iDarpan: How Mirror Worlds and Digital Twins will Revolutionise eCommerce

Published August 30-September 11, 2023

In the Future

This is a short story set in the near future.

Jeni, the eCommerce head at A1 Books who has just recently deployed a revolutionary system, iDarpan, is brimming with anticipation to witness its performance in action.

On a calm evening across the city, Arun, an enthusiastic reader and a regular patron of the A1 Books website, nestles comfortably for his nightly digital exploration. Unbeknownst to him, as his fingers dance on his tablet’s screen, his digital alter ego, ArunDT, springs to life in the iDarpan mirror world, a virtual echo of the A1 Books store.

The digital twin, ArunDT, is a precise replica of Arun’s literary predilections and buying tendencies. Interacting in real-time with A1 Books’ virtual catalog, ArunDT began to browse through a collection of the latest mystery novels—Arun’s favourite genre. It is a step ahead of Arun, anticipating his actions and ready with the next books Arun needs to buy.

Back in the physical world, Arun activates the A1 Books app on his tablet. A personalised message welcomes him, followed by a meticulously curated list of mystery novels, including beloved authors and promising newcomers that cater to his literary palate. His eyes sparkle with delight, wondering how the website has intuited his reading mood. He eagerly adds several books to his cart.

Back at the helm, Jeni observes the seamless operation from the control centre, a satisfied smile gracing her face. iDarpan is performing flawlessly. Its AI has astutely predicted Arun’s preferences, enhancing his browsing experience and increasing his likelihood of making a purchase. The promise of a superior, tailored customer experience is being fulfilled.

As Arun lingers on the site, real-time insights gleaned from ArunDT enables A1 Books to offer timely recommendations. The moment Arun concludes perusing a book description, a tactfully placed pop-up emerges, suggesting another novel that might pique his interest. Intrigued, Arun engages, eventually adding two more books to his cart than he had initially anticipated.

The results delight Jeni. Arun’s browsing session duration has lengthened by 15%, and the value of his purchase escalated by 30%. The credit, she knows, is due to the iDarpan system and ArunDT’s predictive accuracy. iDarpan is fulfilling its promise as an experience gamechanger and a profit amplifier – PxL, as Jeni affectionately dubbed it. It is the dawn of a new eCommerce era.

That night, Arun drifts to sleep, his dreams populated by the enthralling narratives and new worlds awaiting him in his new books. In the digital expanse, ArunDT emulates its counterpart, metaphorically shutting down for the day. However, in a few days, when Arun will be ready to reach for his tablet again to uncover more literary gems, ArunDT would be primed and prepared, steps ahead, armed with a perfectly tailored just-in-time email to seamlessly transition Arun’s commerce experience to his inbox. For Jeni, iDarpan is her secret weapon, working tirelessly to ensure that Arun and every customer’s experience with A1 Books is nothing short of exceptional.

The rest of this series is about how to bring this future to life.

The Idea

Recently, I came across this in Scientific American:

Imagine Martha, an octogenarian who lives independently and uses a wheelchair. All objects in her home are digitally catalogued; all sensors and the devices that control objects have been Internet-enabled; and a digital map of her home has been merged with the object map. As Martha moves from her bedroom to the kitchen, the lights switch on, and the ambient temperature adjusts. The chair will slow if her cat crosses her path. When she reaches the kitchen, the table moves to improve her access to the refrigerator and stove, then moves back when she is ready to eat. Later, if she begins to fall when getting into bed, her furniture shifts to protect her, and an alert goes to her son and the local monitoring station.

The “spatial computing” at the heart of this scene is the next step in the ongoing convergence of the physical and digital worlds. It does everything virtual-reality and augmented-reality apps do: digitize objects that connect via the cloud; allow sensors and motors to react to one another; and digitally represent the real world. Then it combines these capabilities with high-fidelity spatial mapping to enable a computer “coordinator” to track and control the movements and interactions of objects as a person navigates through the digital or physical world. Spatial computing will soon bring human-machine and machine-machine interactions to new levels of efficiency in many walks of life, among them industry, health care, transportation and the home. Major companies, including Microsoft and Amazon, are heavily invested in the technology.

