2022: Recent Advances in Personalized AI-Driven Marketing
Have you bought something online? In today’s world, the answer is almost certainly yes. As a result, you have provided your payment information to a vendor, and indicated your interest in a certain type of good or service. However, your data then will be indexed across various different corporate databases. Companies share their data with other companies and vice versa in order to develop a holistic profile of a certain consumer. Additionally, public records and other online records are added to the sum of a given consumer profile. Usually this creation of a consumer profile is done in the pursuit of creating a three-dimensional view of a certain buyer in order to better develop a picture of his or her spending habits and consumer preferences.
Although this practice is a product of the 21st century and therefore not a truly new practice, certain new developments have taken place in recent years, such as the entry of machine learning and data science into the area of consumer data. The use of such sophisticated techniques allows for a deep analysis of consumer profiles.
One company, PreciseTarget, has specialized in the analysis of this mass accretion of data. PreciseTarget aims to renovate digital marketing by leveraging data analytics to provide insights about certain consumer segments. AcquisitionAI, a service of PreciseTarget, seeks to analyze the data set of customers by potential value to determine which customers provide the highest value to the company. ConsumerInsights, another service of PreciseTarget, seeks to evaluate the best acquisition targets for a given company based on the company’s sales record.
The type of deeper personalization pursued by PreciseTarget allows companies to retune their marketing strategies in order to more accurately and finely define a certain consumer. Before the advent of such AI-guided marketing techniques, companies usually marketed goods and services to their previous customers based on those customers’ past purchasing habits. Now, with the aid of machine learning, companies are able to make predictions about consumers’ spending habits that extend beyond the relatively narrow scope of past purchases. Much of this analytical power involves typification, where consumers are assigned to a certain category and and a number of subcategories. Marketing suggestions are then made about that particular consumer based on the categorization of their profile.
Despite its success, machine learning is not a magic bullet. The distinction between general AI and specialized AI should be made clear. While a general AI would be capable of performing a variety of tasks similar to a way that a human could, a specialized AI can only perform a narrow range of tasks for which it has been specifically designed. Thus far, a general AI has proved impossible to develop, and so only specialized AIs have been developed. Even with these specialized AIs, the term ‘AI’ is somewhat of a misnomer as these technologies are not intelligent, lacking the ability to distinguish the map from the territory. Such programs are also susceptible to bias and flaws, and the deep analysis of consumer activity requires human feedback from those with marketing expertise. This supervision of AI is now a generally accepted standard, since human intelligence has been proven necessary to qualify the output of machine learning and data science activities.
What might the future of AI and marketing hold? A personalized consumer experience may be on the horizon. Venkatesan, professor at the Darden School, notes how Coca Cola created a personalized experience at advanced drink dispensers by allowing consumers to consume specialized drinks based on their personal preferences. The dispenser selected drinks for athletes based on a formula analyzing their levels of hydration, among other factors. According to Venkatesan, the diversification of the consumer experience will create a better experience by varying the features of the given product or service in response to the needs of a given customer. Although the advanced drink dispenser may seem like a basic example, there are other cases that possess greater sophistication.
Digital commerce has already bgun personalizing the consumer experience in a variety of ways, so further examining the application of personalization technology as it stands may be useful in order to predict the future course of such technology within the context of other applications.
The future of AI is bright within the specialized field of personalized marketing. The Customer Data Platform Resources determined in a survey of 2,500 adults that approximately 81% of consumers do not unconditionally oppose the idea of machine learning, data science, and other specialized AI tools being used to promote marketing for the purpose of the delivery of personalized recommendations and promotions for consumer products and services.
According to CDP, 58% of people prefer product recommendations based on prior search history, while 43% prefer replenishment recommendations based on past purchases.
What room is there for AI-assisted marketers to grow their customer base and improve their techniques? Recent strides in technology has made marketing extremely accurate; 32% of consumers rated personalized recommendations as extremely accurate. In the same study, 58% of people rated the personalized recommendations as somewhat accurate. Although this is a fairly good result, it indicates that recommendations are not perfect. However, 10% of consumers did rate personalized recommendations as inaccurate, indicating the presence of a disgruntled minority.
A study by Gartner indicates that such a rosy view may be slightly premature. A survey of marketing leaders in late 2020 indicated that only 17% of digital marketing departments use machine learning, data science, and other special-AI techniques broadly across their marketing system (in an integrated fashion). [Another study found a similar but slightly higher figure of 19%, and a study with stricter criteria found that only 6% of marketers were fully utilizing the state-of-the-art AI-related marketing tools at any given time]. Furthermore, 63% of marketing leaders indicated frustration or challenges with delivering personalized experiences to customers.
However, AI can be useful tool for digital marketing when properly implemented. One study found that 26 percent of respondents indicated that adopting specialized AI tools enabled their businesses to gain a competitive edge in the marketplace.
Despite the lack of companies using specialized AI in a broad fashion, a large number of companies have only very recently adopted the use of this technology. A study by Salesforce Research, entitled State of the Marketing, indicated that the share of marketers using AI increased from 29% in 2018 to 84% in 2020.
