Digital Company news South Africa

Making the most of marketing insights with machine learning

It's been almost a year since OpenAI's large language model, ChatGPT, was released for public use. Since then, the world has become well and truly enamoured with all things artificial intelligence (AI). While many of us believe that this is our first real chance to experience what AI has to offer, the truth is that AI - including subsets of this technology, like machine learning - has actually been around for quite some time.
Charl Fourie
Charl Fourie

In fact, machine learning is what powers the chatbots we interact with when placing customer service calls, it is what enables the predictive text that corrects our typos and it is even the foundation of the TV show recommendations we get while browsing through our favourite streaming service.

For modern marketers, machine learning can be used to make sense of vast amounts of customer data so that they can better understand buyer behaviour, identify patterns and make more accurate predictions, says Charl Fourie, head of business intelligence at Sprout Performance Partners. The major benefit here is scope and scale.

“By teaching a computer system how to make accurate predictions based on the data you feed it, marketers can gain important insights in a way that simply would not be possible if a person, or even a team of people, were analysing the information manually.” As such, there are significant efficiency and accuracy gains – as well as time and cost savings – that marketers can enjoy by embracing machine learning, he adds.

Unfortunately, many marketers are not using the full power of these algorithms, which means that they aren’t tapping into the insights hiding in the vast amounts of consumer data they collect, says Fourie. There are a number of reasons for this. For starters, if the data isn’t of the right quality, if you don’t know how it is structured and stored and if you don’t understand what data you do and don’t have, your machine learning investments won’t bear much fruit.

“If you want the data you have to deliver any real value, it needs to be cleaned up and consolidated so that it actually tells you something useful,” says Fourie.

From what he has observed, he further notes that some of the bigger brands, with deep pockets and large teams of data scientists, are using machine learning to their advantage but the majority still have a long road ahead.

For smaller businesses, Excel spreadsheets might still work well enough but as soon as the amount of data you’re trying to process gets too big – and your Excel spreadsheet starts crashing – it might be time to look to newer and more advanced technologies. Luckily, many of the tools on the market are user friendly enough for someone without too much technical knowledge to get by. For example, if you’re already using Google Ads, feeding this data into a Google algorithm is simple enough. He believes that problems arise when you try to do something more advanced, like add data from various social media platforms, things can get a bit too complex for a marketer with minimal tech understanding to handle.

For those looking to experiment with machine learning, here are a few tips.

Understand your data

“This may sound fairly obvious but if you don’t know the who, where, why, what – who has access to it, where is it, why are you keeping it and what is it – your data won’t add any value to your business,” notes Fourie. Before you approach an agency or third party provider to help you on your machine learning journey, take the time to audit your data so that they don’t have to waste time searching for information and tidying everything up.

Identify use cases

“It’s a bad idea to bring in machine learning just because of the hype. Conversely, if you understand what machine learning offers and what it can do, you’re better positioned to use it for your unique needs and challenges.” For example, with the end of third-party cookies on the near horizon, marketers won't be able access the same amount of customer data that they did in the past. If you’re a marketer who uses third party cookies a lot, it makes sense to get your own data in order before this long relied on data mining resource is no more, he advises. Similarly, if you’re looking to deliver more personalised experiences, machine learning enables you to provide more relevant and customised content, deals and recommendations.

Just start

In my experience, says Fourie, people are worried about technologies like machine learning because they don’t understand them. And because a lot of companies still don’t fully grasp the potential, they are afraid to get started. But it’s important to experiment so that you can figure out how they can add value and how you can incorporate them into the work you do. “These tools aren’t going anywhere,” he concludes. “So if you don’t get stuck in soon, you really do run the risk of being left behind.”

About Sprout Performance Partners

Sprout Performance Partners stands at the intersection of partnership, digital evolution, and data-driven brilliance. Our mission is to foster genuine collaborations with businesses, elevating their digital marketing maturity to unparalleled heights. By harnessing the potency of advanced data enhancement, we craft bespoke strategies that propel brands forward. With Sprout, it's not just about marketing; it's about growing together, ensuring our clients not only navigate the digital realm but dominate it. Your success is our shared journey.

Let's do Biz