Personalized: Customer Strategy in the Age of AI - By Mark Abraham
Date read: 2025-06-21How strongly I recommend it: 8/10
(See my list of 150+ books, for more.)
Go to the Amazon page for details and reviews.
Good overview of marketing personalization with viewpoints on AI. Given how fast AI is changing, I wouldn't be surprised if the AI section becomes out-of-date very quickly.
Contents:
My Notes
Personalization is creating experiences at scale that get fine-tuned with each successive interaction, empowering customers to get what they want—better, faster, cheaper, or more easily.
At its core, personalization is about speed. Speed in getting to know the customer throughout the customer journey, and speed in constantly improving the experience based on that knowledge.
Personalization, done right, goes beyond pushing customers to buy specific products. It’s about making customers’ lives easier.
The Five Promises of Personalization:
To empower your customers, you must first determine which parts of the customer journey are most critical for personalization.
Many companies proceed in the wrong order. All too often, they put the data and technology in place before articulating how they want to empower the customer.
Questions to ask when starting to map out your personalization strategy:
Arti Zeighami, the former chief data and analytics officer at H&M Group, recommends starting with “a couple of simple use cases, based on actual business problems, that would deliver the most value to the organization.”1 These should be projects that would be highly visible and could be scaled up quickly to win advocates for the new ways of working that personalization requires.
Types of Customer Data - Any information the customer willingly shares when explicitly asked is known as zero-party data. Though zero-party data includes all the information a customer provides when creating their account, it can be collected at any point. For instance, recording a customer’s call to customer service or virtual chat generates information about the customer that the company might use to personalize, including whether to send a follow-up communication, what tone to adopt in that message, and how to route a future call from that customer.
First-Party Data - Transactional and behavioral data created by recording actions the customer takes—browsing different pages on the site, hovering over specific items, making purchases, and so on.
Second-party data comes from a company’s trusted partners—other businesses that collect first-party data from customers who give their explicit permission to use that data more broadly. A classic example is a co-branded credit card. Merchants might send offers to their cardholders, generating data in the process, and then share that data with the credit card company, under the terms of the permission granted by the customer when they signed up for the card.
To further enhance data already in hand, companies can obtain third-party data through several different means. Most companies tap into what is available from paid media platforms (i.e., Meta & Google).
Many companies are putting in place chief data officers (CDOs) and establishing data stewards across the organization. CDOs define the data governance process. The data stewards—the actual data users, such as a marketing executive or the head of call center operations—help agree on common definitions for different types of data, such as a “lapsed customer.”
Three “intelligence” components of the personalization tech stack: targeting intelligence, experiment design and activation, and next best action orchestration.
Targeting Intelligence Types of Data:
Experiment Design & Activation - Decides which customers are eligible for which experiences based on the targeting intelligence, the available content, and the key rules
Next Best Action Orchestration - This layer prioritizes actions across channels; for example, it might skip sending a paid ad when a customer has already clicked on an email.
Regardless of your data and technology approach, prioritize scaling your customers’ personalization response data. This data provides critical insights into what works and what doesn’t with your customers, and it’s a key source of competitive advantage.
So, to craft a sound content strategy, start by asking: What do we want to personalize? For which customers? At what stages in their journey? How will doing so enhance the customer’s experience? and What business goals will it enable us to achieve?
In personalization, the possible combinations of attributes of content relevant for an individual customer interaction are so vast that companies are increasingly using AI to scan the content and create the tags automatically at the most granular level possible. A whole class of software solutions is emerging that enables and automates AI-driven tagging.
Companies need to establish the right governance processes to ensure that the many content components they develop are appropriate, free of bias, and meet compliance requirements.
Upskilling - Creative directors and their teams will need to excel at a range of skills vital for rapidly testing content and learning from the testing process. Such skills include knowing how to measure success using digital metrics and understanding how to apply the resulting data to optimize the impact of the company’s creative assets. Beyond traditional writing and design skills, teams will have to be cross-functional, and include data and analytics experts, data scientists, martech experts, and software engineers.
