In this article, Professor Richard McFarland describes how he and his colleagues designed and piloted a method of automated adaptive selling for online retailers, which permits e-commerce sites to use the same influence tactics that salespeople use in face-to-face interactions. Their innovative research recently won an Emerald Literati Award from Emerald Publishing.
Scenario 1: You walk into a store in search of one specific item and found yourself leaving with many more items in hand--the salesperson just seemed to ask the right questions and give helpful answers.
Scenario 2: You’re idly browsing Amazon and it seems like the site is seeing straight into your heart. You find yourself adding much more to your cart based on the products that they ‘recommend for you’ than just the item you set out to buy originally. But how did Amazon know just how to appeal to you??
In scenario 1, you’ve experienced a strategy called adaptive selling, meaning that salespeople adapt their sales strategy to you.. In scenario 2, the retailer has used a customized shopping experience based on your individual browsing habits.
The “art of sales” has been a hot topic for decades: a quick Google search returns over three billion hits for the phrase and there are countless articles, podcasts, and videos dedicated to how one can master selling. Thanks to this interest and to dedicated research, we now know a lot about what makes for a successful salesperson. With the advent of the digital age, however, the art of sales becomes a bit more complicated. Salespeople often use adaptive selling, meaning that they adapt their sales strategy to that specific customer, a process that involves asking questions, reading the customer’s reactions and adjusting accordingly for each customer. Online retailers have been slow to follow suit, though many do already use Web customization to some extent, for example the ‘Suggested products’ you see on the sidebar. Less common is the use of the aforementioned adaptive selling online, marking a missed opportunity for retailers to personalize the shopping experience for each customer in a completely unique way, mirroring the way that salespeople do face-to-face interactions. Decades of research have shown that this kind of selling strategy is highly effective, making it desirable for online retailers as well. By piloting an innovative way to use automated adaptive selling online, Dr. Richard McFarland (ESSEC Business School) and his colleagues Maurits Kaptein (Jheronimus Academy of Data Science, Tilburg University) and Petri Parvinen (FAF University of Helsinki, DIEM Aalto University) show that it is possible to implement automated adaptive selling online and that doing so is more effective than using existing Web customization tactics.
What is adaptive selling?
Adaptive selling is a strategy that salespeople use to adapt their sales strategies and tactics based on the customer. The process is called the ISTEA model (1): impression formation (I), strategy formulation (S), message transmission (T), evaluation of the effect of the sales attempt (E), and subsequent adjustment (A). During impression formation, the salesperson sizes up the customer, taking into account factors like their emotions, their relationship, and their needs and preferences (I). Based on what they glean during that stage, they make a strategy tailored to that person, selecting which influence tactics will be most effective (S). Next, they deliver their message in the way that they have deemed most adapted to that person (T). Then they evaluate its effectiveness, observing the customer’s reaction (E) and adapting their technique if necessary (A).
Salespeople who are better at practicing this adaptive selling process are typically better performers and typically have better customer satisfaction (2). Dr. McFarland and his colleagues focused on three of the six tactics listed in the influence tactic taxonomy (3): social proof, scarcity, and authority. Social proof makes use of the common wish to adhere to social norms, for instance by listing bestsellers. The scarcity principle means that people are often drawn to things they think are rare, so by saying that a product is ‘a limited time offer’ or ‘almost out of stock’, it becomes more desirable. People also tend to listen to authority figures, hence the authority principle: by saying that ‘4 out of 5 dentists recommend this toothpaste’ or having a celebrity endorsement, the product becomes the ‘right’ choice in the eyes of the consumer. It is important to note that while individuals tend to be consistent in which influence tactics they find most relevant and that there are differences across customers In other words, if customer A shows that they are especially receptive to social proof, this will likely be the case in all kinds of different sales settings, but it does not necessarily mean that the social proof will be equally effective for customer B.
By using influence tactics and the ISTEA framework, salespeople can tweak the in-person buying experience in a way they think will be most effective with the customer.
How can adaptive selling be applied online?
Given that adaptive selling is, by nature, quite personal, it might seem a bit counterintuitive to apply it to a more impersonal setting online, where the interpersonal feedback is entirely removed from the equation. This can be seen as an advantage, as the possibility of human error is removed and the process can be automated. It’s an essential step that’s been overlooked by many retailers, despite the increasing popularity of shopping online. Dr. McFarland and his colleagues sought to address this gap. They partnered with three companies to evaluate their automated adaptive selling method, measuring success by comparing the customer click-through rates under different conditions.