…As is true of virtual and augmented reality, spatial computing builds on the “digital twin” concept familiar from computer-aided design (CAD). In CAD, engineers create a digital representation of an object. This twin can be used variously to 3-D-print the object, design new versions of it, provide virtual training on it or join it with other digital objects to create virtual worlds. Spatial computing makes digital twins not just of objects but of people and locations—using GPS, lidar (light detection and ranging), video and other geolocation technologies to create a digital map of a room, a building or a city. Software algorithms integrate this digital map with sensor data and digital representations of objects and people to create a digital world that can be observed, quantified and manipulated and that can also manipulate the real world.

As I was reading this, I started thinking about the possibility of digital twins in marketing. I know the idea is used in manufacturing, but what about marketing? This is what I will explore in this series. Could brands and marketplaces who have huge data about us create a digital twin for each of us? Could they then use that to predict our next best action? Could the digital twin help them better personalise their offerings to us? In some ways, the adtech platforms are perhaps already doing this when they show us ads. Could this be expanded to martech? The way Generative AI predicts the next word using large language models, could a “large customer model” help marketers predict the next step in customer journeys? Like what some of the chess software does, it could even simulate various paths customers are likely to take. What would such a future of “digital twins” look like? I think of this as “iDarpan”, inspired by an idea of “Mirror Worlds” that I had come across 30 years ago. [Darpan means mirror in Hindi.]

Mirror World – 1

As I was thinking about digital twins, my mind immediately went to a book I had read almost three decades ago, “Mirror Worlds”, by David Gelernter. It was published in 1991 – before the Internet and Web as we know it came into our lives, and just around the time when I was leaving NYNEX and embarking on my entrepreneurial journey. From the book’s description:

Technology doesn’t flow smoothly; it’s the big surprises that matter, and Yale computer expert David Gelernter sees one such giant leap right on the horizon. Today’s small scale software programs are about to be joined by vast public software works that will revolutionize computing and transform society as a whole. One such vast program is the “Mirror World.”

Imagine looking at your computer screen and seeing reality–an image of your city, for instance, complete with moving traffic patterns, or a picture that sketches the state of an entire far-flung corporation at this second. These representations are called Mirror Worlds, and according to Gelernter they will soon be available to everyone. Mirror Worlds are high-tech voodoo dolls: by interacting with the images, you interact with reality. Indeed, Mirror Worlds will revolutionize the use of computers, transforming them from (mere) handy tools to crystal balls which will allow us to see the world more vividly and see into it more deeply. Reality will be replaced gradually, piece-by-piece, by a software imitation; we will live inside the imitation; and the surprising thing is–this will be a great humanistic advance. We gain control over our world, plus a huge new measure of insight and vision.

In this fascinating book–part speculation, part explanation–Gelernter takes us on a tour of the computer technology of the near future. Mirror Worlds, he contends, will allow us to explore the world in unprecedented depth and detail without ever changing out of our pajamas. A hospital administrator might wander through an entire medical complex via a desktop computer. Any citizen might explore the performance of the local schools, chat electronically with teachers and other Mirror World visitors, plant software agents to report back on interesting topics; decide to run for the local school board, hire a campaign manager, and conduct the better part of the campaign itself–all by interacting with the Mirror World.

Gelernter doesn’t just speculate about how this amazing new software will be used–he shows us how it will be made, explaining carefully and in detail how to build a Mirror World using technology already available. We learn about “disembodied machines,” “trellises,” “ensembles,” and other computer components which sound obscure, but which Gelernter explains using familiar metaphors and terms. (He tells us that a Mirror World is a microcosm just like a Japanese garden or a Gothic cathedral, and that a computer program is translated by the computer in the same way a symphony is translated by a violinist into music.)

Mirror Worlds offers a lucid and humanistic account of the coming software revolution, told by a computer scientist at the cutting edge of his field.

Here is Gelernter’s idea from his book: “A Mirror World is an ocean of information, fed by many data streams. Some streams slowly represent hand-entry of data at computer terminals; they flow slowly. Others are fed by automatic data-gathering in and monitoring equipment, like the machinery in a hospital’s intensive care unit, or weather-monitoring unit, or traffic-volume sensors installed in roadways. These streams may be so fast-rushing that they threaten to overwhelm the main programs with information tidal waves. The solution is to connect Mirror Worlds to fast-rushing data streams via a sort of software hydroelectric plant. Such programs are designed to sift through complex floods of data looking for trends and patterns as they emerge. They are constructed as layered networks. Data values are drawn in at the bottom and passed upwards through a series of data-refineries, which attempt to convert them into increasingly general and comprehensive chunks of information. As low-level data flows in at the bottom, the big picture comes into focus at the top. A Mirror World is a two-faced duality. You can look at it as a datapool, as a detailed historical archive; or you can look at it as a datafilter, capturing and synopsizing the current state of a complicated system right now.”

