Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Artificial intelligence is set to change the way we work, but its full potential remains untapped. In marketing analytics, AI promises to revolutionize the industry:
Given the potential for transformational gains, the widespread adoption of AI should be the norm in marketing analytics. Why not? What are the barriers to this shift? More importantly, what can organizations and their teams do to change this? Here we provide practical answers to these questions.
Let’s start with the blockers, as highlighted in IBM’s 2023 AI Adoption Index. They identify five key barriers:
These challenges are significant, but we see them as obstacles rather than insurmountable obstacles—obstacles that can be overcome with a case-based approach to AI deployment.
Over the past year, we’ve used this approach on nearly a dozen brands, achieving rapid value and significant performance improvements. Here is the tutorial.
Dig deeper: The path to smarter marketing based on artificial intelligence
Sometimes the use cases are self-explanatory. For example, a large retailer we work with is facing a churn problem where an AI-based approach to churn prediction could deliver significant business value.
Other times, the most relevant use case is not so obvious. In these cases, creating a catalog of use cases helps prioritize opportunities. This catalog lists potential AI-enhanced use cases and ranks them based on impact, scale, and effort required.
Here are some basic AI use cases in marketing analytics that we’ve come across:
These examples illustrate how AI can deliver significant business value. Once use cases are defined, the focus should shift to overcoming implementation hurdles.
Dig deeper: Artificial Intelligence and Machine Learning in Marketing Analytics: A Revenue-Based Approach
The first hurdle can be overcome by focusing on a high-value use case with little effort, as highlighted in the case catalog approach. For example, our churn prevention strategy for one client involved using AI-driven customer intelligence to trigger email messages for high-risk customers. The solution was seamlessly integrated into existing workflows and demonstrated how targeted use cases simplify scaling efforts.
The complexity of the underlying data is the most common obstacle we encounter. The aphorism, “Don’t let the good become the enemy of the great,” is apt. Data is never perfect. The best approach is to put aside the pursuit of perfection and focus on the data that matters.
Web interaction data and customer transaction data are two types of data commonly available in most businesses. They are particularly powerful for building AI-driven segmentation models for propensity, engagement, loyalty and churn. Additionally, AI-enabled data preparation and cleansing can automate tedious tasks and enable faster and more comprehensive data access.
Dig deeper: 4 ways to fix bad data and improve your AI
Spending problems often stem from a fundamental misunderstanding of value creation. Implementing AI in marketing analytics requires investment. This can range from a modest $50,000 to start up to seven figures for more ambitious projects. However, this expenditure is an investment, not just an expenditure.
ROI can be predicted, quantified and measured. Focusing on specific use cases makes it easier to build a strong ROI business case to justify the investment. For example, AI-driven segmentation and scoring typically delivers a 10% to 15% improvement. A brand investing $20 million in outbound marketing could see an annual return of $2 to $3 million, making a compelling case for investing in AI.
Expanding the pool of available expertise can address limited skills. While few professionals have both the technical skills and the subject matter knowledge to deploy AI for marketing analytics, the issue is primarily an internal enterprise issue. The solution is outsourcing expertise.
In a rapidly changing environment where specialized skills are scarce and essential, it is often impractical for businesses to develop these capabilities in-house. Partnering with a specialist to build custom AI marketing analytics applications is the most effective, low-risk approach. These efforts can eventually become owned assets, but without the immediate burden of building and implementing them internally.
The last block, ethical concerns, stands aside from the previous four. While ethical considerations in AI are serious and impactful, we did not see them as a significant barrier to the adoption of AI in marketing analytics. A more common blocker is a practical one: legal and compliance issues.
Legal and compliance teams are particularly concerned with generative AI, where concerns about inappropriate or off-brand content, as well as copyright and intellectual property risks, can significantly slow down or even halt AI initiatives.
Ultimately, each organization must establish its own governance and controls for AI adoption. Focusing on high-impact, low-risk use cases has worked well for starters. For example:
Artificial intelligence is transformative and will revolutionize marketing analytics. A use case-driven approach provides a clear roadmap to overcome barriers to implementing AI in marketing analytics. This measured strategy paves the way for sustainable AI integration, increases internal team trust, and fosters AI expertise within the organization.
Marketing analytics leaders who adopt these strategies will be well-positioned to increase performance, streamline operations, and cultivate a responsive, data-driven culture that is ready to harness the potential of AI.
Dig deeper: Why causal AI is the answer to smarter marketing
Contributing authors are invited to create content for MarTech and are selected for their expertise and contribution to the martech community. Our contributors work under supervision editorial office and submissions are reviewed for quality and relevance to our readers. The opinions they express are their own.