Notice: Trying to access array offset on value of type null in /home/marketdiag/public_html/wp-content/plugins/really-simple-facebook-twitter-share-buttons/really-simple-facebook-twitter-share-buttons.php on line 318
Warning: count(): Parameter must be an array or an object that implements Countable in /home/marketdiag/public_html/wp-content/plugins/really-simple-facebook-twitter-share-buttons/really-simple-facebook-twitter-share-buttons.php on line 514
Eric will present “Influence in the Machine Learning Age” at CRC, October 22-24 in Orlando.
In the classic sequel, Back to the Future 2 (1989), Marty McFly travels into the wild world of 2015 where he discovers that skateboards have evolved into hoverboards. In 1989, that’s what Hollywood thought our future might hold. But flash forward to 2019, and many things look more or less the same. Despite the prediction of airborne kids, we’re all still watching TV (lots of it), driving cars, paying to go to the movies, stressing out over our daily lives, waiting in line at the airport and the DMV, and riding skateboards the way we always have – with wheels on the ground. When you look at it that way, nothing’s really changed.
Of course, that’s a small part of the story, because everything has changed – more specifically, technology has changed everything. We are in the midst of an ongoing digital revolution that has shifted, and will continue to shift, our relationship with the world around us. We are at a critical inflection point where the ‘Math Men’– the owners of data and algorithms – are sitting around the same table as the ‘Mad Men’ –the owners of the creative process. We are in the midst of a new machine age that is fundamentally changing consumer expectations, including the way we both create and consume content. Storytelling of all kinds will never look the same again.
Machine Learning in the Digital World
One of the greatest fields of innovation of the 20th century is machine learning. Simply put, machine learning is a branch of computing that trains systems to perform different tasks through algorithms that learn from data and from examples. Many of marketing’s current buzzwords – terms like artificial intelligence, virtual reality, augmented reality, and personalization – are built on the shoulders of machine learning. They wouldn’t exist without it. More than ever, machine learning is a really big deal.
As it applies to content creation and advertising, the great promise of machine learning is clear: Automation (e.g., programmatic buying), personalization (e.g., dynamic banners, videos, or experiences), and creativity (e.g., machine learning shifting how content, including ads, gets made). The greatest indicator that machine learning is here to stay lies in the financial numbers. We have moved from theory to scalable revenue-generating solutions for machine learning in a very short time frame. Sources estimate that the Machine Learning as a Service (MLaaS) market size is estimated to grow from $900 million in 2017 to $8.3 billion by 2023. While most industries will eventually see machine learning impact their top and bottom lines, today it is advertising, which represents one of the most proven paths forward.
It all started with games?
On the surface, it’s easy to point to the ad industry as a place where machine learning is already in play. As digital advertising is starting to surpass ‘traditional’ advertising in terms of spend, we are finally seeing what all the fuss is about. The holy grail for the ad industry these days is personalization as connected to attribution. This is the process of delivering highly relevant and valuable communication to customers, and then understanding in great detail the impact and results of those efforts while also gleaning new insights. Recent studies have shown that targeting and personalization can have a significant impact on advertising effectiveness, enabling savvy marketers to reach the right people with the right message at the right time. But all this is really just the tip of iceberg. The field of machine learning is getting far more sophisticated, with hints that data scientists can work around the same table with creatives on much more than just search or display advertising. Just how sophisticated has the field gotten? As it turns out, the answer to that question has more to do with gaming than it does with advertising.
One of the most famous examples of this kind of technology took place in 1997 during the chess match between Deep Blue (an IBM Machine) and Garry Kasparov, the world’s best chess player at that time. IBM’s Deep Blue was able to search through 200 million possible positions per second to determine the best move. And Deep Blue won in a matchup that became a significant cultural moment for computer science. However, much has changed since 1997 and chess, while a complex game, has a finite number of possible moves, very unlike the real world. Fast forward to 2016, and researchers had moved on to a new fascination – the ancient game of Go. Whereas Chess is a clear game of logic, Go can most accurately be described as a game of feel and intuition. In Go, there are more possible variations of a game than there are atoms in the Universe, making it impossible for any machine to compute all potential outcomes. Can a computer master intuition and creativity? Researchers at Google’s Deep Mind thought so, thus they built a machine called “AlphaGo” that took on Lee Sedol, the world’s best Go player in a series of 4 matches. AlphaGo beat Sedol 4 to 1.
Similar to Go, in the real world there are an infinite number of imaginable outcomes, in addition to the uniqueness of human error. The idea of a machine accounting for everything seems impossible and foolish. But that’s exactly like the challenge facing autonomous cars in 2019. In order to ‘train’ these cars, you have to get them out on the road to learn. With every trip, the machine becomes a better driver. The machine eventually becomes a better, safer driver than a human. This is could very well be the future of driving. Get ready for it.
