Do you pay attention to reviews? Which? Do you click thumbnails or do you prefer to scroll through images?
Machine learning models determine the answers to all of these questions (and many more) by making changes to the ecomm site accordingly and increasing your conversion rate.
Conversion rate optimization plays a vital role in lowering your customer acquisition costs by utilizing the customers you already have now by increasing their value by how much they spend in one session.
What are the consequences of these changes?
At worst, shopping addictions worsen and customers buy things they may not want and return them. So, in this case, the hassle of returning and restocking and the environmental impact of shipping.
At best, customers have a pleasantly efficient shopping experience and come back for both the experience and to purchase more.
So, getting back to the potential dangers: the goal is clearly defined–commerce; the ways in which the site uses the customers’ data and changes the site are not as clearly defined, though many sites have cookie policies that explain how data is used.
We, at TUP, promote clear and available explanations of how data is used, what’s captured and, most of all, accountable measures of its application.