Google plots downfall of last-click attribution, ascension of machine learning
- Google Attribution, a new free service currently in beta, was designed to bring together data from all of a marketer's channels in one place and provide a complete view of performance, the company revealed yesterday during the keynote at Google Marketing Next, to listeners of a live stream of the annual event and in a blog post.
- Another big focus of the keynote was the many ways that machine learning is enhancing Google's offerings, including new developments like helping marketers tie YouTube TrueView video ads to store visits and bringing audience insights to search ads.
- Also revealed during the gathering of more than a thousand marketers was the arrival of location extensions for YouTube video ads so marketers can help consumers researching a purchase find a nearby store where items are available and the rolling out of store sales measurement at the device and campaign levels.
Of the several announcements made at yesterday’s event, Google Attribution was one that seemed to get the biggest response from the audience. This response suggests not only the headache that cross-channel and cross-device attribution presents but also marketers’ faith in Google’s ability to potentially make some headway on a thorny issue that has been a challenge for years and whose importance only grows as channels and devices continue to proliferate. Google execs also expect the Attribution news to be a big deal: “This year, we’re solving the attribution problem,” said Bill Kee, Group Product Manager, Attribution, during the keynote.
Google’s efforts to address attribution are not purely altruistic. The company is likely betting that if marketers have a clearer view of what’s driving results, they will invest more in display advertising, an area some are expecting to retreat as consumers demand less intrusive experiences. In fact, Forrester recently forecast that CMOs could pull up to $2.9 billion in display ad spending next year.
“Let’s be blunt: Google’s renewed advocacy for upper-funnel marketing will make more revenue if more advertisers focus less on the last click before purchase, and start focusing more on top-funnel activity, namely display ads,” said Chaitanya Chandrasekar, CEO/Co-Founder, QuanticMind in an email comment provided to Marketing Dive.
Because of challenges inherent in tracking consumers across channels and devices, many marketers rely on last-click attribution, which often doesn’t provide a full picture as the consumer path to purchase becomes increasingly complex, involving research on a smartphone, reading reviews on a tablet or desktop and purchasing in-store. By giving the last click full credit for driving a sale, marketers fail to properly value the other steps along the journey to purchase.
Google Attribution, which will be rolled out to marketers over the next few months, is being positioned as a first by the company because of how it enables marketers, at no additional cost, to measure the impact of their marketing in one place, track consumers as they move between devices and is integrated with ad tools like AdWords, Google Analytics and DoubleClick Search.
Google Attribution can display traditional last-click attribution measurements but, by switching to data-driven attribution, the service will leverage machine learning to determine how much credit to assign to each step in the consumer journey. The data can be used to optimize marketers’ display and search campaigns.
The machine learning advancements announced yesterday also have the potential to significantly enhance marketers’ ability to make sense of the wealth of data in a multi-channel digital environment.
Over the past few months, Google has been ramping up its use of machine learning and mapping technology to help retailers measure store visits. Yesterday’s news furthers these efforts by bringing store visit measurements to YouTube TrueView campaigns so brands can measure the impact of video ads on store visits.
Machine learning helps Google analyze trillions of search queries and activity across millions of websites to figure out when consumers are close to making a purchase. This in-market audience data is now being made available to DoubleClick users for optimizing their search efforts.
“Smart use of data science, combined with machine learning, can help advertisers decrypt data into profitable, actionable insights that tie everything together,” wrote QuanticMind’s Chandrasekar in an email. “When Google itself is beating that drum, this is a strong affirmation that intelligent automation and big data are continuing to be the leading forces in digital.”