Farewell 2017, a year in which marketing technology really took off and where both consumers and marketers alike started to experience the potential of artificial intelligence in our daily personal and professional lives. Thanks to our new voice-operated companions Alexa and Siri, and advanced analytics tools based on machine learning becoming increasingly accessible, we all caught a glimpse of the exciting future driven by AI.
Or did we?
Certainly, interest in AI among marketers this past year was the highest it has ever been. It was difficult not to stumble upon some discussion about AI and marketing in every conference, blog, trade press article or pitch. That's a very healthy interest, and we should encourage it into the new year. However, AI hasn't yet lived up to all its promises (and hype). I don't know about you, but there's a 50-50 chance when I ask Siri something that I will get an intelligent response. Half the time it fails to understand what I've just said (which is deeply troubling to me given that there are no voice recognition issues when my three-year-old son says "Hey Siri, what's the weather like?"). But voice recognition matters aside (which aren't truly what AI is about anyway), we are still in the nascent stages. We're not quite there yet, which is perfectly fine. This year we will see more advances and the computer science and software engineering will keep on getting better and better. That's exciting, but from a professional marketing standpoint, there are a few things that we need to keep in mind as (hopefully sophisticated) users of AI, machine learning and advanced analytics tools.
1. Data quality is paramount
We all know "garbage in, garbage out" (GIGO). This needs to be a marketing mantra in the age of AI-enabled marketing. The learning that happens in AI, particularly in advanced systems based around deep learning and neural networks, requires a lot of data. So collecting a lot of data is important, that is, quantity matters. But the quality of the data needs to be taken into account (just like in the good old days of traditional "quant" market research). In fact, data quality is paramount. There's nothing magical about the algorithms behind AI systems, it is just data crunching in sophisticated ways. And if the data are highly noisy or poorly measured, it is harder for the systems to identify meaningful patterns that allow it to make good predictions upon which your decisions can be informed. So focus on data quality more than ever. Collect as many data points as you desire (after all, storage is cheap these days), but make sure that every variable you capture is properly measured, understood by all the relevant people in your organization, thoroughly documented and actually used in business decisions. If you get the data capture, particularly measurement, right, then you stand a greater chance of getting more value out of your analytics and AI systems.
2. Fluency in all things AI, machine learning and advanced analytics is essential
Fluency for all marketers in your organization is necessary. Fluency does not mean technical proficiency, however. It simply means that everyone needs to have a reasonable understanding of what they are talking about when it comes to these methods and tools and how they can be applied to marketing practice. This will likely require some training and "upskilling" but it is worth it. As these tools become more common it means that understanding what they can-and cannot-do is required of everyone on a marketing team. It also makes marketers less susceptible to relying on technical experts to explain everything, which can be wasteful of time as well as lead to confusion and misunderstandings that get in the way of great work. Ultimately, leaders need to push fluency in this respect, and need to realize that although not everyone needs to know how to use a hammer, they do all need to know what the hammer is, what it does, how it can be useful and when it shouldn't be used.
3. Automation is great, but not everything needs to be automated
In 2017, a lot of the discussion about AI in the popular and business press revolved around the use of AI to automate common and repetitive business tasks-as well as in the near future complex tasks like driving cars, buses, trains and trucks. Robots, in other words. This is exciting and definitely a lot of digital marketing activity is now automated, based on decision rules (algorithms) that marketers set up and run for things like programmatic ad buying. It is also pretty common nowadays to read reports about some new startup that claims their AI system can write better emails, Facebook posts or creative copy than expert human copywriters. Great, we've automated a bunch of marketing tasks and have used smart software to do it. There's nothing wrong with that, but as we embark on a new year it is worth remembering that we should resist the temptation to automate everything. Automation works best when the software has an advantage over humans. For example, when it can buy ads in milliseconds to optimize an ad buy, or do personalization at scale for emails in a direct marketing campaign. But just because we can automate creative processes that go into making great ad campaigns, for instance, doesn't mean that we should. It is better to use automation when needed and free up human time so that extra effort can go into marketing tasks, strategic thinking and creative processes where human effort is most valuable. Use AI to automate when it makes sense to do so, and don't be driven by the urge to automate just to lower costs.
4. Consumers are absolutely critical for AI success in marketing, so don't forget them
Ultimately, consumers (or customers or clients) are what will determine the success or failure or your efforts to use AI and advanced analytics in your marketing operations. Consumers, however, mostly do not understand AI and how it can be used by their favorite retailers and brands to make their shopping experiences better. When the lay-consumer thinks about AI they think about driverless cars, robots and tech giants like Apple, Facebook and Google. They don't understand predictive analytics, they don't know how algorithms affect the news they see and the ads that are targeted at them and they aren't up to speed on the latest methods and trends being used in the industry in efforts to make marketing and advertising both better and more efficient. Some consumers see magic and appreciate it without thinking too hard about it. But others see opportunities for misuse or worse (which has not been helped in 2017 with news of how ad platforms like Facebook were manipulated by nefarious entities to influence elections). Consumer distrust of "smart" AI-based systems, therefore, is a big issue and it can undermine your efforts to build an advanced, robust, efficient and effective analytics system. If consumers don't trust the organizations and systems that use their data, then marketers won't have quality data. Over the coming year, therefore, I think we as a marketing and advertising industry must do a better job of helping to inform and educate the public on these technologies. And, perhaps even more importantly, we need to ensure that we do everything we can to make our data-driven practices as transparent, accessible, open, honest and value-creating as possible. I don't think we face a crisis of consumer trust in data-driven, analytics-based marketing yet, but we do need to develop trust so that consumers are more savvy with respect to these matters.