Online purchase

Learn about the algorithms when planning your next online purchase

A good salesperson is a kind of resourcefulness. It’s about quickly assessing your customers and presenting your products in terms that resonate with their unspoken needs and wants. As artificial intelligence and machine learning increasingly intersect with e-commerce, these invaluable human skills are finding algorithmic analogues, not just at the point of sale, but throughout the customer journey.

The results will be familiar to online shoppers around the world. Netflix and Amazon algorithms use data from every customer click to refine their recommendations and drive consumption. Tech giants are also deploying consumer activity data to fine-tune their email and social media marketing and more. This is probably just the beginning, as the technology’s predictive prowess continues to improve.

For major online travel agencies (OTAs), such as Expedia Group and, algorithm-based personalization is a growing priority. Digital natives increasingly want a travel booking experience that maximizes both personalization and convenience. In addition, the balance of power between OTAs and inventory providers could shift in favor of the latter, due to trends in greater consolidation in the hospitality industry. Faced with the prospect of declining commissions as deals are renegotiated, OTAs are rushing to offer increasingly accurate projections of individual travellers’ willingness to pay.

Jack Chua, director of data science at the travel site hot wire (owned by Expedia Group), exemplifies the talent driving the personalization of online marketing through AI. The margins of AI All in Dubai earlier this year, Chua and I discussed the profound strategic implications of this technological revolution.

AI and economic theory

Chua previously worked on pricing models at Amazon. Estimating price elasticities for a global online marketplace that includes millions of products can be a complicated project. The purchasing habits of certain products can defy basic economic intuition, for example when the price falls, demand sometimes falls instead of rising. This is usually due to variables that affect demand and are correlated with price but not observed in the data. These variables can be unstructured and user-generated, such as customer reviews or product photos.

For Chua and his team at Amazon, making the model work with, not against, the grain of economic intuition required overlaying an “optimization layer, or some sort of regularization function” that ensured that their model machine learning was consistent with economic theory. One of my recent articles, “Ideas Markets: Prize Structure, Entry Limits, and Design of Ideation Competitionsfollows a similar paradigm. It combines decades of economic research on game theory with data from large-scale crowdsourcing competitions to estimate the parameters guiding participant behavior and provide a framework for optimizing incentives.

Chua advises leaders to understand the economic structure of the problem when applying AI to business. Instead of just plugging the data into a general purpose machine learning algorithm, they should recognize how additional information such as text and images can also contribute.

Three tips for deploying AI

According to Chua, there are three pillars for the successful implementation of algorithmic business solutions.

First, unsurprisingly, you need skilled quants on board. “You have to hire the right experts to be able to develop that capability,” he said. “Think of data scientists, statisticians, economists, even business leaders and operators to run the program. You have to have that staff in place.

Second, hard skills must be complemented by an appropriate focus on soft skills. Cultivating a culture that understands and appreciates what AI and machine learning can bring to the organization is essential. Chua lamented that too often he hears managers say things like, “Data science is just basic analytics.” “When leaders think that data science is not much more than a simple SQL query, I think it leads to a lack of thought and determination of the real impact of data science on various parts of the value chain,” Chua said.

Third, Chua urges business leaders to remember that data science and AI are essentially “optimization technology” that works best at scale. Companies should progressively implement data science solutions to strengthen and amplify the processes that are already creating value. “If you’re a startup and you don’t yet have a product that generates millions of dollars in sales, it’s harder to justify building that 20-person data science team that’s going to cost you 10 million dollars a year,” he said.

Identify opportunities

Right now, Chua sees plenty of opportunities for a growing number of start-ups trying to “democratize AI, put it in the hands of commercial operators with no domain knowledge” instead of going the route. more familiar with selling AI as a service. .

Many companies have been working on AI improvements for commonly used equipment that have looked the same for decades. Chua mentioned the example of weeding robots with computer vision systems that can distinguish weeds from crops as they roll across fields. The machines would use up to 20 times less herbicide than standard methods which typically involve covering entire fields with chemicals linked to negative health outcomes. The logical extension of this innovation would be to integrate it into machines already used by farmers, such as tractors.

“Use cases like where we’re merging the high-tech space of AI with brick-and-mortar or traditionally non-AI domains are a really exciting next step,” Chua said.

Pavel Kireyev is Assistant Professor of Marketing at INSEAD.

Did you find this article helpful? Subscribe to our weekly newsletter.

Follow INSEAD’s knowledge of Twitter and Facebook.