How Learning Where Shoppers Browse Online Helps Businesses Make the Sale
I’m a die-hard fan of travel websites. Whenever I have a trip to plan I immediately look for the best-deals on sites like Expedia and TripAdvisor. More often than not, I browse the sites just to satisfy my curiosity with no intention of booking a trip. I check out airfare to places I’d like to go and marvel at how inexpensive it is to travel at certain times of the year.
I am not the only one. Deal-rich travel sites attract a lot of visitors but only 1 percent of users actually purchase a plane ticket or hotel after browsing.
Many people (myself included) just like to look at the bargains and store the information away for later use. The challenge a lot of sites face is how to guide users to their ideal product as quickly as possible, while leaving the browsers to be flooded with advertisements, increasing the site’s profit.
Bryan Balin works for SmartTravel, a subsidiary of TripAdvisor, and has helped solved this problem for travel sites. SmartTravel now uses an algorithm that creates a picture of the user and the probability of her making a purchase. The data used to determine their profile includes click speed, time of day, location and number of previous visits to the site. The algorithm takes the information collected and calculates the revenue gains and possible risks of showing the user advertisements. The entire site is able to adjust itself in milliseconds to better suit the user.
The algorithm used at SmartTravel is an example of “machine learning,” a new approach to business economics. Arthur Samuel set the concept of computers teaching themselves in motion back in 1959. A machine repeats a single task as instructed by a programmer, measuring its success rate on some scale. The computer then changes its process to measure whether it does better or worse each time. These actions create a loop until the computer “learns” how to effectively complete the task.
One fewer thing to worry about.
hiQ Labs has predicted machine learning will transform business. Companies will be able to better woo senior staff members by using the concept to collect data on variables like pay scales, organogram and job titles. It can compare this information against data from other companies and determine who might defect and why. Managers can narrow down who might be considering leaving their organization and create a tailored solution that benefits them, instead of a generic blanket solution that covers all staff in an attempt to keep one or two people from leaving.
Scott Nicholson, a Stanford economics PhD and advisor to hiQ Labs, has commented that in the new world of economics it can be difficult work presuming who are the buyers and who are the browsers.
Platforms are trying to learn more about customers to provide them with better links, better services and better information. Poynt, a cashless payment terminal, has begun finding ways to connect store owners and customers, as well as shops themselves. Store tills are being linked together so owners can see how their sales (anonymously) compare to local competitors, allowing them to make specific changes to boost sales and profit.
A personalized online experience.
By singling out each site visitor as an individual, the user’s needs are tailored to and met, creating more accurate results and increase sales for the company. As a frequent shopper, having ads and content designed for me within seconds makes me feel more confident in my purchasing decisions.
Economics is trying to become more personalized in the new business world, and could be setting new standards for future generations.
Source: How Learning Where Shoppers Browse Online Helps Businesses Make the Sale
Via: Google Alert for ML