Everything New Is Old Again
For a technology that will undoubtedly shape our future, it’s curious that the use of AI in marketing seems so circumscribed by the past.
Others have already observed that the boom in marketing AI is driven by technology catching up with theory; that we now have the processing power, the storage, and the data sets needed to allow algorithms developed some years ago to finally deliver.
But if AI is to fulfil its promise of allowing marketers to conduct one-to-one conversations with customers in real time, we’re going to have to address some very familiar problems.
The first of these is the assumption that what marketers want to say is the same as what consumers want to hear. I became aware of this problem when I first joined the industry in 2000, at a time when advertisers were using aggressive online advertising formats such as pop-ups, pop-unders, road-blocks to counter falling click-through rates, and internet users were responding by installing pop-up blockers in droves. Sound familiar?
Around that time I interviewed GM O’Connell, the founder of Modem Media, one of the first digital agencies in the US. His analysis of the situation was pithy and exact: “You can’t annoy someone into liking you”.
He also had a solution, which he called “advertising as a service”. This would be advertising so precisely targeted and so timely that it wouldn’t feel like advertising at all. Instead it would be so helpful that consumers would be happy to allow it through their increasingly sophisticated electronic defences.
Now, AI proponents will argue that this is exactly what the technology delivers. But there are a number of problems. The first is that – as an industry – marketing has pursued the approach of trying to annoy people into liking us so assiduously that virtually a whole generation is now lost to us behind ad-blocking software. The task of persuading them back by arguing that this time it’s going to be different will be all but impossible.
A second, linked, problem is that we are only at the beginning of AI being able to deliver “the right message to the right person at the right time”. And while I firmly believe we’ll get there, we need to ask ourselves how many more ad blockers will be installed before we do.
To be fair, this isn’t simply a problem of over-eager marketing. Consumers want more targeted and relevant marketing messages, but it’s very difficult to see how personalised your online experience actually is. This is the problem addressed in Eli Pariser’s 2011 book The Filter Bubble. You might rail against the retargeted ads that follow you around the internet, but at least you were interested in those products at some point. What’s all but impossible to recognise is the amount of irrelevant advertising that you’re not seeing.
There’s also the question of privacy. For around ten years, surveys have shown that around two-thirds of people want more relevant, more targeted advertising; roughly an equal number don’t trust the technology by which such personalisation will be delivered.
These are the reasons why some people – including me – see an alternative approach to AI use as being so compelling. This approach, based on the ideas of the personal information economy, would see individuals holding their own data and releasing it to the market when they wanted to make a purchase in a reverse auction. For years the stumbling block has been the technology required to interface between an individual’s data store and the companies that might want to fill their expressed need for jeans, say, or a TV. But now that gap is being filled by the rise of AI-driven intelligent agents such as Alexa, Siri or Cortana. Rather than putting personalisation into the hands of marketers, the AI revolution might just turn marketing on its head, giving total control to purchasers.
If this doesn’t happen, AI will still transform marketing beyond recognition – we just have to make sure our customers are still paying attention when it does.
Michael Nutley, Journalist, editor and content strategist