It boils down to delivering the right message, at the right time, on the right channel. Photo Illustration: Yuliya Kim; Sources: Getty Images
Netflix’s Bill Nye Saves the World was keenly anticipated by nostalgic fans in their 30s who love science, and based on the ratings after two seasons, the show does not disappoint. In fact, Bill and his correspondents hash out some pretty relevant scientific issues on the show, particularly as it pertains to tech.
In one of the first episodes, the show explores the world of AI, discussing its benefits and problems. One of the more hair-raising issues brought up had to do with the fear that AI might become smarter than us, take over and kill us all, with one particular panelist reminding us about the 2015 incident where a Twitter bot autonomously tweeted out a death threat.
So, if AI is malfunctioning and, oh, tweeting out death threats, how is it being used effectively in business? Well on this end, there haven’t been any terrifying revelations (thank goodness), but in terms of getting the best out of the technology, it is still very much a learning curve.
Within the last five years or so, big data has exploded onto the marketing scene and Mar Tech-focused companies have been cooking up new ways to harness its power. The introduction of AI and machine learning has certainly shown excellent results for numerous campaigns. That is, if companies are financially fit enough to go through a lot trial and error.
Marketing automation is still not an exact science
Marketing automation at its best is not a plug-and-play kind of thing. It involves a lot of testing and retesting and while AI-powered technologies are definitely out there, they are still in their infancy and tend to require a lot of marketer supervision.
“When we talk about how to use learning and AI with marketers, we talk a lot about real-time experiences with data,” says Thom Gruhler former CMO of Microsoft Windows and CEO of Fjuri, who points out that the most important aspect of automation is the quality of the data. That is, after all, what ultimately determines how the machine learns.
According to Gruhler, machine learning initiatives are the most mature in direct response. Generally, marketers look at which current customers responded to which ads and those data sets are then used to find “lookalike audiences,” which are likely to respond similarly. From there, more potent audience groups are cultivated. This is how Harley Davidson grew their sales by 2,930 percent in three months.
Delivering the right content to the right people
Any seasoned Mar-Tech pro will tell you that success boils down to delivering the right message, at the right time, on the right channel. There are numerous platforms that claim to do this, but Gruhler points to the likes of SpongeCell and Amplero, who are using AI with the capability of automated message construction with body copy.
These technologies are dynamically selecting campaigns over and generating lots of creative messages and they tweak content based on the responses they receive—where an ad should go, what channel, which ad unit, which company and a ton of other parameters.
Marketers know that they need to get comfy with automation because it’s becoming the norm. Many are still using the old standbys Salesforce and Marketo, others are opting for some of the 5,000 newer Silicon Valley-made platforms and still others are using a combination of all of the above in the hopes of providing the best customer experience. And therein lies another scientific problem to solve: getting these programs to work together effectively.
So, more solutions are being made. Companies like Tray have tried to accommodate the need for multiple automation softwares, by providing a platform in which marketers can drag and drop whatever platforms they’re using all into one place. There are also numerous consulting companies out there whose sole purpose is to stitch the automation story together—and it’s important that they do.
As top consultants will tell you…
