The Drum Awards Festival - Extended Deadline

-d -h -min -sec

IBM Watson Microsoft Google

Marketers must learn to hack people instead of algorithms


By Lisa Lacy, n/a

May 23, 2017 | 6 min read

The onslaught of data and machine learning will potentially reveal some uncomfortable truths about consumers – and this will have profound consequences for brands and marketers.

Machine learning will reveal some uncomfortable truths about consumers and marketers will have to adapt.

Machine learning will reveal some uncomfortable truths about consumers and marketers will have to adapt. / Pixabay:

That was according to Lexi Mills, managing director at Marquis Communications and consulting vice president for digital marketing collective Manyminds North America, who spoke Monday (May 22) at the Inbounder in New York.

The technology we use is creating epic amounts of data – to the tune of 2.5bn gigabytes in 2012 – with Google processing over 40,000 search queries every second, Mills said.

And while Google originally rejected machine learning, a type of artificial intelligence (AI) in which machines learn when exposed to new data without being programmed, it has since changed its tune. In addition, Google owns properties like YouTube and Google Maps and “can measure the ripple effect of our plans”, Mills said. In other words, Google knows a lot more about consumer habits and context – and that means PR and SEO strategies of yore, like, say, building links or creating infographics, are no longer enough – and this is particularly true as machine learning takes off.

“If you want to continue to optimize over the next five to six years, the ripple effect is important – you need to understand everything about consumers more than ever before,” Mills said. “We are moving away as marketers and publishers from hacking algorithms to hacking people.”

The media, too, is changing – increasingly preferring formats like video over text, and, in some cases, blatantly seeking clicks – even when that comes at the expense of more noteworthy stories – because traffic pays the bills. And there is a lesson therein for marketers seeking exposure for their brands. In fact, Mills called it a “key aspect of optimization from my perspective now”.

And as an increasing percentage of content is produced by machines, Mills said we are creating bubbles on social media that become self-locking.

“Facebook finds we only click on things we agree with,” she added.

That, in turn, exposes some truths we may not like.

Look no further than Microsoft’s Tay chatbot, which, Mills noted, “started off awesome…and then she started chatting to people”. Less than 24 hours later, Tay was racist, sexist and anti-Semitic.

IBM Watson had a similar problem when the team behind it fed it the Urban Dictionary to make it more colloquial.

For her part, Mills was able to have some fun with a more lighthearted consumer mistake with the site, a parody bathroom products site she created on behalf of “when we realised that over a million British people searching for new bathrooms were popping ‘bathroom sweets’, rather than ‘bathroom suites’ into their preferred search engines”.

The effort itself generated a lot of earned media, but also showed when it comes down to literacy versus chocolate, chocolate wins – and if marketers strategically go after whatever is trivial, they’ll win.

“Reward is based on metrics, not morals,” she added.

And this reveals perhaps another uncomfortable truth: Google can tell you how to boil an egg, but not the meaning of life, Mills said.

“We need to keep our social data bubbles open. If you want to be racist, you should be it by choice at least,” she added. “We have a new set of brand problems arising. Our technology is smarter than we are and it is going to get smarter, better, faster.”

Pointing to the infamous example in which Target knew a teenage girl was pregnant before her father, Mills said “this is going to be a big…problem. Google is getting really smart…do you have a strategy in place for when this happens to your clients? It is going to be a really big problem and absolutely no one is prepared for it. This will be a big problem in our era. We have no backup plan. We need to rethink it.”

And it’s not just data – it’s also short attention spans, which mean marketers have to tell brand stories in headlines. And that’s not so bad when brands are selling candy or shoes, but is harder when it’s advanced technology and/or a product with no existing framework in which to understand it. That means marketers have to start thinking in really short bites and use retargeting to get back to consumers and tell the next chapter of the brand story.

That was part of the inspiration behind the faux IoT handbag the iBag2 for comparison tool to educate consumers about financial wellness. The project included assets like video, infographics and GIFs and “worked because I accepted the truth of where our market is”, Mills added.

The same is true of the 3Doodler, which Mills called a “glorified glue gun”, but, she added, “‘Glorified glue gun’ is not clickable.” And after positioning it as a 3D printing pen, Mills said the project generated over $1.5m on Kickstarter.

In fact, Mills recommended going to Kickstarter to see what’s emerging as if it is trending there, it will be trending in consumer technology soon thereafter.

And as we move toward a world in which Google may send push notifications about a nearby burger joint to consumers stumbling out of a nightclub at the end of the night, it’s even more important than ever for marketers to be great storytellers and to be charismatic, she added.

IBM Watson Microsoft Google

More from IBM Watson

View all


Industry insights

View all
Add your own content +