Even in uncertain economic times, marketing technologies continue to flourish. Overworked marketers facing a shortage of hours in the day are discovering enormous efficiency gains with a variety of automated solutions.
Marketing analytics is growing faster than any other segment in this category, promising everything from perfecting the customer journey to proving — to even the most skeptical of CEOs — that marketing is a revenue-generating corporate hero. But the danger is that amid the hype and hope of what analytics can do for marketers, we lose sight of the single most important component of the entire system: the human being.
This notion of “human-in-the-loop” computing has already become prevalent in other discussions: The Internet of Things, augmented and virtual reality and several other topics that might appear on a Silicon Valley buzzword bingo card. But only recently have we begun to explore the importance of a human-in-the-loop approach in marketing analytics — and it’s essential that we do.
What is human-in-the-loop computing?
According to Wikipedia, human-in-the-loop computing is defined as a computing model that requires human interaction. The primary benefit of a human-in-the-loop model is that it allows a human being to change outcomes, rather than relying on the automated process itself to determine the outcome.
The earliest examples of human-in-the-loop models were flight and driving simulators, which couldn’t have been developed without including a real person in the development process, since the simulator needs to incorporate the signals of the pilot or driver into the program.
A more modern example of human-in-the-loop computing would be Google’s self-driving cars. While some at Google would prefer to cut the human out of the loop, regulations currently require a steering wheel in the car and a driver behind that wheel — just in case the computer has a hard time predicting the movements of other drivers or a particularly difficult judgment call needs to be made. (Good thing, too, considering a Google self-driving car had its first accident on Valentine’s Day this year!)
Marketing is no less sophisticated and requires even more human input than flying a plane or driving a car, and so it only makes sense that marketers should be kept in the loop in any marketing automation system. The real question is, where and how should they engage?
Marketing analytics should aim for intelligence augmentation that enhances human judgment by eliminating irrelevant and misleading data inputs.
The very real, tactical challenge is that most marketers’ online calendars look like solid brick walls of time commitments, sometimes two or three layers thick with conflicting appointments. They don’t need more meetings, more data, or more presentations in their lives.
They need a reliable filtering mechanism to make sure they’re actually focused on the decisions that will ultimately impact their customer’s experience, their company’s performance and their own success. That’s the real promise of marketing analytics — filtering out what doesn’t matter and giving you real-time, quantitative visibility into the decisions that do matter.
Put another way, the goal of marketing analytics should not be the creation of artificial intelligence that replaces human judgment. Instead, marketing analytics should aim for intelligence augmentation that enhances human judgment by eliminating irrelevant and misleading data inputs.
Marketing analytics should operate in concert with, not in place of, the already time-tested algorithms of human curiosity and human intuition.
Here are some ways to make sure your own marketing team is plugged in at the right touch points in your own marketing solution and some warning signs that might indicate it’s time to up your engagement.
1. Focus on the right question
The only thing worse than the wrong answer is the right answer to the wrong question. And nothing beats pure, organic, free-range human curiosity when it comes to formulating questions that lead to impact.
Make sure you’ve got an enlightened brain watching your marketing analytics signals for signs that you might want to explore something further.
For example, let’s say your automated social listening software is telling you that customers are increasingly frustrated by wait times at your restaurant. A human with a lack of curiosity might take those findings at face value and assume the data was referring to the time it takes the server to bring the food after taking the order. This might lead to hiring more cooks.
But a deeper dive, sparked by human curiosity, might reveal the complaints were actually about the time it takes to get seated in the first place — a gripe that impacts revenue when people walk out in frustration before ever sitting down. The solution here might be better wait time management, remote wait time notifications that untether diners from the hostess desk or expanded seating options.
2. Read between the lines
Humans far outpace computers when it comes to understanding context, and language is all about context. Be wary of any marketing analytics system that relies on successfully teaching a computer the nuances of how people verbally communicate with each other.
Instead, look for systems that rely on the computer to do what it does best: perform predictable computations at massive scale and speed, filter out junk and look for quantifiable correlations. Then rely on the human brain to do what it does best: understand language, key into nuances and draw conclusions from patterns.
There are moments when logic is simply not your friend — and those moments are a daily reality for marketers.
For example, if a marketing analytics solution reveals that people can’t really tell the difference in the effectiveness of one kind of over-the-counter medicine over another, then the temptation might be to focus instead on packaging or pricing.
But if those same language patterns reveal that people are using different words with different emotional connotations when they’re describing those medicines, a human in the loop may be able to discern that it’s the emotional states and cultural nuances that actually differentiate their product.
And in that case, the key to driving more sales lies in aligning brand creative and demographic targeting with the hearts and mindsets of target customers.
3. Cultivate the courage to act
How many times have you lost an argument with a friend or loved one, even when you knew with mathematical certainty you were right? There are moments when logic is simply not your friend — and those moments are a daily reality for marketers.
Having the right data isn’t enough to drive action. Many of an organization’s most important decisions, particularly ones about marketing and brand, rely on trusted relationships between people.
What’s more, as a recent Deloitte report reveals, a growing number of companies who’ve invested in big data and data analytics solutions still don’t trust their data enough to act on it.
Let’s say you’re a car company that’s looking to drive sales in a specific automotive category. Marketing analytics may reveal a huge opportunity to differentiate and drive sales in your category by incorporating certain components into the car interior that you’d only expect from an adjacent category — say, luxury interior components in an otherwise economy-class vehicle.
Making those changes on the assembly line isn’t cheap and may fly in the face of your brand identity. The data says make the change. But your gut says, wait a second, is this the right call? It’s hard to trust numbers on a page.
However, if the people in charge of brand and product strategies are themselves part of the exploration of options, it’s a different story. Instead of receiving a report that says, “Rethink everything you think you know about your cars,” decision-makers now have the option of being in charge of the exploration of issues.
Most importantly, marketing analytics solutions must never claim to reveal absolute causality. That’s just impossible.
But with intuitive visualizations, marketing analytics solutions can deliver a virtual lineup of the most likely suspects of causation, then allow human intuition to assign causation and choose the course of action to pursue.
This is how you solve for the trust gap, and this is how you enable data-driven instincts and trusted relationships to land on the right decisions. And ultimately, enabling better decisions with the right combination of human intuition, informed by the voice of intelligent data, is what marketing analytics is all about.
The bigger challenge is, how do you nurture your organization’s curiosity, courage and collective will to act on what you discover?
コメント