Want to be a true strategic partner? The key is insights

Over the course of my career working with marketers and agency partners, it's inevitable that I get asked, "Tell me something I don't know." For some, this request can be frustrating or daunting. The truth is, it's one of my favorite to answer. The ability to share something new about your business, your existing customers or even your prospective customers is one of the ways to earn the title of "strategic partner." And the key to this is insights.

Similar to "innovation" and "creativity," the word insights today has become so overused that it's been commoditized and devalued. Yet do people really know what an insight is? Oxford defines an insight as "the capacity to gain an accurate and deep understanding of someone or something." This is a high bar. For our industry, it requires a deep understanding of consumer behavior. Kevin Drew Davis, the chief creative officer of Wunderman Chicago & Canada, has a simple way to bring it to life: A fact (or data point) is that people tend to feed their pets twice a day. An observation (or information) is that they tend to feed them at breakfast and dinner time. An insight, however, is that people feel guilty eating in front of their pets.

Can't you just Google this kind of information? Well, just because you can Google something doesn't mean you're getting an insight. If we're going to respond to the request to "tell me something I don't know," then we must become obsessed with understanding as many dimensions as possible about consumer behavior. And that starts with data. Not all data is created equal, however — "garbage in, garbage out," as the saying goes. Given that it comes directly from the consumer, first-party data is vastly more valuable than third-party data, for example. An internet-scale first-party data set that is unduplicated, real-time and not panel-based is a prerequisite to analyzing the many behavioral dimensions of your audiences.

In today's data-driven world, it's now mission critical to differentiate among data, information and insights. Over the past decade, billions of dollars have been spent on data-related initiatives, from building first-party data assets, to buying of-the-shelf audiences built using third-party data to implementing data management platforms that typically end up with slow or stale data. However, ask any brand today what it needs most, and the answer is the same—it needs insights. For some reason, despite all the money invested, we continue to be data rich but insights poor.

I believe that we're at a tipping point where it's no longer possible for people to sift through the continuous deluge of data searching for insights. Smart executives know we now need to leverage machine learning to inspire human learning. We need to use AI to understand the patterns in the data so we can unlock and discover the insights that are not intuitively obvious.

Forrester predicts that AI will spark an insights revolution and that insights-driven firms will gain disproportionate competitive advantages in the marketplace. The downside? There will be many charlatans calling themselves "insights-driven" when they are not.

At a time when brands need insights more than ever, buzzwords won't work. Real AI is hard. You can't fake AI. AI requires cloud-based computing power, software that learns and, most importantly, a constant stream of real-time data at scale (ideally lots of first-party data).

Here, too, we must manage expectations about AI. Despite all of the transformative business impact AI will have, most companies still underestimate the time, the talent and the resources needed to implement and maximize it for use cases such as insights. Indeed, as businesses embark on their journey to create AI-driven insights, they are realizing the need for organizational change. To harness the power of AI, you will need to reimagine how teams and functions are organized, breaking down the many silos in order to get one's own data house in order. It's only a matter of time before companies recognize this is a necessary prerequisite to applying AI techniques such as supervised, unsupervised and reinforcement machine learning.

While it won't happen overnight, I'm excited that we are seeing the shift toward AI-driven insights. Our rallying cry should be to challenge the conventional wisdom and to question everything.

While so many aspects of our industry are being transformed by technology, insights have continued to lag. I believe the AI-driven insights revolution is just getting started. Let's all raise the bar on what is considered an insight. If someone says they have an insight, challenge them. Is it really an insight, or rather a fact or an observation? Did this insight come from first-party data? Was AI used to discover unknown patterns in the data?

It's time to question everything that we know about insights. Only then will we be in the best position to reply to the request, "Tell me something I don't know."