Duke University, Fuqua School of Business MBA students Katie Hall and Torren McCarthy as well as Holly Larson contributed to this post.
The age of artificial intelligence (AI) in marketing is upon us. The August 2019 CMO Survey confirms this, with corporate marketers reporting a 27% increase in the implementation of AI and machine learning (ML) in marketing toolkits compared to just six months prior. This large uptick is just the beginning of marketers’ commitment to improving customer insights, personalization, targeting, and predictive capabilities, among other applications. Marketers project a 59% increase in the use of these tools in the next three years.
A deeper look at survey data, however, reveals substantial heterogeneity in firm adoption of AI. Perhaps even more important, we see a range of technical and organizational challenges that are limiting the full adoption and organizational impact of AI-driven tools.
Which Companies are Adopting AI Tools for Marketing?
Despite large aggregate increases in current and projected implementation, not all firms are adopting AI or using AI across the same types of marketing activities. Chief marketing officers (CMOs) at higher revenue companies report stronger current and projected adoption rates, as they use their greater resources to leverage data to increase customer engagement. This effect is most pronounced with companies with annual revenue of $1B or more. The same trend holds true with companies that have a higher percentage of internet sales. These firms are harnessing the skills of data scientists and abundant data about their online prospects and customers to optimize the customer experience, with the goal of driving revenues and profitability.
Across industry sectors, we see that corporate adoption rates are highest in transportation, technology, and education and lowest in manufacturing, mining, and energy. In addition to these verticals, adoption rates are markedly higher at B2C companies versus B2B companies, with B2C services companies leading B2C product companies in AI application.
Popular AI Applications in Marketing
While the presence of AI grows overall within the marketing function, certain applications have been more widely adopted than others. Big picture, we see that the top three AI applications all involve a focus on driving customer value with more effective marketing. Specifically, over half of respondents are utilizing AI technologies for content personalization (56.5%) and generating customer insights using predictive analytics (56.5%). Nearly half are utilizing AI to target customer decision making (49.6%). These applications vary quite a bit by sector with B2C companies dominating in their use of predictive analytics for customer insights and targeting decisions—especially B2C services companies. Content personalization, on the other hand, is the focus of the other sectors. Overall, B2C companies are investing in applications such as optimizing programmatic advertising and media buying, improving marketing ROI by finetuning marketing content and timing, and implementing conversational AI for customer service.
The Top Marketing Use Cases for AI Right Now
As marketers look to use AI in content personalization, predictive analytics for customer insights, and targeting decisions, here are some key considerations to guide their journey:
- Content Personalization: AI can be an effective tool to gather deeper insights and understanding of customers. By gathering the right data, marketers can develop more personalized offerings of products and services across customer channels. However, mistargeted personalization can result in customers feeling frustrated at best and mistrusting brands at worse. Research by Accenture finds that poor personalization and lack of trust cost companies $756B in 2017 as 41% of fed-up customers switched companies.
- Predictive Analytics for Customer Insights: Predictive analytics provides the opportunity for companies to deliver relevant and timely offers to consumers. This makes AI an important tool for sales and marketing organizations to boost customer acquisition, loyalty, and retention. However, machines using predictive analytics can only predict outcomes they have been programmed to recognize – meaning they can miss critical information or jump to faulty conclusions—Target’s prediction and targeting of a teen’s pregnancy is a case in point. Companies should take time to fine-tune AI models before putting them into wide practice.
- AI for Targeting: Using AI to develop audience segments and reach users is a logical way to deploy the technology. A recent survey by the Data & Marketing Association found that 62% of marketers said that improving audience segmentation to support better ad targeting is one of their top campaign management priorities. However, nearly half of consumers are concerned about how companies are using their personal data. New laws are being implemented to restrict such usage, including what data organizations can collect and use, how data must be stored, and providing customers with opt-out and deletion rights. With regulations varying by country and even state, companies seeking to utilize AI for targeting must have a keen understanding of requirements and their obligations before proceeding. However, data from the August 2018 CMO Survey finds that marketers are not worried that their company’s use of online or third-party data could raise concerns about data. In that survey, only 10% reported that they were very worried about privacy.
Organizing to Maximize the Impact of AI in Marketing
Beyond the challenges of specific marketing applications of AI, firms also suffer from a set of organization-wide challenges that can impede their ability to use AI in marketing and more broadly. We discuss each need and point to solutions below.
