Customer Journeys: ML to the Rescue

How AI Promises to Orchestrate Personalized Messaging to Customers Everywhere

The most recent No Brainer Podcast centered on the move away from third-party data cookies and how AI may resolve it.

Data protection laws like General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA) have forced browser companies to shed third-party cookies, forcing digital marketers into a new territory. As the year progresses, marketers dependent on electronic commerce must collect first-party customer data and better interact with them to convert.

In an ironic twist, while generative AI continues to dominate the headlines (hello, Sora), machine learning (ML) and predictive analytics are the engines driving the promised new marketing era. Generative AI works at the end of some models to generate personalized content based on data analysis and recommendations.

ML is the primary algorithm behind insights and predictive outcomes that inform human strategy. Humans and larger complex AI models use insights to create personalization frameworks and incentive or campaign recommendations. Then they use generative AI to create personal communications within the guidelines. Call it the revenge of ML.

Traditionally called customer data platforms or CDPs, this product segment is evolving quickly to meet the AI-fueled first-party data world. Sure enough, CDP vendors, CRM providers, data analytics service providers and new AI vendors are all promising solutions for a personalized future, albeit using machine learning algorithms to parse through wide swaths of data. 

Some may argue that the tools used are more automated than actual AI; however, technology vendors are marketing their data mining tools as AI. Marketing organizations care less about technical arguments and semantics; instead, preferring to focus on whether or not hyped outcomes are attainable for their marketing department.

For marketing leaders, it is a race against time. By the end of the year, the days of tracking customers across the websites, pulling all their browsing data, and sending them offer after offer wherever they go will end. Instead, their first-party data becomes paramount, and the pressure to execute against each lead becomes more critical. To succeed, marketers must attain a new level of precision and personalization to convert.

The remainder of this article will examine the promise of AI – specifically ML – to help marketers better extrapolate first-person data. Next up, we will examine how AI is creating personalized digital experiences for a better post-party cookie era. Finally, in the subscribers-only section, we will list several CDP and AI vendors working on actual solutions and how accessible they are to organizations. 

Unifying First-Party Data

With third-party cookie data becoming much less relevant in the immediate future, marketers must focus on unifying their first-party data to build better, more effective personalized customer experiences. That means going beyond the typical web-based CRM data to integrate and unify mobile apps, email, texting, call center, analytics, and brick-and-mortar data.

Of course, unifying data is simple to say and hard to perform for many organizations. It requires a data governance strategy to establish baseline standards for data quality. Data governance policies can be extensive, and while informing first-party data collection, they also include second and third-party data collection and enterprise and employee use guidelines. 

In most cases, unifying data requires extensive data cleaning. Often, brands cannot meet that ideal and settle for a policy and strengthen their data hygiene as time progresses. Data cleaning is one of the most time-consuming rote tasks in a data science team’s list of activities. Unfortunately, if executing a more personalized form of marketing at scale is the way of the future, then dirty data is the fastest way to reduce or, worse, tank a company’s conversion ratios. 

Machine learning comes to the rescue, of course. More and more companies use AI to assist with more rote data cleansing tasks. AI-assisted data cleaning includes deduping information across data sets, resolving conflicting information and errors, and using probabilistic programming to resolve uncertainties with machine learning algorithms. Generative AI can be used as a data-cleaning app to help fill in missing information based on prior examples, creating synthetic data. Machine Learning can train the generative AI to improve its capability.

Data remediation and cleaning are not the most straightforward tasks for enterprise marketers to allocate resources. Often customized to the enterprise, data scientists need to be involved, and the human resources arms race for capable AI services is not simple. Third-party professional services and data cleaning houses are abound. 

Given the importance of clean data to businesses’ future, it figures to be a booming enterprise. According to Grand View Research, the data cleansing market is expected to reach $26.7 billion by 2028. Recruitment, procurement, and acquisition become central to success here.

Currently, in larger enterprises, data science teams govern AI cleansing processes and oversee the resolution of conflicting data sets and challenges. The data cleansing process may change shortly if data cleaning algorithms develop and strengthen. One of the most significant barriers to achieving faster data readiness is unstructured data – free-form text, imagery, audio, and video. 

Small and medium enterprises may need more resources to clean and unify their data to the best standards effectively. That said, they should still pursue data governance, cleaning existing data gathering policy, and unifying data wherever possible. Long-term, vendor data-cleaning AI solutions will likely evolve and offer alternative solutions for down and up-market buyers.

