Home to more than 12 models, SEAT is Spain’s most successful carmaker brand and employs more than 15,000 people worldwide. A member of the Volkswagen Group, SEAT has its headquarters in Barcelona and is present in 80 countries on all five continents. SEAT sold around half a million cars in 2018, making a record turnover of almost 10 billion euros. SEAT aims to be a key player in digital transformation and smart mobility; hence, it strongly promotes all of its processes and production activities to be adapted to the digital environment. getML team has been critical to supporting this goal by adding a machine learning layer to SEAT’s media system.
In a world of fragmented attention, getting noticed is a key to push the needle, and to stand out in the crowd. Furthermore, customers are approachable via different channels. From here, SEAT is following the multichannel marketing approach for further growth and recognition. SEAT runs data-driven ads across multiple networks via Google Marketing Platform DV360 that unifies advertising and analytics at one-stop. Having said that, the journey of traditional programmatic advertising is a tedious and complicated mission. It takes time, embraces costly bidding system, produces huge amount of data, pushes advertisements to the wrong people, with unmanageable amount of adjustment possibilities, therefore more room for errors. That is why AI-driven programmatic media optimization is taking over, resulting in automated processes, budget optimization, and better targeting. SEAT strives to improve audience targeting, and bring the maximum customer engagement, yet with high-cost effectiveness. To do so, SEAT and getML teamed up for an AI-driven programmatic targeting model. This advertising model has been designed to automate and optimize media buying, to the right people at exactly the right time. In other words, keeping a constant eye on SEAT while following smart and precise bidding, with less manual tasks and guesswork, and enhanced targeting.
Maintaining visibility and turning a quality audience into customers is very time consuming and often takes a big dent in the online advertising budget. Today's digitally-focused media campaigns are interspersed with several KPIs, and complex targeting, especially among large spendings. SEAT is tracking tens of different variables such as “geography, gender, and site placement” to capture and target the ideal audience out of the wide net. So, even more data with innumerable values and millions of complex data points, that exhaust a lot of staff time and grow unmanageable scenarios. This means, crunching through large datasets with insights scattered across various sources. It implies that the traditional media buying is likely to block the optimization and effectiveness of digital marketing campaigns due to the manual, slow and tedious processes. Hence, it has to be automated, flexible to adjust, in a real-time, brings deep consumer insights, and speaks to the right audience, so that ads are pushed to the right users, at the right time, and at the right price. The programmatic media buying enables companies to do just that.
The starting point was analyzing historical campaigns as well as the general behavior of the audience groups, to best understand the market and customers in different places and across different variables. The programmatic algorithm collected and evaluated data to perform media buying decisions at later stages. Afterward, getML programmatic optimization model was deployed to automate the buying process and effectively arbitrate the budget with real-time and cheaper bidding, thus reaching more users who are likely to become customers. Subsequently, trained models and key takeaways were applied to practice, seeking ongoing media optimization, and recommending the best potential set-up of programmatic campaigns. After all, getML remodeled SEAT digital marketing efforts and generated 2X viewable impressions, as well as conversions that are 47% cheaper, inducing further projects to enable further optimization.
The collaboration with getML team provided SEAT with an automated media buying model that surpassed the traditional media buying process. Using getML, SEAT got conversions 50% cheaper than the manually optimized media approaches. By identifying the best performing variables, the viewability of ads served by getML optimization model were doubled. Finally, getML Suite continues to provide SEAT with programmatic media model to run targeted and effective marketing campaigns.
The year is 2020. The world is entirely dominated by Deep Learning applications. Well, not entirely... This post highlights the blank spot on the map.
Automate Machine Learning on time series with getML