Top

A Downside to Big Data and Data Mining for Marketers

Mitchell Goozé | Sep 21, 2017 175 views 4 Comments

Share on LinkedIn

By now anyone who is a marketing professional knows that analytics, big data, data mining etc. are key tools that must be mastered. Unfortunately for too many it has become a crutch or excuse to stop thinking. Another downside is that too many marketers spend even less time than they did before “observing the native in its natural habitat.” They are relying almost exclusively on “data” to drive their decisions. I am a fan of data driven marketing, just not exclusively.

Another “downside” to this reliance on data is unintended consequences. Facebook has been selling ads for a while. One of the things businesses like about advertising on Facebook is the ability to target specific audiences. Facebook really does have way more information about their users than most users want to realize. That information is provided to advertisers to target. So far, so good.

However, this reliance on data by the people at Facebook caused a hugely embarrassing situation for them: turns out you could target people who hate Jews specifically. Turns out one of the targeting options was field of study. Over 2,000 people had listed their field of study as “Jew hater.” In addition the algorithms that Facebook used to help their advertisers target the right people included “why Jews rule the world” and “burn Jews” as fields of study that could be correlated to Jew haters.

Did Facebook offer this service on purpose? Of course not. In fact they have removed field of study as a targeting choice. Previously, their data had allowed people to exclude certain ethnic groups from their ads. They stopped that eventually as well.

How did this happen? Reliance on big data and suspension of critical thinking in my opinion. Data is good. Data without critical thinking and first hand observation as a supplement, can be misleading.

Mitch

Print Friendly, PDF & Email

Republished with author's permission from original post.


Recent Editor's Picks:


Categories: BlogCustomer AnalyticsThink Tank

175 views

4 Responses to A Downside to Big Data and Data Mining for Marketers

  1. Chip R. Bell September 22, 2017 at 6:03 am (186 comments) #

    Mitchell, this is a very good and important blog. It is important that we remember customers are humans, not computers. And, as humans, they are ever-changing, not static. As Tom Peters wrote, “Customers perceive service in their own unique, idiosyncratic, emotional, irrational, end-of-the-day, and totally human terms. Perception is all there is!” Big data is valuable but incomplete. Imagine managing your interpersonal decisions and relationships at home with only big data available gleaned through surveys and demographic data! It all might be accurate but ineffective!!

  2. Michael Lowenstein September 22, 2017 at 7:08 am (1310 comments) #

    There is always potential danger, via missed opportunities or pitfalls, with over-reliance on any single source of decision-supporting input or insight. The best marketing decisions are usually based on converging multiple streams, or sources, of information, some of which is structured, some unstructured.

  3. Andrew Rudin September 24, 2017 at 3:04 pm (231 comments) #

    Hi Mitchell: thanks for posting this thoughtful blog. I believe many companies get wowed by possibility of good outcomes, and overlook the negative. But project management practices must include identifying what’s uncertain, and when the risks are significant, making sure measures are in place to mitigate them. This is especially important with information technology.

    An armchair analysis of the scenario you described with Facebook suggests that the potential for nefarious use was not understood or identified, and that the company’s risk-analysis rigor was possibly inadequate.

    This serves as an important ‘lesson-learned’ for other companies adopting new procedures, processes, and technologies: every alleged solution always creates new problems. It’s important to determine what they might be.

  4. Mitch Gooze September 25, 2017 at 5:09 pm (15 comments) #

    Thanks guys, appreciate the feedback

    Mitch

Add Your Comment (All comments are reviewed by moderator, no spam permitted!)