Social media data mining: Garbage-in, garbage-out?

gigoIt’s human nature for people to strive for the most flattering public persona … while confining the “true reality” only to those who have the opportunity (or misfortune) to see them in their most private moments.

It goes far beyond just the closed doors of a family’s household. I know a recording producer who speaks about having to “wipe the bottoms” of music stars — an unpleasant thought if ever there was one.

In today’s world of interactivity and social platforms, things are amplified even more — and it’s a lot more public.

Accordingly, there are more granular data than ever about people, their interests and their proclivities.

The opportunities for marketers seem almost endless. At last we’re able to go beyond basic demographics and other conventional classifications, to now pinpoint and target marketing messages based on psychographics.

And to do so using the very terms and phrases people are using in their own social interactions.

The problem is … a good deal of social media is one giant head-fake.

Don’t just take my word for it. Consider remarks made recently by Rudi Anggono, one of Google’s senior creative staff leaders. He refers to data collected in the social media space as “a two-faced, insincere, duplicitous, lying sack of sh*t.”

Anggono is talking about information he dubs “declared data.” It isn’t information that’s factual and vetted, but rather data that’s influenced by people’s moods, insecurities, social agenda … and any other set of factors that shape someone’s carefully crafted public image.

In other words, it’s information that’s made up of half-truths.

This is nothing new, actually. It’s been going on forever.  Cultural anthropologist Genevieve Bell put her finger on it years ago when she observed that people lie because they want to tell better stories and to project better versions of themselves.

What’s changed in the past decade is social media, of course.  What better way to “tell better stories and project better versions of ourselves” than through social media platforms?

Instead of the once-a-year Holiday Letter of yore, any of us can now provide an endless parade of breathless superlatives about our great, wonderful lives and the equally fabulous experiences of our families, children, parents, A-list friends, and whoever else we wish to associate with our excellent selves.

Between Facebook, Instagram, Pinterest and even LinkedIn, reams of granular data are being collected on individuals — data which these platforms then seek to monetize by selling access to advertisers.

In theory, it’s a whole lot better-targeted than the frumpy, old fashioned demographic selects like location, age, income level and ethnicity.

But in reality, the information extracted from social is suspect data.

This has set up a big debate between Google — which promotes its search engine marketing and advertising programs based on the “intent” of people searching for information online — and Facebook and others who are promoting their robust repositories of psychographic and attitudinal data.

There are clear signs that some of the social platforms recognize the drawbacks of the ad programs they’re promoting — to the extent that they’re now trying to convince advertisers that they deserve consideration for search advertising dollars, not just social.

In an article published this week in The Wall Street Journal’s CMO Today blog, Tim Kendall, Pinterest’s head of monetization, contends that far from being merely a place where people connect with friends and family, Pinterest is more like a “catalogue of ideas,” where people “go through the catalogue and do searches.”

Pinterest has every monetary reason to present itself in this manner, of course.  According to eMarketer, in 2014 search advertising accounted for more than 45% of all digital ad spending — far more than ad spending on social media.

This year, the projections are for more than $26 billion to be spent on U.S. search ads, compared to only about $10 billion in the social sphere.

The sweet spot, of course, is being able to use declared data in concert with intent and behavior. And that’s why there’s so much effort and energy going into developing improved algorithms for generating data-driven predictive information than can accomplish those twin goals.

Rudi Anggono
Rudi Anggono

In the meantime, Anggono’s admonition about data mined from social media is worth repeating:

“You have to prod, extrapolate, look for the intent, play good-cop/bad-cop, get the full story, get the context, get the real insights. Use all the available analytical tools at your disposal. Or if not, get access to those tools. Only then can you trust this data.”

What are your thoughts? Do you agree with Anggono’s position? Please share your perspectives with other readers here.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s