- AI should be driven by strategy. Companies often lack a clear strategy for AI technologies. Adoption should not be driven by fear of missing out (FOMO), but by clear company objectives that specify what outcomes AI will help the company achieve and what data it will offer to support marketing decisions. In this instance, teams can take a page from corporate Internet of Things applications which tie to the broad corporate and technology strategies and create stair-step gains over time.
- AI should take an enterprise-wide view. Organizations should catalogue where AI opportunities exist across functions, any current or required data sources needed to capitalize on AI, and areas of overlap. This effort will lead to a shared view of these important investments and help the firm build AI capabilities—not just pockets of smart marketing inside the company that entrench organizational and data silos.
- AI should be built into organizational capabilities. AI ideally should be an end-to-end corporate competency that spans functions. As such, it requires a shift in mind-set around the execution of core capabilities. Experts suggest that navigating these initial barriers effectively requires a clear direction on how AI will not only benefit the organization, but also how it will impact and be integrated into the company’s roles, functions, and culture.
- Smart AI requires smart marketers. While AI can be utilized to personalize content across a broad set of customers based on certain attributes, it is important that this process is monitored by humans who apply creative expertise and ensure contextual alignment to achieve desired business outcomes. For example, when working with AI-enabled campaigns or marketing automation tools, there is a temptation to “set it and forget it.” This strategy only works when marketing is not expected to exhibit diminishing marginal returns or when competitors are not expected to respond—both unlikely. Human monitoring of these tools over time will bolster their effectiveness by identifying new opportunities and addressing issues as they arise. There is a looming concern that AI will replace the value that is now currently provided by human beings. CMO Survey results suggest that this concern is overblown with only 1.7% of marketing leaders reporting that new technologies are replacing marketing employees in their companies by “a great deal” and 57.6% reporting that this is occurring “not at all.”
- Strong AI demands strong data. It has been reported that 67% of respondents at firms with a strong digital base say their organizations have embedded AI into standard business processes, compared with 43% at all other companies. While this effort takes time, it sets companies up for long-term success—much like Amazon’s analytics capabilities and digital platform enables it to deliver continued propulsive growth.
- AI is only valuable if marketers use resulting marketing analytics. AI that produces marketing analytics is only worthwhile if it is used. While this is a self-evident truth, CMOs report using marketing analytics in decision making 39.3% of the time, representing a 29.3% increase since 2013. However, this number is still too low. Companies will have to work harder to build AI-driven data and decisions into their standard operating procedures so that they systematically produce a real payoff.
- A gap in talent makes the use of AI challenging. The CMO Survey also found that only 1.9% of marketing leaders reported that their companies have the right talent to leverage marketing analytics—much of which power AI applications. To overcome the talent gap, organizations should do whatever it takes to ensure they are properly staffed, whether that is hiring external talent, building capabilities in-house, or buying or licensing capabilities from large technology firms. Perhaps even more important, companies should work to align their AI strategy with their analyst talent strategy so that human and machine learning complement one another.
- Successful AI requires top management sponsorship. Given that marketing applications of AI provide clear cases that demonstrate the impact of these technologies, the CMO can lead change by collaborating with the CIO and company leadership to build out an analytics function that has the potential to transform functions across the organization. Business owners should be closely involved throughout the development process so they can better understand what is being delivered.
- Take AI baby steps. To prove the business value before making sweeping changes, incremental steps should be taken. For example, demonstrating the value of AI by generating insights from existing data can highlight the business importance of integrating existing platforms and capturing more comprehensive data, thus securing business buy-in for expanding AI’s use at the firm.
- AI should address “what’s in it for me.” Many corporate workforces are experiencing change fatigue with the relentless adoption and refinement of new business models, advanced technology, and customer-centric engagement strategies. Corporate marketing leaders should address this challenge by demonstrating how learning new AI capabilities can improve marketers’ skill sets and knowledge, professional opportunities, and paychecks. New AI competencies should also be tied to professional development plans, so that employees are held accountable for gaining new skills.
As the August 2019 CMO Survey demonstrates, AI has passed the tipping point of adoption, with nearly all firms reported some level of usage and many targeting multiple applications for exploration and adoption. It is not surprising that the most popular applications are customer facing because AI provides tools to personalize the customer experience—increasing customer value and sales over time.
Putting AI into practice is hard work. Companies will need to develop new organizational competencies and cross-functional operational processes to capitalize on its full potential. For those that do, a bright new future of marketing awaits, with predictive capabilities and precision targeting that have the potential to deliver unprecedented results. CMOs and their marketing teams should, therefore, stay the course and commit to continuous improvement, knowing that today’s incremental gains with AI will lead to significant and sustained advantage in the months and years ahead.