IBS.it: how to use AI to lower marketing costs and increase conversions
IBS.it is considered to be the company that pioneered e-commerce in Italy when its first online purchase was paid by credit card. It was also the first to sell - in 2010 – the first e-book. Set up in 1998, it is now an e-commerce business that sells cultural entertainment and leisure time products: Italian and English books, music, videogames, toys and household items. Every day, the company receives 8,000 orders and more than 25,000 products leave its Assago warehouses. In 2018, more than 1 million people made purchases on IBS.it, which had more than 65 million visits.
Why did you opt for AI?
“Online competition has recently increased in a big way and focuses more and more on personalizing supply and anticipating customer needs. It is also fundamental for us to optimize our marketing costs, addressing these only for customers who are really interested in buying. This led to a need to focus on artificial intelligence, with incisive solutions that can make the most of, and support human intelligence through technology."
How does the technology work?
"Datrix developed a Google Cloud Platform solution by using TensorFlow software to analyze our collection of historical data and anticipate customer demand. The predictive module enabled us to identify those consumer segments with more propensity to buy and to quickly increase the conversion process. We now manage to invest in advertising that targets a small group of customers, who have been identified through machine learning algorithms which observe almost 1,000 different behavior patterns that are typical of users who are about to buy. This would have been impossible to analyze without the help of technology. The integration with the various Google products turned out to be quick and easy. When the segments of interest are calculated, they are exported on Google Ads so as to launch personalized campaigns within a few minutes”.
What were the results?
“The most important results were definitely higher sales and lower marketing costs. The results of the project, which took off in July 2019, were very significant after just 3 months: the conversion rate (first-time or returning buyers) increased fivefold, leading to higher sales compared to the manual procedure, meaning without the help of AI”.
Would you recommend it to other entrepreneurs?
“Of course I would, because the excellent conversion rate is not just an arrival point, but the start of a journey. In fact, with machine learning the results are bound to get better because the suggestions proposed by the algorithms will get more accurate over time. In addition, Google Cloud Platform has enabled us to speed up the development of a stable, effective solution at competitive costs”.