French advertising technology company Criteo has unveiled plans to invest €20m in AI research over the next three years. Focused almost exclusively on AI for online computational advertising, the investment is to be spread between AI talent acquisition, primarily in Paris, infrastructure development, and collaboration with academia. At an event held in Paris on June 27, returning CEO Jean Baptiste Rudelle also announced the launch of Criteo’s AI Lab in Paris and Grenoble in France and Palo Alto in California. The goals of Criteo’s AI Lab are as follows:
understand consumer behavior and context better
accelerate the AI research cycle from concept to testing to product development
foster AI research innovation via partnerships with academia
create a pool of AI talent in Paris.
Making AI more practical is at the core of Criteo’s AI Lab
Engineers and researchers with AI experience are in high demand with many large companies including Facebook and Google recruiting at a global level, including Paris. The company has been using machine learning for its recommendation engine for 13 years and therefore already has strong expertise and assets in AI. Criteo’s AI Lab is set to grow from 67 to approximately 150 employees in the next three years. While significantly smaller in size than the large tech firm, the lab is very focused on practical applications of AI and is set to merge teams of engineers, product developers, and researchers on joint projects. This should generate some appeal to potential European applicants, but whether this will deter them from working for Facebook or Google remains to be seen.
Criteo aims to alleviate media companies’ dependence on GAFA (Google, Apple, Facebook and Amazon). The technology developed through this AI program will help create better targeted advertising to the 1.4 billion internet users it reaches – smaller than Facebook’s 2.2 billion monthly active users in 1Q18, but the largest open omnichannel platform available. Criteo saw $615bn of e-commerce transactions in 2017 and serves over 75 billion ads every month, driving $29bn of post-click sales annually. Criteo’s client base is very large and influential and the company has become a very strong alternative to the ad networks dominated by American tech giants such as Google and Facebook. The results and/or products that will come out of the company’s AI Lab could have a disruptive impact on the markets in which its clients operate. Making the company’s new ad products more effective could provide significant growth in advertising revenues for internet publishers, and more particularly for retail companies.
One example that Criteo’s AI Lab has been working on is the creation of a unified product catalog across multiple retailers. Criteo uses its own image- and text-recognition AI engines to uniquely identify products on retailers’ websites. This will, for instance, help to avoid advertising products to customers who have already bought them, which can be annoying and may lead to ad blocking.
As explained by CEO Jean Baptiste Rudelle, the lab offers the company the ability to make a long-term investment on AI from a business sense while accelerating AI research for more immediate practical results. Think of it as the equivalent of a “turbo engine” for focused AI development. This is something that not many companies the size of Criteo (2,800 employees, €2bn revenues) can afford to do – a testament of both the importance of AI in the ad tech sector and the audacity of Criteo’s management team. The difficulty for Jean Baptiste and his team will be to effectively select and implement the AI concepts that the lab will generate within the three-year time frame they have set. This is not something that has been done before and, if it proves successful, it should make them more competitive with the likes of Google and Facebook.
2018 Trends to Watch: Media & AI, ME0002-000796 (September 2017)
Digital Advertising Forecast Report: 2017–22, ME0002-000788 (August 2017)
Ronan de Renesse, Practice Leader, Consumer Technology and Media & Entertainment