Mirror World – 2

I asked ChatGPT and Bard to summarise Gelernter’s Mirror Worlds idea.

[ChatGPT]

The book presents several key ideas:

Mirror Worlds: The core concept introduced by Gelernter is that of “Mirror Worlds.” These are software models of some part of reality, such as a city, a financial market, or an industrial process. These digital reflections of reality can be interactively explored and manipulated, providing real-time, operational understanding of the systems they model. They effectively offer a detailed and comprehensive overview of complex, real-world systems.
Prediction and Planning: Gelernter argues that Mirror Worlds can be instrumental in decision-making processes, serving as crucial tools for planning and prediction. For instance, if a city is modeled as a Mirror World, city planners could run simulations to predict the effects of different policies or decisions, ultimately leading to better outcomes.
Data Visualization and Accessibility: Gelernter suggests that the traditional, text-based interface of the computer world would evolve towards more graphical, interactive, and intuitive representations, making data more understandable and accessible. This is in line with the creation of Mirror Worlds, as these environments would require complex data to be visualized in user-friendly ways.
Social and Political Implications: The book also delves into the social and political implications of Mirror Worlds. Gelernter envisions a future where these software models could promote transparency, allowing ordinary citizens to explore the workings of various systems such as the government. However, he also warns of the potential for misuse, including privacy breaches and information overload.
Ubiquitous Computing: Although not explicitly named as such in the book, Gelernter essentially predicts the rise of ubiquitous computing, where computing is embedded into everyday objects and activities. The accessibility and interactivity of Mirror Worlds would be part of this pervasive digitization of the world.

It’s important to note that while Gelernter wrote this book in 1991, many of the ideas presented have since found parallels in modern technological developments, such as the construction of digital twins, sophisticated simulations, augmented reality, and virtual reality.

[Bard]

Mirror Worlds is a book by David Gelernter that explores the potential of software to create virtual representations of the real world. These representations, called Mirror Worlds, would be updated in real time and would allow users to interact with them in a variety of ways.

Gelernter argues that Mirror Worlds have the potential to revolutionize the way we interact with the world around us. For example, a Mirror World of a city could be used to track traffic patterns, plan routes, or visualize crime data. A Mirror World of a company could be used to track inventory, manage finances, or simulate product development.

Gelernter also argues that Mirror Worlds have the potential to change the way we think about the world. By providing us with a virtual representation of reality, Mirror Worlds could help us to understand the world in new ways and to make better decisions.

Here are some of the key ideas of Mirror Worlds:

Mirror Worlds are virtual representations of the real world that are updated in real time.
Mirror Worlds can be used to interact with the real world in a variety of ways.
Mirror Worlds have the potential to revolutionize the way we interact with the world around us.
Mirror Worlds have the potential to change the way we think about the world.

“Mirror Worlds” was published in 1991, and many of the ideas that Gelernter discusses have since become reality. However, there are still many challenges that need to be overcome before Mirror Worlds can reach their full potential. For example, Mirror Worlds require a lot of computing power and bandwidth. They also need to be able to accurately represent the real world in a way that is useful to users. Despite these challenges, Mirror Worlds are a promising technology with the potential to change the way we interact with the world around us.

[Rajesh]

Mirror Worlds is an idea I have been fascinated by for a very long time.

