Commentary

Are Current Technologies Truly Capable Of Data Wrangling?

by Tyler Loechner , Staff Writer, May 8, 2013

I was invited to watch a Festival of Media Global 2013 panel called "Algorithm Versus Man" via a Google+ Hangout, but unfortunately the panel time was changed while I was still asleep here in the United States. A week later, the panel was made available to me via Dropbox. So much for being able to watch in real-time with today's technology.

Oddly enough, what today's technology is truly capable of was a major focus on the panel.

The moderator was Mark Palmer, founder, Maverick Planet. The panelists included George John, CEO and founder, Rocket Fuel; Ricky Liversidge, CMO, DG MediaMind; and Sameer Singh, VP & head of global media planning, strategy and buying, GlaxoSmithKline.

Palmer opened up the panel by giving two movie examples of what machines could become. The first is the lovable, kind, animated Wall-E. The second is the ultimate science-project-gone-bad example: Terminator. Palmer cited and IBM study that claims the biggest worry of global CMOs is a "lack of control of data." So he asked the panelists whether or not the industry is exaggerating what the current technologies are capable of when it comes to data wrangling.

John said that computers learn over time, and that the technologies available present "a great opportunity for marketing." When it comes to the machines learning, John said, "They just can."

"They just can?" Palmer challenged in a mocking voice.

John later explain how the machines "just can" learn. Essentially, he said, the machines put together a giant list of impression served and whether or not they worked. He said that the machines look for patterns and can readjust accordingly.

"We all got here with machines," said John. "Planes, trains, cars." He argued that once you get comfortable with machines and what they can do to help, they are no longer menacing.

There is obviously some disconnect, though. If all of these technologies are truly capable of simplifying data enough, would the number one worry of a CMO be the lack of control of data? There's a disconnect somewhere.

big data, data management, real-time

4 comments about "Are Current Technologies Truly Capable Of Data Wrangling?".

Check to receive email when comments are posted.

Henry Blaufox from Dragon360, May 8, 2013 at 4:18 p.m.
Tyler's subjects in the discussion seemed to address two related issues. First, what the "machines" - computers processing massive amounts of data and reporting trends and correlations discovered - can do. Second, what people - in this instance CMOs - want from the machines.
The CMOs worry about a lack of control over the data. Is this another way of stating that by and large they don't trust the information they are receiving from their systems, so are wary of acting on it?
This lack of trust in the data presented is not new, and not unique to marketing.
Reply

Michael Lynn from ECD Consulting, May 8, 2013 at 4:43 p.m.

To follow up on the above comment. In order to be confident about the output you have to be just as confident about the input. But because so much of the data is based on unknown/unexplained algorithms, how can you trust it? The machine may get better...but at what? Garbage in, garbage out?

George John from Rocket Fuel Inc, May 8, 2013 at 6:07 p.m.

Hi Guys. George John from the panel here.

The moderator was paraphrasing the IBM CMO survey, but he paraphrased in a way that unfortunately packed in some meaning that wasn't originally there.

IBM asked CMO's which opportunities they felt most unprepared to manage. The #1 answer was "data explosion." So there's no undercurrent of things being "out of control" or not trusted, just that it's a huge opportunity that they are unprepared for. And that's ok. They'll figure it out, and smart companies will build technology to help them.

Linda Dorman from Consultant, May 8, 2013 at 6:18 p.m.

Machines aren't menacing but perhaps they should be perceived as tools that help us make better decisions rather than a replacement. Yes, we "all got here with machines" but when I get in my car, it still doesn't know where I want to go or the route I want to take based on my past travel patterns. As one former mentor once explained to me, "just because you can doesn't mean you should".