Implications For Creativity: Advertising, Brands, and Experiences
If machine learning is getting this sophisticated, what does this all mean for advertising and storytelling? Way more than you might think. In a vast oversimplification of the traditional advertising creative process, insights are generated through research (planning), creative is developed and tested (writing, art direction, production, copy testing), and media is planned and bought. The exotic video shoot happens, and then – BAM – you’re on stage at Cannes. But this type of creative linear model is totally upended when considered through the value proposition of machine learning. Imagine for a moment a creative process with an infinite number of possible narratives and experiences, fully automated and intelligently executed. Practically speaking, what does evolving toward this new world mean, and what are its implications?
This may not be a very popular opinion, but there is a growing school of marketing professionals who are realizing that people might not love brands as much as we’d hope – and this trend is growing. For example, a recent study shared by L2 showed that the percentage of affluent consumers who can name their favorite brand has declined significantly across industries. Another study showed that no one would care if 77% of brands died overnight! The truth is, people don’t actually think about brands as much as marketers do. We all have real life to worry about. Trivial things like family, health, financial concerns, political conflict, personal growth, and love. It’s becoming evident that what customers really want from brands is for them just to work fluidly so they can experience what really matters to them. Hence, Amazon Go is born. As of March of 2017, Amazon Go wasn’t even a reality yet. However, by October of 2018 reports were calling it “the future of retail.” Amazon Go is made possible by machine learning – and it’s not going anywhere.
Incidentally, this behavioral change is a driving force behind the growth of connected devices like Alexa and Google Home. 20% – that’s 1 out 5! – Google searches are now voice searches, and that number is climbing. This trend toward ambient experiences is evident not only in how we shop and search, but in how we choose to connect and optimize the world around us. People want technology that just works without being told explicitly what to do. And if we have to instruct or ‘teach’ technology, we want that process to be as seamless as possible. Who is willing to wait more than 5 minutes for an Uber to arrive? Even the act of ordering one feels like a pain – why doesn’t Uber just know? That’s why technology titans like Google, Microsoft, Amazon, Facebook, and IBM are investing heavily in location-specific experiences powered by data and machine learning. The commercial implications are vast. Imagine a not-too-distant future where customers enter any retail environment, interact directly with products or content they want, and then just move on with their lives. What this really means for brands – let alone for advertising – is very much an exciting work-in-progress.
Storytelling as we know it is forever changed.
Shopping is great, but what about content creation? Can machine learning be applied more directly to guide the creative process itself? Early indications suggest an emphatic yes. The same technology that powered Alphago beating the world’s best Go player inspired filmmaker Oscar Sharpe to build a machine that could write screenplays based on the input of thousands of scripts and texts. Again, this involves the same mechanisms of machine learning discussed above. The resulting project, entitled Sunspring, is just about as weird and awkward as one might expect. Of course, Sunspring is on its way to becoming a cult classic.
A year earlier, in 2015, a creative planner from McCann’s Japan office (Shun Matsuzaka) set himself an ambitious task called the “creative genome project.” His goal was to build the world’s first AI creative director capable of directing a winning TV commercial. Matsuzaka and his team collected a database of highly-awarded Japanese ads for the machine to learn them, and then worked with Mondelez to capture all the mandatory creative elements that needed to be conveyed. They created two very different ads – one developed by AI and another developed by an esteemed art director – and put the two spots up for vote in a public competition. The human-directed ad ‘won’ by a narrow margin (54%), but both spots worked. And, frankly, the AI-directed ad is far more entertaining. This isn’t the last we’ll be seeing of machines in the ad industry.
Machine Learning has also been used to create content that gives consumers that “just for me” feeling. While Netflix’s Bandersnatch has recently been grabbing headlines, the idea of using artificial intelligence to create choose-your-own-adventure type content is not new. In 2014, for example, the filmmakers called “The Daniels” released a project called Possiblia that let viewers switch to alternative narratives broken down by the second. The short film tells the story of a couple breaking up, except that the viewer can influence what paths the couple goes down frame by frame. Enabling this level of interactivity meant that there were more alternate timelines of this 6-minute film than the number of seconds since the Big Bang!
Finally, the math men are building machines that actually get sentimental, too. An Israeli startup called Beyond Verbal has built machine learning technology that is able to determine people’s emotions and state of mind just by listening to them speak. Anyone can see this in action with a very famous interview clip featuring Steve Jobs. Considering the trajectory of voice based engagement, it’s not difficult to imagine applying this to any type of entertainment experience. A world where we experience dynamic, automated content based on our personal, individual states of mind is not as far away as we might think.
So, are the robots really taking over? That might be overstating things a bit, but there is no doubt that the machines we’re building are getting more powerful, stronger, and more insightful. In a recent New York Times article, Steven Strogatz (a prominent mathematician) hypothesizes that a future evolution of AI – he dubs it AlphaInfinity – could possess intelligence at a level that humans, try as they might, won’t be able to fathom. Strogatz suggests the possibility that such a super intelligence could (quite easily) solve all our human problems from healthcare to global warming to flight schedules – and that maybe, just maybe, we’d all sit back and simply marvel at this incomprehensible intelligence. Internalize that for a moment. Now, just sit back and prepare yourself for when it happens. We’re in this together.
NewsBusinessEric Solomon – Blackbird