How artificial intelligence can be applied in marketing.
According to Klaus Schwab, Founder and Executive Chairman, World Economic Forum, “AI is disrupting almost every industry in every country.”
Broadly speaking, AI is the emulation of human intelligence by machines. It can be defined as the ability of machines to learn from the data and situations provided to them to enable them to make intelligent, informed and improved decisions. The difference between the old (automated assembly lines) and new robots is that the old ones could only perform those functions for which they were programmed, based on a fixed set of rules; new robots can make intelligent decisions for which they were not programmed, based on their learning from the surrounding environment and the data provided to them. From a marketer’s point of view, the benefits of AI can be divided into two segments:
1 Acquiring vast amounts of data
This has to do with the ability of machines, cameras, sensors or hardware to acquire more information from their surrounding environment in real time. Computer vision (or more specifically facial recognition, image processing and object recognition) is a function of AI. Similarly, speech recognition is also a part of AI. Amazon’s Alexa, Google’s Virtual Assistant and Apple’s Siri are based on AI.
2 Prescriptive and predictive analytics
The power of prediction is based on the large amounts of data that can be gathered though AI. This is the most powerful function of AI, which has the ability to transform entire industries through accurate predictions. In fact, prescriptive and predictive analytics will be the most important tools for brands.
Trax Retail, a Singaporean start-up founded by Israeli scientists, uses computer vision to track FMCG products on the shelves. This not only ensures their visibility, it also gives insights about how shelf placement and product placement relative to other brands impact sales.
How marketers can leverage AI
Vast amounts of data can be acquired and used for data-driven marketing in ways not possible before. In the indoor DOOH environment, the engagement rate of ad viewers can be measured through computer vision to create programmatic ads – an environment where machines decide which ad to serve to which audience based on their engagement with the ad. Engagement levels are measured through the time spent looking at the ad and the emotions evoked. Emotion recognition is also used in market research techniques to understand consumer behaviour in retail environments. In 2015, M&C Saatchi, in partnership with Postercope and Clear Channel UK, created the world’s first AI-based DOOH poster campaign for a coffee brand which could evolve to write its own copy and select its own images and fonts based on the emotional reactions of passersby. Similarly, object detection techniques can be used to detect the number of cars on a road and their models to make OOH spending more efficient and accountable. For example, this data can be used to make DOOH ads programmatic-based on income levels (by knowing the type of car), number and speed of cars.
2. Trade marketing
In trade marketing, particularly among FMCGs, a major problem is to ensure the visibility of the merchandise. Trax Retail, a Singaporean start-up founded by Israeli scientists, uses computer vision to track FMCG products on the shelves. This not only ensures their visibility, it also gives insights about how shelf placement and product placement relative to other brands impact sales. Computer vision can be used to acquire customer insights in a store (how they interact with brands, customer demographics, footfall, etc.). Like digital media, it can be used to compute footfall in a store and the rental a brand is paying for its visibility and carry out cost conversion with respect to how many people interact with a marketing promotion tool (e.g. a POP Rack) and calculate ROI on the trade marketing investment.
AI can be used to serve different creative executions of the same campaign (based on different colours, fonts and copy) to different people based on their historical engagement with such types of executions. Such personalised campaigns can improve average CTR’s (click through rates) and viewable impressions, generate more conversation around the brand and bring down the overall cost of media buying, while creating more engagement.
3. Media buying
The most important use of AI in marketing is the power of predictive modelling to generate maximum engagement and conversations from programmatic media buying across all media platforms. The current digital programmatic media buying model is based on the bidding rates across different keywords, channels and audience segments. AI can be used to serve different creative executions of the same campaign (based on different colours, fonts and copy) to different people based on their historical engagement with such types of executions. Such personalised campaigns can improve average CTR’s (click through rates) and viewable impressions, generate more conversation around the brand and bring down the overall cost of media buying, while creating more engagement. Google Marketing Platform developed a machine learning-based tool to increase the number of viewable impressions bought on premium placements. Compared to other impressions that did not use the tool, viewable CPM (cost per thousand impressions) fell by 34% and impressions on premium inventory more than tripled. Thanks to connected TVs, podcasts, radio and DOOH, these machine learning-based ads can be served across the entire advertising ecosystem. In 2017, JWT Canada used AI for their bank and airline clients to change parts of their communication (location or flight price) while still using one idea across their video, display and audio ads. Similarly, by connecting offline and online touchpoints from the moment people are exposed to a brand communication to the moment they make a purchase, AI can predict how many impressions (online or offline) a person needs to be exposed to before they will make a purchase, thus optimising the ad frequency for different audience segments and saving precious dollars for the brands.
Will AI replace marketers?
AI will never replace marketers. However, it will be an indispensable tool in the hands of smart marketers who can sharpen their axe through AI by gaining deeper consumer insights, achieving personalisation and relevance at scale and by serving the right ad, to the right person, at the right time and with the right amount of exposure.
Tayyab.Tariq is CEO, Advertelligent. email@example.com