Mirror World – 3

I first wrote about the idea two decades ago in Constructing the Memex: Mirror Worlds” (in 2003): “Gelernter’s Mirror World seems to describe the blogosphere of today! There is an ocean of information out there, with the RSS streams representing the data streams. The filters are the actions of bloggers. The layers are made up by the circle of bloggers. Think of them as concentric circles of bloggers with the radius determined by their degree of separation from us. So, we can access the vast pool of information through search engines like Google and directories like Yahoo, or we can use the work of bloggers as information filters.” More: “Gelernter envisioned all kinds of Mirror Worlds for hospitals, cities, universities, communities. This is how the blogosphere is emerging. Today, we see blogs created by people. Already, blogs are being embellished by photos from digital cameras. Soon, we can expect sensors to send out information to specific log pages. Each entity whether human or machine or an aggregation can be expected to have a weblog, which is a representation of itself. For all practical purposes, the weblog is akin to a What’s New page of a website. It mirrors the latest state of the entity, with access to the history also. The weblogs become the information sources, publishing their data in RSS, which we can subscribe to as part of our Memex. The interaction of the weblogs with the rest of the blogosphere creates the Mirror Blog.”

I then wrote in an essay on Next-Generation Networks in 2005: “Over the past decade, we have been spending an increasing amount of our time in so-called cyberspace. Companies and individuals have created virtual representations of their products and services. Our communications have also moved to conversing with identities (email IDs, IM monikers, SMSing to mobile numbers) rather than directly with people. David Gelernter’s idea of “Mirror Worlds” takes this to its logical conclusion: we will have a parallel world that we will increasingly inhabit which is a copy of the real world. Today, maps can provide us the spatial copy. But they do not give us the real-time component. That is where a mix of next-generation mobiles, sensors and user-generated content will come in and embellish the other world. So, Mirror Worlds are microcosms of all that we see around us as updated as the real world that they resemble. These Mirror Worlds are accessible to us through screens on the devices we have our mobiles, computers, and perhaps, networked TVs.”

I had also written about it in the context of campaigning and direct marketing in 2009:

During the elections, I was struck by the lack of databases and real-world linkages. Such tools could be great assets for both campaigning and direct marketing. I think of this as a “mirror world”– a virtual replica of the real world along multiple dimensions:

start with maps
add a layer of establishments (buildings, schools, retail outlets, roads, etc.)
overlay this with the voter database that one can get from the Election Commission. The voter database has names of people, their addresses, gender and age.
add the actual voting numbers based on the data published from the EC post-election (can also incorporate historical data to get trends; need to take into account the delimitation)
integrate the socio-demographic and development data that is available from census, various government sites (and collated by independent companies)
finally, buy contacts lists of people with information of their digital identity (email IDs, mobile numbers)
this database can then be continuously updated based on user interactions, thus enhancing people profiles

What seemed like science fiction once is on its way to becoming reality soon. The “Mirror World” idea and its “Digital Twin” companion can transform marketing.

Digital Twins – 1

An idea independent of and yet connected to “Mirror World” is that of a “Digital Twin.”

Wikipedia: “A digital twin is a digital representation of an intended or actual real-world physical product, system, or process (a physical twin) that serves as the effectively indistinguishable digital counterpart of it for practical purposes, such as simulation, integration, testing, monitoring, and maintenance. The digital twin has been intended from its initial introduction to be the underlying premise for Product Lifecycle Management and exists throughout the entire lifecycle (create, build, operate/support, and dispose) of the physical entity it represents. Since information is granular, the digital twin representation is determined by the value-based use cases it is created to implement. The digital twin can and does often exist before there is a physical entity. The use of a digital twin in the create phase allows the intended entity’s entire lifecycle to be modeled and simulated. A digital twin of an existing entity may be used in real time and regularly synchronized with the corresponding physical system. Though the concept originated earlier, the first practical definition of a digital twin originated from NASA in an attempt to improve physical-model simulation of spacecraft in 2010. Digital twins are the result of continual improvement in the creation of product design and engineering activities. Product drawings and engineering specifications have progressed from handmade drafting to computer-aided drafting/computer-aided design to model-based systems engineering and strict link to signal from the physical counterpart.”

IBM: “A digital twin is a virtual model designed to accurately reflect a physical object. The object being studied—for example, a wind turbine—is outfitted with various sensors related to vital areas of functionality. These sensors produce data about different aspects of the physical object’s performance, such as energy output, temperature, weather conditions and more. This data is then relayed to a processing system and applied to the digital copy. Once informed with such data, the virtual model can be used to run simulations, study performance issues and generate possible improvements, all with the goal of generating valuable insights—which can then be applied back to the original physical object. Although simulations and digital twins both utilize digital models to replicate a system’s various processes, a digital twin is actually a virtual environment, which makes it considerably richer for study. The difference between digital twin and simulation is largely a matter of scale: While a simulation typically studies one particular process, a digital twin can itself run any number of useful simulations in order to study multiple processes.”

