'You can only try to travel back in time and stop Terminators': Andreas Misera on AI in advertising
Andreas Misera is a seasoned professional of technical operations with more than 20 years of experience in the Marketing and Media industry. He is in charge of all Criteo's operations around technical solutions, analytics and creative services in Central Europe (DE/AT/CH), Russia/CIS, Middle East & Africa. On May 21, Andreas will speak at Adsider LIVE Conference. We talked with the speaker for a quick interview.
As 2019 became the year of innovations in many industries, how did the technology stack of Criteo improve through the last year?

Innovation has always been a key element of Criteo's business. From the early days, we've been heavily investing in our own R&D department which enabled us to become the world leader in retargeting. That's what we are known for in the industry; however, this perception falls way too short. Today, our product portfolio is no longer limited to just driving conversions. We offer personalized advertising across all channels and devices, and along the funnel.

This includes our recently launched traffic generation solution to attract new users to our clients' websites and our app engagement solutions, to just mention a few. We are already looking to expand our footprint even more, by rolling out our innovative Retail Media business — a solution enabling brands to advertise their products on retailer websites — to more clients and markets and by exploring new areas of business like Connected TV. On top of that, we launched our AI Lab in Paris back in 2018: 100 engineers and researchers are dedicated to making innovations in computational advertising in the AI Lab, in order to drive advertising technology of the future.
What does excellent omnichannel marketing strategy mean exactly?

Omnichannel marketing has been a huge topic for many years now. Think about a world that enables you to combine different marketing activities. Like online display, social media, SEO, email, print, TV or out-of-home, while being able to analyze the real impact of all your activities on e-commerce, offline shopping as well as phone and mail ordering - now you are getting close to what a marketer would call an ideal world. However, Omnichannel is quite complex.

Having an excellent strategy does not only mean you plan these activities properly and well-coordinated; you also need to make sure your infrastructure — the way online and offline shopping systems are set up — enables you to work with all the relevant data.

We see a heavy momentum in Omnichannel activities driven by the current Covid-19 situation. Many users turn to online-shopping due to brick-and-mortar stores being closed or having restricted access, which pushes click-to-collect orders and other omnichannel measures to a new level. It's critical for retailers to not only look at how to compensate offline purchases now online, but to create an omnichannel experience that enables them to serve the wishes of their customers also in the long-term.
'We are known in the industry; however, this perception falls way too short'
How is artificial intelligence used in advertising?

There are many use cases of applying AI in Advertising. The main areas we see are automation and personalization.

Automation because AI can use many different variables very fast to make decisions on the road from inventory availability to showing an ad.

Personalization because many of these variables leading to fast decisions depend on user-specific data.

If you want an ad to fit a user specifically you have to make sure the content and look & feel is good for the user. Scaling this up can only be done with AI.
How does AI change advertising and how does the technology drive marketing results?

AI enables adtech to achieve better performance at scale by being more fine-tuned to what users want and how they react. There are many use cases, from Audience definition via RTB bid values to banner creation. A simple example is building banner content via AI — both the products to be shown and which colors, and the formats and layouts to be used.

Making these decisions based on each user's individual intent and behaviour makes the banner more relevant to the user and thus lets it perform better. Doing this at scale manually is pretty much impossible.

What will you speak about at Adsider Online Conference?

I'll talk about Artificial Intelligence and what it actually means. I'll also talk about Big Data and why it's relevant for AI. I'll explain how AI can be used in advertising to improve performance while taking over tasks that are too often done manually. In a nutshell: AI delivers better results with less manual work — at scale.
'Having an excellent strategy does not only mean you plan these activities properly and well-coordinated'
Why do programmatic solutions become more profitable compared to other types of advertising?

According to Zenith's Programmatic Marketing Forecast 2019, more than two third of digital ad spend in markets like France, Germany, or Russia will be done programmatically by the end of this year. In markets like the UK or US even more than 80 percent. These numbers clearly show that Programmatic became the standard for digital media buying.

One of the biggest advantages for publishers is a bigger competition for advertising inventory, which results in higher prices paid by the advertisers. Advertisers again benefit from direct access to the users they want to address in a specific moment. Players who run Programmatic directly by integration into a publishers' header — like we do at Criteo with our Direct Bidder — benefit from preferred access to inventory. That's why we expect the field of Programmatic Direct to grow even more.

Criteo has headquarters in almost 30 countries. Are there any new markets to enter in 2020?

Criteo's headquarters are in Paris, and the majority of our central teams like product and R&D are based there.

In a fast-changing industry like ad tech, it's crucial that these teams are well-aligned on new developments and changes in order to provide the best support and solutions to our clients all around the world.

On top of that, we have 27 additional offices from New York to London to Tokyo, in order to be close to the clients and with the teams providing direct day-to-day support. Before Corona changed the way we live and work, we regularly also visited our clients in markets where we don't have an office yet.

Becoming too granular is risky, as you also want to ensure that every employee has the required knowledge exchange with the colleagues from the region. We are thus thoroughly considering between regular travel to clients and new office openings. By the way: just today we successfully ran campaigns for our clients in over 100 countries.
'If we talk about Terminator style robots, you can only try to travel back in time and stop them!'
What will you do if one day the machines take over the world?

It depends on the machines that do it and their methodology. If we talk about Terminator style robots, you can only try to travel back in time and stop them!

On a more serious note: AI always works with a defined learning methodology towards a target — this implicates that humans like me and my colleagues have to define the target. In the advertising industry, the machines used for image recognition, user behaviour prediction and all the other relevant topics will never fully replace me in my job. They will rather change the way I do it, reduce manual work and help to scale my activities, so I can focus on the elements machines can only support, but not fully run on their own: the strategic and creative parts of my job.
Read next Ресурс 1