I’m always interested in reading articles about new approaches in my field of artificial intelligence and machine learning, so the recent AdWeek post by Steven Wong and Derek Slater of Ready State suggesting that ‘Machine learning is about to turn the marketing world upside down’ really caught my eye.
Every time I hear or read that machine learning is ‘about to’ do something it makes me smile, particularly in this case as ‘the marketing world’ is completely relevant to my job. As Chief Science Officer for an international marketing company, I am working with these technologies on a daily basis. And in my view, if your marketing solutions aren’t already using machine learning and artificial intelligence then you’re already behind the curve.
Unlike the 2017 scene set out in the original post, more marketing companies are having conversations about human designed personas compared to segments identified by machine intelligence. While correlation isn’t necessarily causation, understanding the similarities in your customer base that lead to a purchase decision is powerful.
There is so much data available on everyone that it’s almost impossible to discern good models using human intuition. If your company is lucky enough to employ talented data scientists then you will probably be getting some good segments, finding new relationships in your data and seeing improvements in engagement in response to targeted campaigns. Machine learning can take this much, much further.
Without a bias on the weight of any piece of data, machine learning can find connections within the data that a human could not see. Rather than a few extra segmentations, you could define hundreds. Instead of ‘if A then B’ predictions, multidimensional data relations can be uncovered, giving far more confidence in the likelihood of engagement and testable results that can improve the engine.
This isn’t science fiction – we are already at a point where marketing applications can and do handle this for you, and it’s up to you to decide how to use the information.
Wong and Slater define machine learning as currently in the initial phase, where it substitutes for repetitive tasks, and their piece focuses on programmatic advertising. My team and I are already well beyond this and moving into the augmentation and modification phase. I’ve had regular discussions where we look at tasks that not only would be beneficial to automate (data analysis, first line customer service etc) but also could provide those ‘wow’ moments.
In my opinion, the article isn’t accurate when it states that machine learning hasn’t ‘hit’ creative functions – there are reports of artificially generated news, and websites are already optimised for gaze detection to place adverts appropriately. I believe this push will continue and it’s important that agencies are prepared. But whether machines will ever achieve the levels of creativity and imagination we achieve as humans is open to debate.
Wong and Slater are correct about the evolution of the technology – it’s moving very quickly and becoming both faster to get results and able to do more complex tasks, so what was state of the art a year ago may quickly become out of date. It’s important to choose a technology partner who is focussed on continual innovation, so you get all the benefits as they become available.
Just as the advent of digital media required a change in focus for marketing agencies, the same is true with machine learning – embrace it and make it work for you.
By Dr Janet Bastiman, chief science officer at SmartFocus
PrivSec Conferences will bring together leading speakers and experts from privacy and security to deliver compelling content via solo presentations, panel discussions, debates, roundtables and workshops.
For more information on upcoming events, visit the website.
comments powered by Disqus