Rameshwar Balanagu writes:

The ability to ingest data in real-time and technological advancements that made it more affordable spurred the adoption of digital twin technology for building digital replicas. Progress in artificial intelligence has led to optimization, better predictability and improved value. Once a digital twin is analyzed and better understood, both proactive and predictive maintenance can be done to respond to and recreate unplanned situations, thereby making timely decisions possible. This can help optimize a company and also contribute to the bottom line by making customers happy through more personalized experiences.

…A digital twin could be considered a lifeline for an enterprise, providing real-time insights and the ability to predict the next best offer or fix leaks in a process. It can also help optimize a network and perform automation akin to surgery to remove blockages or bottlenecks, such as in fulfilling an order in a timely manner. This, in turn, leads to greater customer satisfaction and protects the customer base….It can help customer experience teams by recommending the shortest path to customer satisfaction, highlight pain points, optimize service and suggest areas of automation.

McKinsey: “A digital twin could provide a 360-degree view of customers, including all the details that a company’s business units and systems collect about them—for example, online and in-store purchasing behavior, demographic information, payment methods, and interactions with customer service. It would also generate insights derived from the data, such as the average length of a customer service call. AI use cases leveraging the twin could include customer churn propensity models or a basket of the next products a customer would be likely to buy…When companies interconnect two or more digital twins, they can simulate the complex relationships among different entities and generate richer behavioral insights for even more sophisticated use cases and greater value…As organizations begin to connect these multiple twins of different business domains, functions, and operational ecosystems, the enterprise metaverse could start to take shape. A retailer, for instance, could connect the digital twin of its retail store to digital twins of its warehouses, supply chain, call center, and more until every part of the organization was replicated, sharing insights, simulating scenarios, and enabling automation and AI use cases.”

Amplitude offers this helpful chart:

The Economist (in May 2022): “Digital Twins are virtual representations of a component, a device or even an entire production line in a factory. Fed with data from sensors installed on its physical opposite number, the digital version can be used to plan maintenance, spot any emerging problems and simulate the effect of upgrades and design changes. Aircraft engines in particular are followed by their manufacturers in this way. General Electric, Rolls-Royce and Pratt & Whitney, the three principal firms involved, maintain simulations of individual engines at engineering centres on the ground. These suck in real-time data from their counterparts in the air. At Cranfield University, in Britain, however, a group of engineers hope to go further than this. They propose expanding the idea of digital twinning to produce what they rather grandiosely refer to as a “conscious aircraft”. This would involve creating a digital twin of an entire plane by merging its various monitoring systems, and interpreting the result using artificial intelligence (AI).”

ChatGPT provides a list of use cases for digital twins.

Manufacturing: This is one of the earliest adopters of digital twin technology. It’s used to optimize production processes, troubleshoot problems, predict equipment failures, and reduce downtime. A digital twin of a manufacturing process can simulate different scenarios and their outcomes, enabling more efficient production planning.
Aerospace and Defense: Digital twins have been used to design, simulate, and maintain complex systems such as aircraft and spacecraft. They can also provide real-time diagnostics, predictive maintenance, and performance optimization of these complex machines.
Automotive: In the automotive industry, digital twins are used for virtual testing of vehicle designs before physical prototypes are built. This can save time and resources by identifying potential issues early in the design process. They can also be used for predictive maintenance of vehicles.
Energy and Utilities: Digital twins can simulate and monitor energy generation and distribution systems like power plants, solar farms, and electrical grids. They can be used to optimize energy generation and distribution, improve sustainability, and anticipate and solve potential issues before they occur.
Healthcare and Biomedical: Digital twins are being used to model human physiology in detail, which can be used for personalized medicine, testing treatments, or studying diseases. They can also be used for hospital management to optimize operations and patient care.
Smart Cities: Cities are using digital twins to simulate traffic patterns, utilities, and public services. This can help city planners optimize infrastructure, improve sustainability, and plan for various scenarios such as natural disasters or large public events.
Supply Chain and Logistics: Digital twins can simulate and monitor every aspect of a supply chain, from manufacturing to delivery. This can be used for optimizing logistics, predicting potential disruptions, and improving efficiency.
Construction and Infrastructure: Digital twins of buildings or infrastructure projects can be used to improve the design process, simulate different construction strategies, monitor the health of the structure, and manage maintenance tasks.
Oil and Gas: Digital twins can be used for the design and operation of oil and gas equipment and facilities. They can simulate and predict the performance of oil wells, pipelines, and refineries under various conditions.
Retail and E-commerce: Digital twins can model customer behavior, optimize store layouts, improve supply chain management, and enhance the overall customer experience.