As the scale and breadth of content that companies generate dramatically expands, teams will also need people who know how to manage the prompts for gen AI by which the business generates variations, set guardrails for content design and development, and check content samples for appropriateness and bias.
Looking at the most successful AI transformations of the past decade, we have observed a common thread: what we call the 70/20/10 rule. Seventy percent of the effort in AI transformations involves people: processes, ways of working, incentives, and performance targets. Twenty percent entails getting the data right. The remaining 10 percent is about the technology foundation.
3 Core Measures:
Learning is a function of the number of experiments, multiplied by the time to measure, plus the time to act on the feedback.
We don’t recommend starting a personalization transformation with a full org redesign; that is best left for the second or third year of the effort.
At Spotify, the cross-functional agile teams are famously known as “squads,” with members still affiliated with their respective functional teams (e.g., marketing strategy, creative, operations, analytics).
In reality, it takes a village to lead and manage personalization. Almost every C-suite member will be responsible for contributing to the personalization strategy. Most will also need to change their priorities, operating practices, and performance metrics to fuel progress, as will their extended circle of senior experts and advisers: the general counsel, data leaders, and board members.
CEOs must provide the rationale and the vision: how personalization will distinguish the brand’s value proposition, how the company must rethink its investment priorities, and what new performance targets everyone should aim for.
CFOs need to have a clear understanding both of the critical-path investments for putting the right enablers in place and of how they should be sequenced.
Having a “stand-alone” personalization P&L that measured progress against the overall business case, both in topline impact and costs, was an important tool for informing these discussions.
CSOs (Chief Strategy Officers) provide fresh perspectives on the threats and opportunities arising from market changes, competitors, technology, and regulation.
COOs face a never-ending challenge: making sure that increasing the variation in customer experiences does not create diseconomies of scale. It’s up to them to figure out how to use intelligence and automation to cut the costs of adding more complexity.
CIOs are becoming more like product managers, responsible for partnering to deliver value with the lines of business and support functions—meeting their needs, at cost, while steadily increasing that value.
These relatively new roles (sometimes split between a chief digital officer and a chief data and analytics officer) are gaining ground as top executives and boards realize how critical data and analytics are to their businesses. These leaders must develop a practical (and funded) road map to elevate the value of data assets and build the talent and tools needed to amass real-time intelligence and insights. They also often design and deliver front-line digital experiences. Chief digital and analytics officers (CDAOs) can therefore have a dual function: supporting the entire business with data and analytics while also managing their own channels.
The key to making personalization affordable is making it largely self-funding. Many personalization leaders implement programs where up-front investments are fully funded by the margin from in-year revenue growth.
“It’s important to rapidly put points on the board before you ask for more resources. Create a self-funding mechanism for the next set of use cases. Once the personalization team establishes a track record of value delivery, it becomes easy to secure resources.”
The winning organization of the future will look more like a collection of jazz ensembles than a symphony orchestra. Functional barriers will be reduced. Different specialties will work in more permanent teams around specific customer opportunities. Customer contact will be continuous. Information will be current, rich, and available to all.
In the near future, more companies will be combining smart integration and automation with new ways of working. This will further compress cycle times. The next generation of leaders in digital, marketing, data and analytics, tech, and operations will have grown up operating in agile pods, accustomed to collaborating across functional silos. The personalization leaders of the future will continue to shrink the test-and-learn timelines to hours and grow the number of monthly experiments to thousands, and they will do so with a fraction of the effort required today.
Grow your base of customer relationships and knowledge. Accelerate the velocity of your testing and learning efforts. Look for opportunities to streamline handoffs and automate manual work and unlock the potential of new digital and AI tools. Mobilize to build trust through better experiences, and then challenge your teams to constantly improve those experiences through continuous experimentation. Instead of saying “we are already doing personalization,” ask your teams how you could be doing even more of it.