Here are a couple of ironic mistakes that marketers make: they put too much trust in careful iteration, and they put too much trust in ‘human creativity.’
What do I mean by this?
Well, let’s start with careful iteration.
Marketers assume that the ability to A/B test content makes them scientists, edging closer to a deep truth about their audiences. They assume that the process of analysing audience reactions to the brand’s owned content will give them an in-depth picture of that audience: that they can A/B test their way to a meaningful relationship. This is a mistake; because you can only learn more about the hypothesis your experiment is designed to test.
If your experiment is designed to test only within the narrow parameters of your product, category or competitors, then you are really testing only a very limited micropart of that audience’s interests. For example, if you are Oral B and your content talks about toothbrushes, those people can only ever really give you feedback based on content about toothbrushes, which is a very narrow picture of who Oral B truly is.
Understanding an audience from their interactions across social platforms and the web over time, can give you a more robust and (helpfully) surprising set of insights into your audience. It can identify what they actually demand: the topics, creative formats, image types, and influencers that resonate with them. And it can do this without the huge financial and reputational risks that come with treating your audience as a wall and throwing content at them to see what sticks!
Analysing your audience from their interactions with a broad range of third party content puts marketers in a very powerful position to contribute and lead in the narratives that matter to a specific audience. It allows marketers to embrace an audience-centric, rather than product-centric, approach and to substantiate that approach with specific ideas they know are in high demand from their audience. Most likely the marketer would never have thought of exploring these topics or using these creative formats in their branded content, or would never have been able to justify the rationales for doing so.
Here’s the cruel irony of A/B testing content: it creates a data-driven rationale for making creative decisions, but does so on terribly narrow and unhelpful data – your own!
Using AI and machine learning approaches to understand audiences from their interactions with third party content, brands can think creatively and strategically about the right way to insert their brand into content they know is actually demanded by the audience.
Part of the problem here is a legacy one. It’s a matter of perception about what data can and should be used for. Marketers have traditionally assumed data can be used purely to target audiences, i.e. data has been focused on the question of buying eyeballs after you have made content, and whose eyeballs to buy, and when to buy them. The proliferation of ad tech companies around the world has only added to the orthodoxy. With the recent advances in machine learning and AI, it is possible to do much more.
Big data and AI can answer the question ‘What Content Should I Be Making?’ by identifying the creative aspects of content that resonate with an audience from their interactions across platforms and a diverse spread of types, forms, styles and approaches to content.
To embrace this approach, marketers can’t use the same old targeting data that Google and Facebook help them access. Different data is useful for different things. If you’re trying to drive an e-commerce transaction through display ads, Google’s data is great. It’s based on searches people have made and reveals a specific intent that they displayed at a certain point in time (whether they were searching for a toothbrush, a holiday or an AI tool to help them make better content marketing).
If you are looking to make an impact at the other end of the funnel (and affect product or brand awareness and loyalty) and you want to use content to do that, then you need to understand the audience’s demands and desires from their wider interactions with a broad base of content. We call this ‘Empathy Data’ – to get an emotional picture of what a particular audience cares about, so it can drive a creative output for a brand or publisher.
That is where data is going to be valuable and exciting for companies this year and beyond. When they recognize that data can be used to make better creative decisions, and that better creative decisions can drive awareness and loyalty, then it’s not purely about using data to drive transactions or to force people to watch your content: and that’s progress.
So what about that second mistake that marketers make: trusting in human creativity too much?
First, let’s set the terrain for this debate.
Machines aren’t out to take jobs away from good, hard-working marketers. They are there to support, enable and enhance human marketers and to ensure that they deliver audience-centric, audience-valuable content. The problem is, in our reverence for human creativity we have retarded the development of AI and machine learning solutions that would, ironically, help us do our best work.
Of course, some industries have shed themselves of the burden of this debate and got on with creating better audience experiences. The social betting and gaming space has a proven history of using sophisticated data science to offer better, more personalised experiences for people using their products and services. The insurance industry also has developed some slightly concerning, but impressive, uses of data science to set decisions around coverage and policy applications, with mortgage providers starting to follow suit.
Sadly, the industries that need to up their game are the ones that have relied, falsely, on the supposed divide between human creativity and computational AI and machine learning. Marketers and publishers have been the most obvious examples of this. They have clung onto an image of the ‘Madman’ human genius, and this has led to a slower rate of adoption of machine learning, AI and data-driven tools in the space.
They have relegated data to be used for content distribution, and in pursuit of the goal of forcing audiences to watch content that brands and agencies have guessed they wanted. Instead, as the new breed of marketers are realising, AI and machine learning can be used to make the right content in the first place and to deliver a better, more meaningful brand-audience relationship.
We must not labour under these mistakes anymore. We must face facts. There has never been as much competition for an audience’s time and attention as there is now. Brands are seen as irrelevant, and so is most of the content they make. There’s a real need to adopt a more open attitude towards AI and machine learning and for marketers to use them to do their jobs better, quicker, and with more confidence.
We’ve already seen that the most sophisticated marketers want to use these technologies to improve the quality of their communications. L’Oreal, Unilever, AEG, Universal Music, the BBC, Adam & Eve, and more, are hungry for de-biased audience insights as they move towards an audience-centric model. They’ve got there by shedding themselves of these two major mistakes that marketers make. And they’re much the better for it.