Digital Twins – 3

Unity offers use cases from retail:

Design and planning – Create 3D virtual stores to visualize and simulate the optimal experience prior to construction; utilize planograms and store planning applications to maximize space, improve efficiency, and collaborate remotely.
Sales and marketing – From integrating 3D assets into e-commerce sites to creating virtual showrooms in VR, retailers can use digital twins to increase conversion and make purchase decisions more accurate, which limits returns and mitigates the environmental footprint of e-commerce.
Operations – Digital twins of product SKUs and stores can assist in the development of myriad applications to improve operational efficiency, from autonomous checkout to intelligent in-store navigation.

From “New Paradigm of Data-Driven Smart Customisation through Digital Twin”: “Big data is one of the most important resources for the promotion of smart customisation. With access to data from multiple sources, manufacturers can provide on-demand and customised products. However, existing research of smart customisation has focused on data generated from the physical world, not virtual models. As physical data is constrained by what has already occurred, it is limited in the identification of new areas to improve customer satisfaction. A new technology called digital twin aims to achieve this integration of physical and virtual entities. Incorporation of digital twin into the paradigm of existing data-driven smart customisation will make the process more responsive, adaptable and predictive. This paper presents a new framework of data-driven smart customisation augmented by digital twin.”

A research paper published in May 2022 summarises applications of digital twins across industries. From the section on retail: “The retail sector includes all the stores and shops that sell the products or goods to the consumer/customer directly. The success of any retail business depends on its customer base; attracting new customers is not enough, and maintaining/sustaining the existing ones by providing the best customer support is equally important. Digital Twin [DT] technology in the retail sector offers great potential in terms of consumer experience and marketing. By creating a DT of the customer, retail shops will be able to provide a unique and tailored customer experience based on customers’ patterns of interest. Customer satisfaction can be enhanced further by providing them with relevancy-based suggestions based on their DT without irritating them with multiple recommendations. DT technology enables retailers to turn their products into a dynamic platform that continuously acquires data regarding customers’ needs and buying behavior in order to provide them with better products and services. Combining the data from a customer’s DT with machine learning to understand the customer’s behavior will be beneficial for the customers as well as the retailers.” From one of the referenced papers: “With digital twins, consumer activities can be monitored through their online personas or avatars in online stores and social media. While they represent limited information about an individual, it is likely personal avatars will become an important element to retail in the future through the further development of sophisticated digital assistants.”

Can we create a mirror world with digital twins of customers and predict (probabilistically) what each of them is likely to do next? This is the idea I call “iDarpan”.

Before I get to iDarpan, I want to take a small detour to mention a book I came across recently. Here is its beautiful cover:

From the book’s introduction: “Just as great medieval cathedrals were raised by architects, masons, geometers and bishops to give humankind a glimpse of the infinite, supercomputers are the cathedrals of the information age, where novel worlds of endless variety, even entire universes, can be simulated within these great engines of logic, algorithms and information. You can also re-create the inner worlds of the human body, and not just any body, or an average body, but a particular person, from their tissues and organs down to the molecular machines at work within their cells, their component proteins along with their DNA. The eventual aim of this endeavour is to capture life’s rhythms, patterns and disorders in a computer, not just of any life or an average life, but of one particular body and one particular life—yours.” More: “If we are to create digital twins, how well do we have to know ourselves? To create Virtual You, we need to understand how much data and what kinds are sufficient for a digital twin to be animated by a computer. As Aristotle once remarked, knowing yourself is the beginning of all wisdom.”

The same can be said about the “Virtual You” in marketing and iDarpan.