To implement adaptive selling online, the researchers built an algorithm based on the ISTEA framework that experimentally determines the best sales tactic to use. In the impression formation step, the algorithm forms an impression of each new customer based on previous customers and uses that to calculate the probability of success of each technique, creating a profile of the customer that is then used on their subsequent visits to the site (I). Then the algorithm automatically experiments with the different sales tactics, balancing between using the sales tactic that has previously proven most effective or an alternative that might be more effective for that customer (S).Then a message based on the selected influence tactic is displayed on the webpage (T). Evaluation is based on the click-through rate, meaning the rate at which customer then clicked on the product, either to view more or to add it to their cart (E). Finally, the algorithm adjusts the customer’s profile based on their click-through behavior (A). The system adapts in real time based on the customer’s behavior, meaning that it’s customized on a customer by customer basis, which is new to the e-commerce landscape.
After first testing the algorithm in simulation studies, the researchers tested it in three field experiments, each with three conditions. In the first condition, customers were exposed to the automated adaptive selling method as described above. In the second condition, customers were not exposed to any kind of influence tactic and were instead simply shown the products. In the third, customers were exposed to influence tactics chosen at random, rather than on the basis of the algorithm’s ISTEA system.
Each field experiment was conducted with an e-commerce retailer. In the first, the retailer was a large-scale company selling bath products. In the second, it was a small company selling lingerie, included in order to test the effectiveness of the algorithm in a smaller-scale site. In these settings, influence tactics took the form of text above a product image: either ‘our choice’ (authority), ‘best seller’ (social proof) or ‘limited availability’ (scarcity), or no label in the neutral (no influence tactic) condition. In the third field experiment, the retailer specialized in flash sales. Dr. McFarland and his colleagues complexified the process: the algorithm used influence tactics in multiple spots on the webpage, showing its sophistication.
The researchers found that using influence tactics is more effective than not using influence tactics, even if these tactics are chosen at random. Further, using an automated adaptive selling process is more effective than using randomly selected influence tactics. This shows that at minimum, retailers should incorporate influence tactics on their websites, and that to really benefit, they should implement a system that will adapt the tactic to the browsing customer.
Where can we go from here?
The success of the automated adaptive selling process piloted by Dr. McFarland and his associates is promising news for online retailers who seek to retain customers and increase profits. It shows that the adaptive selling tactics employed by salespeople in offline interactions can be automated and applied online to great effect. This works in a variety of retail settings, implying that it can be widely applied to a breadth of e-commerce sites. Their findings are particularly meaningful since retailers are already trying to make online selling more personal by using alternative forums like instant messaging and social media platforms, highlighting the importance of creating an experience that mimics the interpersonal interaction you would get in person. It also suggests that a one-size-fits-all method is not the way to go, and that e-commerce sites would be better served by developing a sales toolkit that can be adapted to fit the needs and preferences of their individual customers.
The journey doesn’t end here, with work yet to be done to best optimize the selling experience. Given the requirements of the retail partners, Dr. McFarland and his associates measured success by the click-through rates. The method can be further enhanced if browsing and purchasing behaviors are tracked. Future studies could also include measuring physical behaviors [m3] [JS4] like eye movements and facial expressions. Including the full range of the social influence tactic taxonomy is also vital to pilot their adaptation for online use and to identify which ones work best online.
All in all, this is very promising news for e-commerce retailers who are facing the challenge of how best to profit from a booming online market - and perhaps less promising news for those among us on a budget! The automated adaptive selling algorithm piloted by Dr. McFarland and his colleagues shows that adaptive selling can be done in an online format and that this process can be automated to great effect for both the customer’s experience and the retailers’ bottom line.
1. Weitz, B.A. (1978). Relationship between salesperson performance and understanding of customer decision making. Journal of Marketing Research, 15(4), 501-516.
2. McFarland, Richard G., Goutam N. Challagalla, and Tasadduq A. Shervani (2006), “Influence Tactics for Effective Adaptive Selling,” Journal of Marketing, 70 (October), 103-117
3. Cialdini, R.B. (1993). Influence: Science and Practice. Harper Collins College Publishers: New York, NY.