Predictive Analytics Wrong After Data Breaches?
Marketers are in love with big data and predictive analytics, yet they seem to ignore data breach issues. However, what if data breaches started skewing customer information to the point of all marketers’ analytical models being wrong?
Here’s an example of how that could happen: Netflix accounts are for sale on the “dark Web.” The dark Web buyers not only pay a lot less for access to the real users’ accounts, they’re throwing off the data by watching content the customer-of-record may not want to watch. Therefore, any lookalike models trying to capture, for instance, high-worth individuals would be wrong.
In the case of Netflix, this isn’t some nightmare scenario. This is actually happening. And it’s easy to imagine how it could happen to other marketers.
Most of the articles about the Netflix hacks are from the consumers’ point of view.
“Your account could be closed because of something the Dark Web buyer does,” writes AmeriPublications on Dec. 4, “or the buyer could make purchases using your stored credit card information.”
And other outlets, such as BloombergBusiness, are saying marketers may not care because investors don’t penalize them. “The stock market practically ignores” cyber attacks that involve theft of “names, account passwords, email addresses, physical addresses, phone numbers and birth dates.” Netflix stock didn't seem to be suffering on Tuesday, reaching nearly $128 at 3 p.m.
But maybe marketers should be worried about consumer reactions.
“Retailers may be shocked to learn that nearly 70 percent of the consumers we surveyed could correctly identify companies that had been breached,” reads a June 2015 CNBC article. “And they care. When asked how reports of data breaches have impacted their shopping habits, 15 percent of respondents said they generally stopped shopping at breached retailers and 23 percent generally stopped using breached payment methods. Furthermore, our survey found that when a consumer has been a victim of a breach, his or her reactions are even more pronounced.”
However in the meantime, hacks could be inserting data errors beyond the routine ones that already cost companies money. According to a February 2015 post in ITBusinessEdge.com, routine errors involving duplicate customer data entries, incorrect addresses and outdated lists could mean “the average company could waste ‘$180,000 per year simply on direct mail that does not reach the intended recipient because of inaccurate data,’ as Lemonly.com claims.”
So once sabotage enters the picture, how accurate can predictive models be?
Please respond in the comments section below.
Related story: Data, Predictive Analytics Are Worth Billions