Get Across AI Or Get Left Behind

In this opinion piece, software business Station Five CEO Lambros Photios tackles artificial intelligence (AI) in marketing.

With the rise of AI and the drastic change it’s going to bring to marketing, Station Five Founder Lambros Photios warns we’ll need to stay on top of it – or risk being destroyed by rivals.

Data driven decision making is a term we’ve all heard far too much this year. Those implementing data driven decision making are ahead of the curve and on the fast track to success. It’s scientific, it’s statistically backed, and I want nothing to do with it.

For some marketing divisions, this will come as a shock. However, data driven decision making is slow, cumbersome, and quite frankly boring.

But what’s better than a modern approach being implemented by startups, enterprise, and everyone in between?

First, a brief explanation; data driven decision making at its core is a methodology whereby data at our fingertips is used to drive decisions.

This is compared to assumption based decision making, whereby an individual will make assumptions of their target market (or personas) to make decisions. It was very logical to move past anything assumption based, but it isn’t bulletproof.

Data driven decision making is ultimately limited by the marketer’s ability to compute the data available to them (using Excel, or similar).

But how much data can they compute, even with these tools? How long does it take to compute data that could be driving decisions in real time?

What happens to a business when the learnings of an audience are contained to an individual?

Machine Learning: Data driven decision making on steroids

These are all problems that can be resolved with the use of machine learning. For those who’ve kept far away from robotics, automation, and artificial intelligence (with fear I, Robot will become a reality, and Will Smith won’t save us all), I’ll start with a brief explanation.

Machine learning is a sub stream of artificial intelligence enabling software to “learn” from the data it processes instead of just executing basic “rules” against the data.

Just like we learn about the personas we’re marketing towards, machine learning enables a system to learn about its marketing personas. The key difference is that the data a team of analysts processes in a whole day can be processed by a system within a fraction of a second.

Before we push forward, I acknowledge that machine learning is still technically a type of data driven decision making.

The assumption in my terminology (for the sake of simplicity) is that data driven decision making is best used to describe human based decisions. With that out of the way, let’s push on!

Example: The Chess Masters

In December 2017, Google created a machine learning algorithm to learn and consequently play chess.

The algorithm, called “AlphaGo Zero” (the name doesn’t need to make any sense to us, they’re Google!) initially defeated the world champion of the Chinese board game called “Go” in 2016.

With some repurposing, and only four hours of training, the algorithm played chess against the world chess master in 100 games, winning 28, and drawing in the remaining 72.

Coming back to reality; to say this is a bigger paradigm shift than the implementation of data driven decision making would be the understatement of the century.

The issue is that the new technology is out there, but very few organisations are doing anything about it! Those who are using machine learning are seeing huge benefits.

The How: Machine Learning for better marketing results

But how can machine learning drive marketing strategy? I was inspired to put pen to paper after reading the B&T piece by JWT’s Zeina Khodr, who was pointing out the importance of artificial intelligence pertaining to the content experience.

Zeina pointed out that AI still gives many marketers the “creeps”.

This attitude worries me for the sake of these people’s future results, as an open mind (and some degree of cautiousness) can be used to achieve results that are inhuman (literally).

Zeina declared “most are not prepared for how big a game changer this is going to be, and understand very little about the impact AI will have on their business as a whole, let alone their marketing”. I couldn’t agree more.

I find all technical problems are better explained with an example, so here goes!

Situation: I execute the internal marketing operations for a fashion business. The target market includes both men and women aged 18 – 35.

Data Driven Decision Making: The target market is broken down into personas based on 2-3 year age groups (18-20, 21-24, 25-27, etc), gender, location and profession.

I can then class these small groups into user personas. When we combine all variants (for example, 21-24 year old women living on Sydney’s Northern Beaches working in hospitality), we are left with a huge number of groups.

These all need to be understood so the most effective campaigns can be pushed towards individuals within that user persona.

Machine Learning: Jane is a 22 year old woman living in Manly working at Ricky’s, an up market restaurant. Last weekend, Jane purchased a new pair of jeans, and since then has been searching various online stores for a matching jacket before her trip to New York in December (it’ll be cold at that time of year).

From her search behaviour, we know she is willing to spend up to $250 for the jacket. She has an AfterPay account, and we know from her purchasing patterns that lifestyle shots are more likely to lead to a transaction. An AfterPay logo will be watermarked into the lifestyle shot of a $400 jacket, produced as an advertisement for Jane, and distributed on Instagram from 4:00pm to 4:30pm (while she’s on the bus to work).

The jacket selected matches the jeans she previously purchased, and has intricacies subconsciously reminding her of a jacket she owned when she was younger. This was all completely autonomous, and Jane didn’t tell us any of this directly.

With the linking of data sets across social, email and search history (amongst others) this is all very possible with current technology.

Despite this seeming very 1984, some businesses are creating huge benefits for themselves by actively leveraging these strategies. This is the next evolution of marketing, and I’d encourage you to actively take steps in this direction before your competitors do.