Google Attribution Aims To Link Offline Purchases to Online Ads
Machine learning can help marketers better track the success of their ads, whether customers ultimately end up buying online or off, thanks to a new tool unveiled by Google yesterday. However, consumer and privacy advocates are expressing concern about the potential implications of Google's ever-expanding ability to understand user behavior from first search to final purchase.
Google Attribution, whose launch in beta was announced yesterday during the company's Marketing Next conference in San Francisco, is a new offering that lets marketers better assess the effectiveness of their advertising, according to senior vice president of ads & commerce Sridhar Ramaswamy. In a blog post about the tool's launch, Ramaswamy said Google Attribution uses machine learning to provide data-driven insights about each step in the consumer journey.
Tracking Across Devices and Channels
"For the first time, Google Attribution makes it possible for every marketer to measure the impact of their marketing across devices and across channels -- all in one place, and at no additional cost," Ramaswamy said. Rather than just relying on last-click attribution, which misses the impact of most marketing touchpoints, Google's new tool makes use of data from a range of applications like AdWords, Google Analytics, and DoubleClick Search to let marketers better understand consumer actions across devices and channels, he said.
Over the coming months, Google also plans to roll out new capabilities enabling marketers to understand the in-store sales impact of their ads at the device and campaign levels, Ramaswamy said.
For example, large retailers who take email information at the point of sale from shoppers enrolled in their loyalty programs will be able to import data about such transactions into AdWords to better analyze campaign-related purchases by individual stores.
"And even if your business doesn't have a large loyalty program, you can still measure store sales by taking advantage of Google's third-party partnerships, which capture approximately 70% of credit and debit card transactions in the United States," Ramaswamy said. "There is no time-consuming setup or costly integrations required on your end. You also don’t need to share any customer information. After you opt in, we can automatically report on your store sales in AdWords."
'Anonymity Is Anything But'
Even after Google Attribution moves out of beta, the tool will not be available to all marketers, the company noted on its support pages. In general, AdWords users must have at least 15,000 clicks and 600 conversions over 30 days for data-driven attribution to be effective, according to Google.
For marketers that do qualify, transactions will be matched back to their Google ads "in a secure and privacy-safe way, and only report on aggregated and anonymized store sales to protect your customer data," Ramaswamy noted.
However, the advocacy organization behind Consumer Reports does not find those protections completely reassuring.
"Coordinating all those different online and offline data points into one single person's profile has basically been the holy grail of advertising for years now," the organization noted today in an article on its Consumerist Web publication. "That said, 'anonymity' is pretty much anything but. If anyone's looking at your digital breadcrumbs, they can be reasonably sure you are you from shockingly little data."
Companies track and monitor so they can advertise to consumers, Renate Samson, chief executive of the U.K.-based privacy organization Big Brother Watch, told the BBC today. "If we don't want them to do that, take control; don't give your email address for a digital receipt, check the terms and conditions, avoid using loyalty cards and where possible choose to pay with cash," he said.
Image credit: iStock/Artist's concept.
Posted: 2017-06-01 @ 12:01pm PT
After the keynote, a ton of articles have been published talking about how great Google Attribution is. There are some major flaws that I've seen with their model just from what we run with our current clients.