Data hasn’t just become an essential currency for brands; it’s the key pillar supporting the entire digital advertising industry – with ad spend predicted to pass £15 billion in 2019. Information about consumer behaviour, preferences and habits are vital to ensure campaign messages resonate at an individual level, maximising conversions for brands and driving a consistently large share of publisher revenues.
But the necessity of leveraging data doesn’t make it easy to handle. In fact, global research shows improving data usage for segmentation and targeting remains a top priority for half (55%) of marketers and 44% cite gaining a complete, cross-interaction customer view as the greatest challenge keeping them up at night, closely followed by tracking effectiveness.
So what is making this fundamental of marketing success difficult to manage, and what can advertisers do about it?
Taming increasingly unruly data
It’s a well-known paradox that the key benefit of big data — its all-encompassing remit — is also the reason why managing it efficiently is such a hard task. With an estimated 2.5 quintillion bytes of data generated daily by a tech-savvy generation using multiple devices and screens, marketers are overloaded with information. They have access to stacks of first-party data about the habits, attributes and preferences of their own customers, and the ability to add more detail with data supplied by third parties. They know that, if harnessed effectively to drive creative, data can help double revenue growth. And they also know when the application is imprecise or poorly considered it can lead to ad blocking and tarnish a brand’s reputation.
But before large data pools can be utilised as a basis for impactful tailored ads, marketers must translate them into useable insight. And there are many factors that make doing so problematic. For starters, there is the tendency for collection and storage to be siloed; by the department and within isolated systems. Then there is determining how datasets can be analysed at sufficient scale and speed to enable the delivery of personalised ads in real time. What’s needed is a means of taming data and unlocking the insights it contains. And the best solution may lie in another product of the digital revolution: artificial intelligence (AI).
Machines can drive human marketing
Often talked about as the future of business — specifically analytics — AI is already making a significant difference to multiple sectors. Sub-sets such as machine learning (ML), defined as a form of AI where systems learn from experience, are transforming data management and execution – not only capable of processing data from disparate sources at scale and speed but also able to spot trends and patterns that power informed fast decisions. In fact, studies show ML initiatives have helped reduce the analytical run time by up to 90%.
The most prominent ML use case for advertising is boosting personalised creative relevance, often via dynamic creative optimisation (DCO). By assessing varied data sets — including demographic, contextual, and behavioural data — against a catalogue of creative elements, DCO identifies the mix suited to an individual depending on their activity, unique tastes, and place in the purchase path. For example, elements cover the image, background and format type that will align with the device in use, location and weather, as well as language.
But there are also other AI advances going a step further: taking marketers closer to seeing ads from the audience perspective by evaluating scenarios much as the human brain would. For instance, reinforcement learning (RL) is a form of deep ML where algorithms are taught to select the optimal course of action in a particular situation via positive and negative rewards. Unlike the majority of ML tech, it isn’t bound by specific rules and parameters; RL analyses all available options and what they could mean (good or bad rewards) using data about past decisions. So, while there might be an overall goal, algorithms decide how to get there.
In advertising terms, RL can be a valuable tool for one-to-one targeting; pinpointing which ad is most likely to connect with an individual given a previous response, and current context. The best part being that the more RL systems hit the right mark with consumers, the more precise they become: continually learning what works well for individuals and adjusting their approach to enhance experience quality, in-the-moment impact, and campaign results.
Every marketer knows the well-used data-driven advertising maxim: right person, right message, right time. But determining how they can capitalise on the ever-growing masses of consumer-generated information to achieve it isn’t always clear. Fortunately, ML technology provides the tools to streamline and harmonise the complex spider web of data. With smart systems and creative teams in place, marketers can incorporate data insight, creating relevant, connected and personal ad experiences to optimise campaign success.