Consumer Copies

Imagine being able to use Generative AI to “copy” the behaviour of every consumer in the form of a “digital twin” and immerse a digital twin for every customer into a “mirror world.” This is what I think of as “iDarpan”. What would this world of marketing look like? Let’s begin by looking at some of the current thinking.

This is from a 2022 Gartner document entitled “A Digital Twin of a Customer Predicts the Best Consumer Experience”: “A digital twin of a customer is more than the concept, familiar to marketing executives, of a “persona.” Instead of merely collecting data points, it provides context and predictions of future behaviors. It uses both online and physical interactions and it’s dynamic, updating as new information comes in and recognizing that a single person can exemplify more than one persona … and those personas may shift over time. The twin collects enough data to accurately simulate the customer experience — moving them to more positive consumer-business relationship stages by fulfilling expectations and providing unexpected moments of delight.” It adds: “A digital twin can help businesses improve that experience and retain loyalty while also identifying the friction points that lead customers to defect to competitors…The digital twin synchronizes with its physical representation using real-time data inputs and event-stream processing, where action is taken on a series of data points originating from a system that continuously creates data.”

I asked ChatGPT and Bard how the ideas of mirror world and digital twins can come together for marketing.

[ChatGPT]

Mirror Worlds and Digital Twins, both as software models of real-world systems, can provide invaluable insights for marketing in several ways:

Customer Behavior Modeling: Digital Twins could model customer behaviors based on various parameters, which might include demographics, purchase histories, social media interactions, and more. By simulating different marketing strategies or campaigns, marketers can gain insights into potential customer responses and refine their strategies accordingly.
Market Simulation: Mirror Worlds and Digital Twins could simulate an entire market segment or even a global market. By analyzing these simulations, marketers can better understand market trends, identify potential gaps in the market, and predict how changes might affect market dynamics. This could be particularly useful in product development, pricing strategies, and planning market entry or expansion.
Product Development and Testing: In the development phase of a product or service, Digital Twins could be used to simulate and test different design iterations and features in various market conditions. This can allow marketers to gauge potential reception and usability before the actual launch.
Personalized Marketing: With the help of detailed customer models, companies can offer highly personalized marketing experiences. For example, a Digital Twin of a customer can provide insights into the individual’s preferences, enabling marketers to tailor specific messages, offers, or product recommendations.
Immersive Customer Experiences: Mirror Worlds can be used to create immersive, interactive experiences that engage customers in unique ways. For instance, a virtual store where customers can explore and interact with products virtually before purchasing.
Crisis Management and Scenario Planning: Digital Twins can simulate potential crises or unexpected market changes, helping marketers to plan effective strategies and responses.
Supply Chain and Logistics: A Digital Twin of a company’s supply chain can help marketers understand how changes in the supply chain might impact product availability and delivery times, which are crucial elements of customer satisfaction.

[Bard]

The Mirror World and Digital Twins ideas can be applied to marketing in a number of ways, including:

Personalized marketing: Digital twins can be used to create personalized marketing campaigns that are tailored to the individual needs and interests of each customer. This can be done by using data from the digital twin to track customer behavior, preferences, and interactions with the brand.
Predictive analytics: Digital twins can be used to predict customer behavior and to identify potential opportunities for cross-selling or upselling. This can be done by using data from the digital twin to track customer engagement, purchase history, and product preferences.
Virtual reality and augmented reality: Mirror Worlds and Digital Twins can be used to create virtual reality and augmented reality experiences that can be used to promote products and services. This can be done by using data from the digital twin to create realistic and immersive experiences that can help customers to better understand the product or service.
Customer service: Digital twins can be used to improve customer service by providing customer support agents with real-time data about customer interactions with the brand. This data can be used to help agents to resolve customer issues more quickly and efficiently.

Here are some specific examples of how Mirror Worlds and Digital Twins are being used in marketing today:

Nike: Nike is using digital twins to create personalized marketing campaigns for its customers. The company uses data from its digital twins to track customer behavior, preferences, and interactions with Nike products. This data is then used to create personalized marketing campaigns that are tailored to the individual needs and interests of each customer.
Amazon: Amazon is using digital twins to predict customer behavior and to identify potential opportunities for cross-selling or upselling. The company uses data from its digital twins to track customer engagement, purchase history, and product preferences. This data is then used to predict which products customers are likely to be interested in, and to create targeted marketing campaigns.
Walmart: Walmart is using virtual reality and augmented reality to promote its products and services. The company has created a virtual reality experience that allows customers to walk through a Walmart store and see products in 3D. The company is also using augmented reality to allow customers to see how products will look in their homes.

