I recently stumbled on an excellent blog post in which futurist Gerd Leonhard let loose with a few thoughts about how marketing will develop over the next decade. Among the many points he made, Leonhard stated:
“Companies are going to try to predict how people feel about their brand, and then adjust in real time by changing features, and starting new conversations with customers in real time. All of the companies of the future will have one big job: to make sure that the customer feels cherished and safeguarded. As Amazon calls it, ‘customer delight,’ will be the number one mission. If you screw that up, everyone will leave.”
This quote is spot-on when it comes to using real-time data to steer the conversation you’re having with customers, but in it, Leonhard also repeats a big misunderstanding regarding the purpose of the predictive models he cites.
Predictive models aren’t designed to measure ambiguous platitudes such as how people feel about a brand. Predictive models use past customer responses to precisely inform future customer interactions—a point best exemplified by looking at publishing in the social sphere, where creating customer delight is much less important than creating customer understanding.
What’s Wrong with Customer Delight?
Chasing delight can lead to shallow social messaging. Just ask any marketer who posts overly general status updates along the lines of “Sunny days sure are great, aren’t they?!” That sort of delightful message may get a lot of “likes,” but it won’t convert customers into taking meaningful actions.
Using predictive publishing to chase customer delight won’t fix this problem. Instead of relying on gut feelings to tell you what people will “like,” predictive models will let you use the data to put immediate social responsiveness over real marketing impact. As a marketer, you aren’t looking for attention; you’re seeking ROI on your actions. Using a more efficient process to super-charge a faulty approach won’t magically fix your campaign; it will just magnify your useless results.
Customer Understanding Beats Customer Delight
But just as predictive publishing will magnify useless results when improperly applied, it will also magnify meaningful results when set to a worthwhile task. Leonhard hints at the real purpose behind predictive publishing in the above quote. Aside from “delight,” Leonhard also utilizes the word “conversation” when he talks about predictive publishing, and “conversation” is a much more accurate word to use when you’re talking about predictive publishing’s communication enabling capabilities.
You see, predictive publishing operates conversationally at its core, running data on your customer’s previous social actions (likes, shares, text-based responses, timing, etc.) and uses this data to create predictive models that help you determine the best way to communicate with them in the future. Once you publish another social action using these predictive models, you will accumulate more customer data, and this data will focus your next action even further, creating a feedback loop that teaches you how to communicate effectively with your customers.
In this way, predictive publishing mirrors the dynamics of a one-on-one real-world conversation. When you speak with another individual, you take in and evaluate interpersonal data (body language, verbal cues, etc.) to assess their engagement and whether or not what you’re saying lands with them. Once again we’re talking about social feedback loops here, where you’re taking in data, evaluating it, and adjusting your position accordingly in an effort to reach shared understanding. The only real difference between the way predictive publishing works and the way a real-world conversation unfolds is that predictive publishing lets you hold an intimate conversation with countless people, all at once.
Of course, all this is a little academic and high-minded, so let’s get to the meat of the question here: How can you use predictive publishing to create such a social conversation?
Well, let’s say you have your social action wrapped up and ready to go. You want to make sure it reaches the largest audience possible. Based on responses to previous posts, predictive publishing can tell you the best date and time to put up your social action to maximize the eyeballs that land on it.
Or, maybe you have some information you want to impart but you aren’t sure whether it should come across as neutral, light-hearted, or serious. Predictive publishing can help you find out what tone your customers respond best to and whether it’s appropriate for your current message.
And here’s the big one: You know what you want to say, you know the tone you want to use, and you know when you’re saying it, and when you’re just about good to go, predictive publishing can help you fine-tune the wording on that message to ensure it really hits. Fine-tuning may sound like a small thing, but any marketer with a background in split testing understands the potentially massive impact of changing just one word in a field of copy.
With these common and powerful applications of the technology, it’s obvious predictive publishing has more to do with telling you how to communicate with your customers than telling you what to tell them. And while you’ll experience a certain amount of delight in being heard, I would argue predictive publishing ultimately offers something much deeper than that.