One way that companies and marketers might seek to make the most of AI could be by expanding the use of the tool outside of its traditional main focuses, entertainment and shopping, as well as outside its traditional realms of ecommerce and advertising. One example is that of a dieting application called DayTwo. This program analyzes the user’s blood to determine what the user ought to eat that day. One user of DayTwo found success using the app to address his longstanding affliction of diabetes. He managed to lose one hundred pounds in two months with the aid of the application, and he credited his success to its use. Another company, ZOE, offers a similar product. Although these companies have yet to prove the potential for long-term commercial success, their growing user base indicates that they have had at least some success.
This example of the use of AI to leverage personalized marketing techniques within the healthcare and wellness scene is not unique to the often imprecise realm of dieting and nutrition. At the Massachusetts Institute of Technology, researchers have developed machine learning technologies that enable medical providers to more efficiently provide breast cancer radiography services to the broader population. These algorithms were developed in response to existing problems within the current framework, such as a false positive rate of up to 19%, and the occasional failure to identify cancer at the optimal stage. The algorithms sort patients into categories based on risk factors, thus reducing the risk of failing to diagnose late-stage cancer before it becomes advanced, and additionally reducing the chance of false positive by indexing cases against the typical patient to develop a predictive profile. Although the algorithm does not provide 100% accuracy, it represents an improvement over previous techniques. Such developments as these may provide greater confidence to patients seeking such medical services, even though healthcare may be traditionally be seen as a field where data science-driven personalization would be immediately relevant.
Ballpark numbers of special AI improvement related benefits suggest that adopting a personalized marketing strategy can reduce customer acquisition costs by up to 50%, increase revenues by up to 15%, and increase the efficiency of expenditures by up to 30%, in an ideal situation.
Some statistical studies have also indicated some concrete benefits of specialized AI within personalized marketing. One study by Forrester Consulting found that personalization-at-scale results in an average of a 5.63% increase in sales revenue, a 10.26% increase in order frequency, a 2.57% increase in average order value, an 11.22% decrease in the cost of marketing, a 10.82% increase in click-through rates, a 2.69% increase in conversion rates, and a 13.25% increase in cross-selling and up-selling opportunities.
Clearly, although some progress has been made, many improvements remain to be undertaken. Any declaration of a revolution in AI-drive personalization is premature. The general hype surrounding the concept ensures that expectations continue to soar above the level of what has actually been achieved to date. Nonetheless, a clear examination of the technologies that have actually been implemented indicates that improvements have been made, although they have not been revolutionary. Additional confounding factors exist in the measurement of the efficacy of a certain marketing strategy or technology. At any rate, the type of hyper-personalized products envisioned by the drink machines of Coca Cola remain anomalous today; the concept of a truly personalized experience continues to elude almost all industries and product lines, despite the desire for such having existed since at least 1980. After all, isn’t the notion of the factory, the assembly line, the interchangeable part, the algorithm, the machine, and the sorting box antithetical to the very idea of personalization? A deterministic system implies limited and clearly defined causes, which also imply limited and predictable outcomes. Yet out current tools are unable to provide a unique experience for each person, as the cost would be too high to the producer. Only a truly visionary marketer would be able to rise above this contradiction and heed the call to this challenge.
Sources:
Report: Consumers Open to AI in Marketing, But Privacy Concerns Remain
https://www.forbes.com/sites/louiscolumbus/2021/02/21/10-ways-ai-and-machine-learning-are-improving-marketing-in-2021/?sh=226847d714c8
https://www.gartner.com/en/newsroom/press-releases/-gartner-says-63–of-digital-marketing-leaders-still-struggle-wi
AI Marketing Statistics: Data Shows Emerging Technologies Are Here Now
https://www.drift.com/books-reports/state-of-marketing-ai/#Marketing+AI+Use+Cases%3A+Key+Findings
https://business.adobe.com/resources/retail-personalisation-with-ai.html
https://www.nytimes.com/2022/03/14/well/eat/ai-diet-personalized.html
https://www.ibm.com/watson-advertising/thought-leadership/how-ai-is-changing-advertising
https://tenovos.com/resources/blog/ai-content-personalization/
https://news.mit.edu/2022/seeing-future-personalized-cancer-screening-artificial-intelligence-0121
https://www.sidley.com/en/insights/publications/2022/07/artificial-intelligence-considerations-for-femtech-and-beyond
https://www.publicissapient.com/insights/make-it-personal
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwj3wJiLq_b4AhU4DkQIHUfjArI4MhAWegQIHBAB&url=https%3A%2F%2Fwww.marketingaiinstitute.com%2Fblog%2Fmckinsey-ai-will-have-the-most-impact-in-marketing&usg=AOvVaw3AgxaknzzVkG1cBXJUCOuU
https://www.wbrc.com/2021/10/06/companies-are-tracking-your-personal-data-without-your-consent-what-you-need-know/
https://www.businesswire.com/news/home/20211110005933/en/PreciseTarget-Launches-the-First-AI-Based-Customer-Acquisition-Product-for-Retail
https://hbr.org/2021/07/how-to-design-an-ai-marketing-strategy
https://ideas.darden.virginia.edu/AI-powered-marketing
https://99designs.com/blog/web-digital/ecommerce-personalization/
https://www.businessperspectives.org/index.php/journals/innovative-marketing/issue-348/how-artificial-intelligence-can-change-the-core-of-marketing-theory