To win customers, you must stand out. But it’s not the product that should stand out – it’s how you spread the word. Many of us have had great ideas, like a new killer app that’s going to change the world. We set up a website, invest in building the product, and wait for the customers to come – and no one does. Search engine optimization is the first hard-learned lesson, followed by email marketing (hot and cold), and then social media marketing and the battle between inbound and outbound marketing tactics. But it all boils down to the fact that the best ideas go unnoticed without proper marketing, and marketing is the sole process that generates revenue for your business.
So we need to get marketing ‘right’ if we want to have any chance of succeeding, but the problem is that marketing is a process that evolves and not a one-time task. We must constantly measure what works, and what does not. Also, personalization is super-important as we are dealing with people, not machines. People have their particular needs, wishes, and fears. In order to sell something, we need to listen more and talk less. Without this, marketing is a waste of money and effort. No one has endless cash to spend on marketing tactics, so we must be intelligent in how we direct our marketing efforts – which brings us to AI.
So, how can AI help?
- By creating profiles, it can find similarities between converted users and not-converted users, thus letting you know which leads you should focus on most (recommendation).
- With content generation, it can create personalized content for every client (even though you are referring to them in bulks).
- By combining and surfacing a number of different parameters, it can give you a holistic view on clients and their expected behavior.
- By automating repetitive, labor-intensive tasks, it can free up your time to focus on what you do best.
Why AI, and What Makes It Different?
Programming everything, especially when it comes to human nature (which is the most unpredictable of all) is a nearly impossible task. That is why we needed to create machines that would learn by themselves – and thus we have AI. AI does not understand what it does, but it does what it is trained to do exceptionally well. So, how do we create machines that learn?
The simple path involved statistics and probability – we ‘trained’ the algorithm by feeding it large amounts of data, then used that algorithm in our predictions. This is a simplified version of what we do in Machine Learning. Deep Learning is a bit trickier, though.
The creation of Deep Learning is motivated by the way the brain and its neural networks work. But to simplify things, imagine an enormous machine with millions of knobs. Each knob affects the output of the machine, and with the right settings, this machine can do just about anything. But the beauty is in the training part: we do not need to know all the possible settings – we can let the machine train itself. Using backpropagation, this machine can adjust its knobs based on the data that is fed to it. Thus, in Deep Learning we have a trainable machine that, given sufficient data, can learn to detect patterns and similarities. This is how a number of techniques, including face recognition, work.
Deep learning is especially useful for unsupervised learning, where one cannot really even describe a problem mathematically to the computer.
AI in Action
According to Avi Goldfarb, when the price of coffee falls, people are likely going to buy more of it. That is a simple prediction, but we can also predict that people will buy less tea, and that they will buy more sugar and cream. Looking at multiple parameters and finding correlations are important in building accurate prediction models and finding what matters most. This is what makes Netflix’s and Amazon’s recommendation engines so powerful.
But beyond prediction, we also have the issue of action. Action means actually making use of a prediction, to drive a certain behavior in prospects. Here is another area where AI can do wonders – by creating a prediction and action loop, AI can continuously improve its marketing approach. RTB House uses this to refine the ads displayed to users – if the user did not engage, it will change the recommendation on the ad. Its deep learning algorithm knows how to forecast, adapt and recommend the right ad at the right time – leading to improved click-through rates of up to 41%.
While AI is a time-saver for businesses, it also raises the level of professionalism. Because the AI has been trained, the people who use it are benefiting from a system where the proper approaches are built in. “People doing their own marketing is king, but people don’t do their own marketing. People aren’t entrepreneurial, or educated like that,” says Lodewijk Veldhuijzen, founder of Anything App, who’s focused on a solution where anyone can monetize their time over the phone while machine learning provides visibility and connections between the right parties. “The search engine learns which queries are often used in which regions. Which profiles are receiving clicks, but not calls. Or which ones are receiving calls. Your tendency to go for a specific price range. Your tendency to call close, or far away. Who people like you are calling. Continuous learning will mean we can accurately show things that should interest you, considering your demographic and geographical location.”
PersonaPanels approaches the topic from the opposite angle – while the other methods focus on AI techniques to do your marketing, PersonaPanels uses AI to see if those tactics are actually working. It does that by having an army of simulated customers interact with your website and show you the result. The models are based on how real-world visitors tend to interact with websites. So instead of relying months on A/B testing and collecting the results, you can just make the change, have the system evaluate it, and then use the result that worked best in the real world.
What Path to Take?
Going back to our ‘killer app’ example, we now have a tool that makes our marketing smarter. AI is not another layer of effort or expense – rather, it’s about making what you were already doing smarter, faster, and much more cost-efficient. AI is not black magic, but rather a necessary ingredient that lets you focus on the right prospects while automatically making you appear more professional and relevant. After all, it is a machine that has learned by observing thousands of samples – more than most of us humans do.