The good news here is unstructured third-party data has become more valuable thanks to generative AI, with the help of human guidance to resolve open data questions. The ability of LLMs, diffusion models, and other generative AI algorithms to use unstructured data is fantastic. At the same time, using unstructured data incurs risks, including bias, data privacy, and ethical uses of the information. Generative AI can work as part of the cleaning solution, but it is an evolving role, with capabilities and best practices developing each month. While important, generative AI outcomes need to be verified by humans, as discussed across many industry conversations.

From First-Party Data to Personalized Marketing

Marketers can use artificial intelligence to glean powerful insights from first-party data to fuel more engaging personalized communications and strengthen campaign tactics. The total view of the customer journey triggers many possible improvements for marketing campaigns, including the oft-promised and rarely delivered goal of one-to-one personalized marketing. 

The ability to deliver greater personalization, even one-to-one marketing, is the holy grail of marketing. Since the days of Dale Carnegie, sales and marketing executives have focused on customer-centric selling, making buyers feel special. Today, it’s still essential. Salesforce Research surveyed over 6,700 consumers and business buyers globally, and 84% of customers say that being treated like a person, not a number, is very important to winning business.

Unfortunately, while effective for high-end B2B and consumer sales, personalized selling is not scaleable for most organizations yet. Instead, brands rely on marketing to convey that sense of personalization. Mass advertising pre-Internet focused on lightning bolt moments provided by electric campaigns that struck chords with vast swaths of customers. While brand promise is still paramount for any sound marketing strategy, we have finally reached an era where we can achieve a semblance of personalized marketing communications if systems and operations are built well.

Specifically, working with a unified customer journey founded on first-party data, AI allows for:

1. Greater sophistication in segmentation: Smaller niche audience marketing is suddenly realizable with insights about that segment and generative AI tools to create customized content.

2. Deeper insights about customer behavior and needs: By revealing greater insights into customers, marketers can deploy messages and tactics that may be more effective than thought or missing altogether from the customer journey. Conversely, marketers can deprioritize tactics that are over-emphasized.

3. More personalized recommendations and targeted offers: Based on a customer’s behavior, brands can use probability scoring to better communicate with them with unique custom-built offers and content. Again, generative AI tools built on personalization rules – pre-established branding and messaging frameworks that effectively prompt the AI – provide the necessary scaling to meet this need.

4. Improved customer experiences before and after the sale: Brands can use their deeper insights to understand conversion points better and foster stronger relationships with those customers to encourage repeat business and prevent churn.

This list provides some of the most common next steps or results from the unified customer journey. However, this list is incomplete and will always be so. Imagination is the limit for use cases, and marketers who want different outcomes should explore them. Mapping the vision to pragmatic data sets and AI models and training said models is the key to actual implementations.

The four examples are the best marketing practices most sophisticated digital marketing organizations have deployed. For example, anyone who uses AmazonPrime is familiar with the benefits or the results of this kind of precision communication from a retail organization. 

Here are a few more examples based on the Association of National Advertisers AI Use Case Compendium, authored by CognitivePath Principal Analyst Greg Verdino:

Deeper Customer Insights

Mastercard’s “Digital Engine,” a custom-built AI-powered social listening platform, allows the brand to spot emergent micro trends and respond quickly with conversationally and culturally relevant activations. In one example, Mastercard tapped into the buzz around a given celebrity to spin up a two-day campaign featuring behind-the-scenes videos of that celebrity. This micro campaign delivered 100% higher engagement, 254% higher CTR, and 85% reduction in CPC.

Personalized Recommendations and Targeted Offers 

Carvana celebrated its 10th anniversary with a customer loyalty campaign called “Joyride.” The campaign used data about individual customers and their past Carvana vehicle purchases to produce 1.3 million unique AI-generated videos. A variety of Generative AI systems were used to create and animate customized visuals, convert scripted text to speech, and re-voice each video to personalize outreach.

Improved Customer Experiences

7-Eleven data suggests that a business unit using the new AI system has reduced the number of internal product development meetings by 80%, and the company expects to reduce the time required for product planning by up to 90% and better align product distribution with emerging trends and customer needs.

While impressive, potential AI-fueled outcomes are a panacea for deep-seated marketing missteps, such as misaligned offerings, subpar strategy, and creativity. AI represents a series of power tools that allow those with strong marketing skills to execute exponentially faster and more effective marketing strategies and tactics. 

With AI, humans can provide incremental improvements at every step of the journey. The role of the human is to provide the actual strategy, creativity, direction, and quality assurance necessary to direct the tools. Human intelligence will differentiate the winners and losers in the personalization era.

Free readers may also want to check out this article:

Now, let’s take a look at some vendors.