Despite the fact that digital marketing is so fast-paced, there are some processes that simply cannot be rushed. Understanding the insights gleaned from big data and using them to pinpoint your effective targeting channels is one of them. But perhaps not for much longer. Peter Mason, CTO at Addition+, explains how machine learning is about to change the way that brands compete for the consumer’s attention, with an accuracy and cost per action (CPA) not to be sniffed at.
Having been awarded an entrepreneurial scholarship on a programme run by Goldman Sachs, Addition+ started life helping brands and publishers work together in a more targeted and efficient way through technology. “A big part of our early work was helping both sides to trade more effectively using technology, particularly contextual technology,” explains Mason. “This led us to where we are today, with a team that are specifically specialised in using information retrieval technology”.
Information retrieval, the process search engines use to categorise and gather search results, would prove to be key in providing Addition+ with the tools to create its product. “What we were doing was using information retrieval algorithms to identify where the advertising was performing best” continues Mason. But this new strategy for identifying the key performing areas for advertising, presented the opportunity for further innovation. “Human beings cannot follow and track and make decisions on the best context for an advertising message, it’s far too much data. It’s terabytes of information. It’s also not linear, the predictive process requires big data methods.” It’s no longer enough for marketers to assume their audience will be found in one particular segment. “This is why advertising online is often so inefficient. To increase reach, marketers will layer sites and audience segments based on what they think will perform best.”
A Shrinking Pond
The result of this is that competitors are all campaigning for the consumer’s attention within the same space – a system that either rewards the brand shouting the loudest, or drives the potential lead away entirely. When everyone fishes in the same pond, nobody wins. This is where the machine learning comes in. “We’ve built a machine system that doesn’t try to figure out upfront the type of context that will perform best – the campaign learns based on what it sees happening in real time.” The machine learning algorithm draws insights from the campaign that is running in an optimised segment, identifies unique similarities and triggers, and uses these to identify potential new audiences. The ability to pinpoint the most responsive segments means machine learning can expand campaigns with a heightened level of accuracy, keeping the CPA low but greatly increasing the reach.
Spotting The Difference
A key feature of machine learning’s success is the removal of the inherent human bias which impacts the decision-making of marketers. The AI algorithm is able to identify trends and patterns which are imperceptible to a human brain, and use these insights to target similar audiences in real-time. Giving a machine the power over that core decision-making process removes the human ‘middle man’ and makes sure that the data and the conclusions drawn from it do not become disrupted.
“If you’re putting real-time data in the hands of people to make campaign decisions, you’re still just as hindered as you’ve always been,” says Mason. “They simply can’t adapt quickly enough and make the requisite number of changes to a campaign, to increase its performance.” In a market full of ‘lookalike’ targeting and segmentation, many brands find themselves operating in the very same segments as their competitors. “It’s a human being making the decisions, and we all come to the same logical conclusion about what’s likely to work best,” says Mason. “What machine learning allows you to do is see what’s really working, and then scale very significantly, effectively, and discover much wider audiences that are unique to you.”
Getting To Know You
So where does this leave the modern digital marketer? What kind of opportunities are created through this newfound efficiency and resourcefulness? It seems that the introduction of machine learning into your marketing stacks frees your team to focus on that other vital player in the match: the consumer. “The more you understand about what drives your customers, what interests them, what they like to consume, and the quicker you can adapt to that information, the better your relationship is likely to be,” says Mason. “It’s all down to this somewhat overused statement of targeting the right people, at the right time, with the right message. And this is truly moving towards that.”
By keeping on top of the constantly changing requirements of your brand’s audience, you can make sure that you are going the extra mile in order to communicate with your consumer. Identifying the needs of your individual segments has long been a gargantuan big data challenge, but with the help of machine learning, marketers can focus much more on the human element; interpreting the data insights, and improving the overall experience that people have of your brand.