Introduction
We'll begin this book a bit unusually, with a fictional email from the near future:
To: TomSterne@gmail.com
From: New_Vehicle_Marketing@toyota.com
Date: July 20, 2026
Subject: New Toyota Offer
Dear Mr. Sterne,
You asked that we communicate with you on offers of Toyota automobiles and services, and this email is such a message. From public data and information from the Toyota Connect service on your vehicle to which you subscribe, we have observed that:
- Your 2022 RAV4 Prime is now out of warranty, though it has very low mileage of 12,343 miles over four years (as of yesterday) and a perfect service record at our service centers.
- We noted from public vehicle data that you usually purchase a new vehicle when the previous one is out of warranty.
- We also noted from the data that you have a fully electric car, a Hyundai IONIQ 6, registered in Massachusetts, and owned two fully electric cars before that.
- Your RAV4 Prime, a plug-in hybrid, is regularly plugged in to charge by you or other drivers of the car.
- Toyota Connect tells us that 92% of your trips involve fewer than 200 miles.
- Toyota Connect weight and rearward-facing visibility sensor data tells us that your car is often fully loaded.
- Most of your calls to our technical service line have involved scheduled charging at home to save on utility costs.
Given this information, we would like to offer you a reduced loyalty price on the new 2027 Highlander Platinum all-electric SUV. It has more passenger and storage capacity than the RAV4, and its electric range is well over 200 miles. It has a new, easy-to-use scheduled charging system that makes charging your vehicle at the lowest cost almost automatic. Platinum is the highest trim level on the Highlander electric, consistent with that on your current RAV4 Prime. And we understand that you see the benefits of fully electric vehicles.
Given the low mileage and service record on your RAV4 Prime, we are also prepared to offer you the highest estimated value (subject to an appraisal) on your trade-in. Combining the reduced price on the 2027 Highlander EV Platinum ($52,545) and the trade-in value on your RAV4 Prime ($35,700), you would be able to purchase the new car for less than $17,000, not including tax and licensing fees.
We would be happy to bring the new car to your home for a test drive, appraise your current car's trade-in value, and complete all administrative paperwork, all in one short visit. If this offer appeals to you, please let us know by 31 August 2026. If you have questions about it, please visit our ToyoSmart intelligent chatbot on Toyota.com or the Toyota mobile app and provide your name. It is familiar with the details of the offer, the data used to create it, and the vehicles involved.
Sincerely,
Your Friends at Toyota Motor North America
Such a highly personalized offer is technically possible today, although to our knowledge neither Toyota nor any other automobile companies (including digital native car companies like Tesla) makes such offers. We're not picking on Toyota-we both own and like their vehicles-but they market and sell cars (at least in the US) in the same nonscientific way that other companies do. Most companies in other industries also don't use these personalized and targeted approaches. But why not, given that we've been told for several decades that personalized, targeted marketing, sales, and service are the way to achieve better customer relationships? Why not, given the explosion of available data and the amazing capabilities of analytics, artificial intelligence (AI), and autonomous agents?
These are the questions that inspired us to write this book. We've both closely followed developments in the science and technology of customer relationships for the past several decades, and we've been impressed by the technological progress but underwhelmed by the lack of improvement in customer connections and engagement. It was time that the two of us explored the science of how classic data exploration, analytical AI, and generative AI can make the most of data and analytics to boost customer relationships:
Thomas (Tom) Davenport is a professor who has written many books on analytics, AI, and other business technologies, and he's been interested in how technology affects marketing and other customer-facing functions since the 1990s. However, he remains puzzled about the regular bombardments of marketing and sales materials he receives that have nothing to do with his attributes or interest. Jim Sterne has written several books on such topics as ecommerce, web analytics, and AI in marketing-and speaks and advises often on such topics-but he's also convinced that companies can do much better than they are currently doing in these domains.
Of course, technology is always changing, and it has changed a lot in the past couple of years. Jim is convinced that generative AI just might be the key that unlocks the one-to-one marketing door. Tom is certainly dazzled by what generative AI can do but isn't as sure that it will be enough to break things open. But both of us believe that it's not enough by itself to bring about effective and highly personalized customer relationships.
The science of customer relationships is also evolving rapidly. While companies have long relied on intuition and experience to understand their customers, we're now seeing the emergence of more scientific approaches that combine rigorous data analysis, controlled experiments, and reproducible results.
For example, in the case of the fictional Toyota offer in the previous email, it's likely that the content of the email will have been created by generative AI. Perhaps there might even be agents that would go out to different data sources and pull the needed data together to make the offer. But there are also business and organizational challenges that would have to be overcome before such an offer would be feasible.
For example, this fictional offer is from Toyota Motor North America, rather than a dealer. Who would make such an offer-the company or the dealer? Offers and prices are set by dealers at the moment, which is rather inefficient. It would seem that such a personalized approach made from the company overall would rival the knowledge of a dealer salesperson. And there might be no need for a showroom if cars were brought to the customer directly. Changes of this nature would involve difficult and time-consuming negotiations with the existing dealership structure, and they might never happen at all.
The offer draws on a wide variety of data sources and is processed with a variety of intelligent applications (see Figure I.1). Most of these actually exist today, but they would likely be difficult to integrate in order to create the offer. A household might have cars that are registered to different owners within it, for example. Information from service transactions might readily be captured and summarized by generative AI, but service departments collaborating with marketing and sales functions might be more challenging.
Figure I.1 The hypothetical personalized offer process.
In short, it seems likely that the nontechnical changes to create such an offer might be greater than the technical challenges involved. There should be little doubt that the needed technical capabilities will be available by 2026, but the needed organizational and business capabilities may take much longer to become widespread.
Does this offer constitute customer science? We define that term as a concerted and continuous attempt to understand what customers want and need using data and technology, and to offer them only products, services, and advice that are targeted to meet those desires and requirements. The fictional Toyota email appears to be scientifically targeted to the customer's need for a new car. However, we can't verify that the customer wants a new car-the response or lack thereof to the email could provide clues. And we can't label it "customer science" without seeing a long-term pattern of such marketing and sales approaches. One-off experiments seldom lead to scientific breakthroughs, and one-off targeted offers don't lead to great customer relationships over time.
In this book we describe the technologies, the science, the related business issues, and the organizational and behavioral changes necessary to dramatically improve customer relationships. The one-to-one concept of personalization is discussed in detail, with reflection on why it has not come to pass in most organizations in the more than 30 years since it was defined. But we also discuss other aspects of customer-facing technologies, including the voice of the customer, automation of marketing activities, AI agents and related technologies, customer-facing operations, and many other topics with regard to customer issues. Much of the book is focused on improvements in these areas: better AI, better data, better technologies, better analytics, etc. All of these taken together-along with a sincere desire to improve the lives of customers-constitute customer science.
We've tried to strike a balance in the level of optimism about new technologies, particularly AI and generative AI. We aim to take a scientific approach, examining the evidence both for...
