I was standing over the sink, finishing off the last few bites of a coconut cupcake. Through the reflection in the kitchen window, I watched myself bring a mound of sugary, buttery heaven to my face just as the wearable fitness tracker on my wrist flashed the word, "GOAL!" I almost dropped the last bite into the sink out of guilt.
I started thinking about how this little fitness tracker that (almost) could, with all of its sensors and connected-to-the-Web awesomeness, was still utterly lacking in basic intelligence.
For the better part of two years, digital marketers have been waxing poetic about the Internet of things and the promise that networked devices hold in connecting consumers and brands. There are network-connected light bulbs, slow cookers, refrigerators, and more, all tempting marketers with the lure of more data about our customers' preferences and behaviors. As everyday objects in our homes, cars, and even on our bodies become increasingly connected, marketers and brands have unprecedented opportunity to reach customers at the right place and right time.
But, as it turns out, we're really bad at doing anything very meaningful with that data.
If we aren't thoughtful in our execution, our message can quickly degrade into noise, and we risk eroding the trust that we've built with our most loyal customers.
Considering all of this data that we have about our customers (and the vast amount of data on its way through connected devices in the coming months), one would think we'd be able to accurately and consistently target at least some basic messaging. Yet how many of us have received ads on social media that conflict with our core beliefs--beliefs that would be obvious to a 5-year-old reading just a single day of our news feed.
At issue here is our tendency to make overly broad assumptions about our customers based on minute data sets. It's one thing to assume purchase intent if a customer has begun the checkout process on our e-commerce site. It’s something altogether different to assume goals and infer characteristics based on casual, lightweight interactions and keywords.
The challenge in this scenario lies in understanding nonstructured data. It's much easier for us to make sense of past purchase history data than a collection of status updates. Natural language processing is profoundly difficult. Yet that is exactly where an untapped fount of wealth lies.
But it's not just the challenge of language: We still struggle with structured data, too.
For example, I was recently browsing an airline's online reservation system, looking at flights to Costa Rica. I tinkered with a few dates, got a general idea of price ranges, and then went about my business. I never actually started the purchase of any ticket. Two hours later, I received an email with the subject, "Complete Your Reservation!" The email began with "Did you forget something...?"
I get it. It was a nice try--much more sophisticated than other half-hearted attempts I've seen, for sure. But it also left me with the feeling that this airline was far too eager to just sell a ticket. The interaction was too early, too shallow, and assumed far too much.
What could have been done differently? Perhaps the airline could have tucked away the observation that I was planning a trip and that Latin America was in my consideration set. It could have waited patiently for me to engage further, watching as I searched for other destinations, and thus fine-tuning its recommendation. Perhaps after a few interactions it could have followed up with an email for planning the perfect tropical getaway, or maybe it would have picked up on the fact that I was considering multiple destinations, but all within the same time period. The airline could have recommended other itineraries for perfect vacation destinations in early summer.
When we act on small pieces of data, we're wasting money and annoying our customers. Rarely does a single piece of data give us meaningful insight into real intentions and goals. We're too eager to get our message in front of a potential customer. Our intentions are good, but our execution is way off.
From my perspective, we have two options: We can continue doing the same thing that we're doing today, with the risk of becoming even "noisier" as data from connected devices grows exponentially. Or we can dedicate ourselves to addressing the challenge of diving deeper with data and creating more meaningful and rewarding relationships with our customers.