Data analytics lets marketers understand everything they need to know about their clients, from the movies they watch on Netflix to their favorite scoop of chocolate ice cream.
Unless, of course, it isn't. Except when it isn't. Data is ubiquitous, necessary, and beneficial — except when it isn't.
Experts believe that data analytics is approaching a tipping point. Security, privacy, discrimination, and law concerns are colliding with all of the benefits that machine learning and artificial intelligence can provide. Concerns about the coronavirus outbreak and how it has influenced consumer behavior so swiftly exacerbate the problems.
"We're seeing a lot of turbulence in terms of what is the appropriate response." And we're seeing a strategic shift," said Neil Hoyne, a senior fellow at Wharton Customer Analytics and Google's chief measurement strategist.
Hoyne said he's continually speaking with companies attempting to figure out what the future of data analytics solutions will look like. Third-party cookies are being phased out by Google and other internet businesses, depriving advertisers of a wealth of fine-grained data acquired by tracking customers around the web. Companies who are ahead of the curve are already preparing for a post-cookie, post-pandemic world.
"The firms that are going to succeed are the ones that are using data, not guessing," said Hoyne, who talked with other corporate and academic experts at a virtual symposium on "The Use of Analytics and AI in a Post-pandemic World" on Nov. 17. The event was organized by the Marketing Science Institute in partnership with Wharton Customer Analytics and AI for Business at The Wharton School.
The conference addressed a wide spectrum of data analytics services while focusing on what lies ahead in the fast growing field of artificial intelligence.
"We couldn't avoid the obvious in trying to build the program today," said Kartik Hosanagar, AI for Business Faculty Director and Wharton professor of operations, information, and choices. "2021 has been the year of disruption and risks — and, ideally, successful management of those risks."
The Risks And Benefits Of Data
The amazing ways that AI and machine learning aid businesses by automating services, anticipating patterns, and producing recommendations that lead to improved sales and engagement have gotten a lot of press. Amazon's recommendation algorithm, for example, is responsible for a third of its sales, while YouTube's algorithm is responsible for 70% of video views on its platform.
Managers, on the other hand, must pay equal attention to and priorities the risks associated with AI, according to Hosanagar.
Even for companies that are technologically savvy, AI can cause social, reputational, and regulatory issues. The US Department of Housing and Urban Development sued Facebook, alleging that the platform's targeted advertising violates the Fair Housing Act by restricting who can see housing ads. Amazon cancelled a gender-biased recruiting software; Twitter deactivated a Microsoft chatbot that "learned" how to send racist tweets; and the US Department of Housing and Urban Development sued Facebook, alleging that the platform's targeted advertising violates the Fair Housing Act by restricting who can see housing ads.
"These are not insignificant risks for enterprises," Hosanagar continued, emphasizing the significance of building interdisciplinary teams to monitor and examine data for bias on a regular basis.
Unwittingly, humans who build algorithms might introduce prejudice into them. At the conference, Kalinda Ukanwa, a marketing professor at the Marshall School of Business at the University of Southern California, presented and provided a compelling example to emphasize the issue. "Rebecca" asks for a loan with an online bank that uses artificial intelligence to make its judgement. She is rejected down despite having good credit. She is, however, approved if she enters the identical information as before, with the exception of her gender.
While the online bank may see an increase in revenue initially as a result of its ease of use, it may suffer long-term reputational damage. Rebecca may advise her friend "Jim" months later that she shouldn't bother applying for a loan at that bank because she had a bad experience there.
"Algorithm bias can be useful in the short term, but it's counterproductive in the long run since word of mouth reduces customer demand," Ukanwa added.
Despite this, she emphasized the importance of
data analytics solution providers. When it functions well, it removes uncertainty from decision-making and can lead to more equal outcomes. AI, on the other hand, must be closely monitored and changed. There are times when a straightforward solution is available. The bias might have been avoided by simply deleting the gender element from the bank loan case.
Uncertainty Over The Pandemic Has Prompted Many People To Take Action
Since a result of the current COVID-19 pandemic, certain
data analytics service providers that were formerly beneficial have become obsolete as consumers' buying behaviors have substantially changed. According to Google's Hoyne, smart organizations are responding by transitioning from precision measurement to prediction. They're experimenting with what they can do with the data they already have rather than acquiring more. In order to develop more direct links with their clients, they're also moving away from cookie-based third-party data and toward first-party data.
He says that companies are less interested in tracking consumer data from the past since the past no longer matters. Due to rising worries about privacy and legislation, businesses are also looking into ways to make their data more open and dependable to their customers.
These are small changes, not a total overhaul. "They just want to be a little bit better," Hoyne said, describing the approach as "refreshing" for firms because it is more long-term.
Barkha Saxena, Poshmark's chief data officer, cited her organization as an example of adaptability in unpredictable times. Poshmark has always made decisions based on data, and it has implemented an integrated approach that allows it to react fast to market fluctuations. She proposed a framework that other organizations might use to do the same: evaluate the data, implement the plan, learn what worked and what didn't, and repeat.
"This is how you transform data into a useful tool," explained Saxena.
When it came to statistics, she also emphasized the need of having a team mindset. It shouldn't be limited to a particular department, but should be applied to all company procedures.
She explained, "We have a base of incredibly consolidated, reliable, and easy-to-access data, but it is then supplied to all of the teams." "At the moment of the decision, all business users have access to the data."
Comments
Post a Comment