How Volkswagen is using artificial intelligence for ad buying decisions
Cars aren’t the only thing Volkswagen wants to automate. Artificial intelligence is managing the brand’s media buys in Germany and proving to be more effective than its media agency.
Whenever Volkswagen uses the recommendations from Blackwood Seven, a Danish media agency that uses AI and predictive analytics to forecast ad spend decisions, it sells more cars than it would have if it had gone with its media agency’s recommendation, according to Lutz Kothe, the head of marketing for Volkswagen’s passenger cars in Germany. Kothe said his team uses Blackwood Seven’s algorithm to buy the right ads based on sales.
Defining a role for the AI platform, however, has taken Volkswagen a while. Two years ago, the advertiser ditched MediaCom and gave its digital business in Germany to Blackwood Seven. At the time, Kothe wasn’t sure the algorithm would be able to deliver better media recommendations than its agency, but the signs were promising just months into the deal, according to Kothe. Between September and December 2016, Volkswagen used the algorithm’s media recommendations for a campaign for its up! model, which led to a 14 percent rise in orders from Volkswagen’s dealerships versus what those orders would’ve been had the campaign run solely on its agency’s recommendations. In some instances, the difference between the algorithm’s and its agency’s car orders has been as high as 20 percent, revealed Kothe.
Since those early campaigns, Volkswagen has applied the algorithm to all its media strategies in Germany. Every car model there from the advertiser has a different strategy, predicting which media investments will give the best returns for its marketers. Those forecasts are based on Volkswagen’s transactional data — or incoming orders — as well as market data such as fuel price, competitor prices and overall car registrations from sources such as Nielsen. Like other AI platforms, the more data fed into the algorithm, the smarter Blackwood Seven’s recommendations become.
Volkswagen’s marketing team bought recommendations from the platform for display, search and social in 2017. By using the AI platform’s recommendations to pick its media, Volkswagen was able to avoid any hidden rebate costs it might have incurred from buying via an agency. In 2018, all Volkswagen’s media buys will be recommended by the AI platform and then bought by its global agency PHD
Kothe would not comment on the prospect of Volkswagen eventually replacing all its agencies with AI platforms, instead stressing how important AI could be to finding marginal gains in media.
What separates Blackwood Seven from Volkswagen’s agencies, Kothe said, is the AI platform’s capacity to process reams of data. Volkswagen’s marketers pull data from more than 1,400 touch points, yet they relied on the “personal interpretations coming from the agency,” which wasn’t based wholly on the data, Kothe said. The biggest advantage the algorithm gives Volkswagen is the ability to better predict what the brand’s media investments will do, he added. The more precise the recommendations become, the more “we can steer our media activity into the best areas,” he concluded.
This means the brand can invest less for some campaigns and still see a sales uplift, Kothe said. While he would not reveal details on those campaigns, he noted how radio, which had been regarded as somewhat of a stale media for Volkswagen in Germany, has been used more frequently over other media to launch new cars, per the algorithm’s forecasts.
But trusting a machine hasn’t been easy. After two years of working with Blackwood Seven and another year before that of early discussions, Volkswagen’s German team is only now talking to other markets about the algorithm’s performance. While hype continues to build around AI replacing media agencies, the reality is it won’t happen anytime soon. Blackwood Seven’s failed launch in the U.K. is a testament to that; the startup closed its office there last month after just a year, despite offering to give marketers more control of their media spend at the height of the transparency crisis.
Image courtesy of Volkswagen.
Cheat Sheet: Google unveils timeline for a more ‘responsible’ cookie death clock
Google elaborated on its timeline for killing off third-party cookies as part of its promises to the UK's antitrust authority.
How news publishers are using the Olympics and AR to flex their emerging tech storytelling
Big publishers like The Washington Post and USA Today are developing and expanding AR storytelling around the Olympic Games.
‘Weak Sauce’: New industry tool for opt-out from email-based tracking misses ID tech and key players like Facebook and Liveramp
The Network Advertising Initiative's new privacy control is intended to stop email-based audience matching — often referred to as onboarding.
SponsoredHow the ad industry can use its borrowed time to future-proof first-party data solutions
Trent Lloyd, co-founder and head of brand solutions, Eyeota Google’s updated timeline for its Privacy Sandbox rollout, including its two-year delay of third-party cookie deprecation on Chrome, didn’t come as a surprise to many industry observers, given the limited utility of Google’s FLoC and the slow momentum of the Privacy Sandbox in the World Wide […]
Member ExclusiveMedia Briefing: Publishers’ programmatic ad businesses have rebounded to pre-pandemic levels
This week's Media Briefing looks at how the pandemic and the cookie's eventual demise have created the conditions for the programmatic ad market that publishers have been pushing for, with a shift to private buying coinciding with prices pushing past pre-pandemic levels.
‘They will need to use multiple routes’: Shifts appear in the publisher-SSP union, as alternative identifiers proliferate
As the ad tech industry rewires itself around the contours of privacy, supply-side platforms are reinventing themselves (again).