Last month we talked about how the Internet of Things is having a significant effect on how businesses collect data and how marketers are using it as a powerful engagement and personalisation tool. It is, without a doubt, revolutionising the marketing landscape because brands are now able to see precisely what motivates customers, meaning they can hone their products and services to tally with what matters to them.

Now marketers are able to personalise to a degree that was never possible previously, propelling marketing beyond response to anticipation. But with IoT still by the scheme of things in its relatively primitive stages, there is a degree of uncertainty hanging over how best to make it work. Yes, IoT will deliver a gold mine of big data that could potentially be used on a mind blowing scale, but how will the massive influx of performance data actually be analysed?

Deriving Insights from Abundant Data – Finding a Needle in a Haystack?

Ever tried to find that needle in a haystack? The same could be said for deriving insight from terabytes and terabytes of machine data. Is it really possible for a human to review and comprehend the intense plethora of information that is offered up, even at sample size?

In reality, for IoT to deliver on all its promises, the speed and accuracy of big data analysis needs enhancement. If it doesn’t get it, things could start to malfunction, and all those personalised marketing campaigns we have been getting excited about could, quite easily, go pear shaped.

So how can we keep up with IoT generated data and really get to the crux of the insights it has the potential to offer us as marketers? The answer is machine learning.

Deriving insight from terabytes and terabytes of machine data – it’s like trying to find a needle in a haystack.

What is Machine Learning?

Machine learning is a form of artificial intelligence (AI) that enables computers to learn without being programmed. The process involves searching through data to detect patterns, which cause program actions to be adjusted accordingly. Facebook’s News Feed for example makes use of machine learning to personalise a user’s feed. Say you regularly pause scrolling to read or like someone’s posts; your News Feed will, as a result of what it’s learnt by your habits, start to show more of this particular member’s activity earlier on in your feed because it detects that you have an interest in it. It’s the same when your email program learns which folder you usually move mail from certain recipients to, or the predictive text on your mobile phone that suggests the words and phrases you regularly use in your messages: it’s all done through machine learning.

Gartner predicts that the AI world will have a ‘head-on collision’ with the Internet of Things this year and reckons it makes ‘total sense’ that the apps controlling IoT devices will also make use of machine learning. By 2018, says Gartner, AI will be feeding some 6 billion connected ‘things’.

All very interesting. But can machine learning be used to enhance marketing campaigns? And will it change the digital marketing landscape? Absolutely yes on both counts.

Facebook’s News Feed, email programs and predictive text all use machine learning.

Machine Learning in Digital Marketing

You may not realise it, but machine learning, albeit in a rudimentary form, has been in use for many years. How long have you been using the spell check tool in your word processing program? These functions have been using machine learning for decades to identify grammatical and spelling mistakes. Of course the technology has since evolved. More recent examples include online product recommendations. These use algorithms to suggest products to web users that are aligned with what they’ve previously searched or bought. Google’s search browser is another example, frequently using machine learning to understand searcher intent and to make sense of search queries that are marred by spelling mistakes.

Going back to our questions in point, because machine learning analyses consumer behaviours, it can determine many things that are of valuable use to the marketer. For example, things like when email or online ad delivery is most likely to engage and convert. It’s all about putting data in context, which up until now has been something of a challenge. After all, without context, it is impossible to derive meaningful insights.

In this way, machine learning fills a gap that has proved problematic for marketers for some time.

Brand Story Telling Made Easier

InsideBIGDATA, a predictive analytics publication, is of the opinion that machine learning will revolutionise online storytelling strategies. Currently, it is possible to identify what audiences want to hear, but as marketers we still have to build the story ourselves. However, with machine learning, this could well change. Machine learning could, in time, control not just subject matter, but in addition the right tone of voice, the right medium and other vital elements of brand story telling.

In time, machine learning could form an integral part of interactive content, enabling the creation of a truly personalised experience for audiences. And as we discussed in our IoT post last month, personalisation is a powerful influencer.

Machine learning has the potential to revolutionise online storytelling strategies.

These are exciting times. For the digital marketer, there is increasing scope to significantly make an impact with personalised campaigns backed by a profound understanding of what motivates their audiences. The new trends sweeping the digital landscape may feel a little overwhelming, but they really must be embraced if a business is to purloin the competitive edge.

Need some guidance in utilising emerging technologies to meaningfully enrich your digital marketing campaigns? Talk to Figment!