Mirror World, Digital Twin, Metaverse, and Generative AI – all of these technologies can be combined together to reimagine marketing and eCommerce.

A New Era

The foundation of our new era in eCommerce is formed by four essential building blocks:

Mirror World: This is a digital rendering of the physical world, embodying a virtual landscape where millions of customers interact with brands. It symbolises a virtual reflection of our eCommerce business ecosystem, where every customer is unique and is represented by a digital entity with its own needs, preferences, and behavioural patterns. This dynamic setting allows customers, represented by their digital twins, to engage with a brand’s products or services, building a thriving, interactive space.
Digital Twins for Customer Profiling: These software models are personalised depictions of individual customers, amalgamating their demographic, psychographic, and behavioural data. The creation of these digital twins allows brands to tailor predictions for the next best action for each customer, offering personalized recommendations and services that significantly enhance their experience and increase conversion rates.
Metaverse Interactions: As a more engaging term for the Mirror World, the Metaverse signifies an immersive arena where brands and digital twins of customers co-exist and interact. These interactions illuminate customer behaviour, preferences, and purchasing patterns, arming brands with the knowledge to refine marketing strategies, enhance product offerings, and personalise services.
Generative AI for Predictive Modelling: The multitude of data emanating from interactions in the Metaverse is processed using Generative AI. This AI model identifies patterns, deduces inferences, and forecasts based on the data. It helps predict the customers’ next move, simulate diverse scenarios, and provide a test bed for marketers to assess varying strategies.

The combination of these building blocks paves the way for an interactive, immersive, and personalised eCommerce experience, known as the “iDarpan”. This name encapsulates the idea of the digital ecosystem reflecting the customer’s unique needs, behaviours, and interactions. It is, in essence, a mirror image of their digital persona.

iDarpan is a digitally powered twin, enhanced by Generative AI, projecting a comprehensive model of each customer to suggest their next course of action, maximizing their lifetime value. It resonates with elements from the gaming industry and software agents, offering a highly tailored, personalized customer journey. It serves as a test lab for marketers, enabling them to experiment with various paths to enhance customer loyalty and boost lifetime value.

Furthermore, iDarpan seamlessly aligns with the objectives of Martech 2.0, blending in with concepts like Inbox Commerce, Adtech-Style Performance Pricing, and profitable growth for brands. It supports the pursuit of near-zero customer acquisition costs and profitability amplification by focusing on fostering existing customers and maximizing their lifetime value.

Making It Happen

Building an iDarpan system for an eCommerce business involves a comprehensive, data-driven process. Here are the steps brands can undertake to create their personalised digital customer ecosystem:

Comprehensive Data Collection & Seamless Integration: Begin with amassing and integrating diverse data types from multiple touchpoints, including transaction records, browsing patterns, demographic details, customer feedback, and social media interactions. Emphasise on first-party data and enrich it with zero-party data through progressive profiling (via Engagement Footers in AMP Emails) to obtain a more holistic view of customers.
In-depth Customer Profiling: Utilise the amassed data to construct detailed customer profiles that encapsulate their behaviours, preferences, and purchasing patterns. These profiles lay the groundwork for the digital twins and should reflect each customer’s uniqueness accurately.
Crafting the Digital Twins: Leverage machine learning algorithms to forge digital twins based on these intricate profiles. Ensure each digital twin emulates their real-world counterpart’s actions and reactions authentically. Draw inspiration from multi-player gaming environments where software agents govern the entities’ actions.
Constructing the Mirror World: Concurrently, conceive and develop the Mirror World — the interactive realm where digital twins engage with the brand’s online assets. Employ probabilistic models to simulate a multiverse environment, enabling digital twins to explore diverse paths on their customer journeys. Adjust digital actions based on actual customer behaviours for enhanced realism.
Implementing Generative AI: Deploy generative AI models to anticipate future customer actions via a Large Customer Model (LCM). Similar to a Language Model, an LCM interprets actions instead of text and images, examining customer activities on digital properties and through push channels. These AI-generated predictions inform CRM teams and recommendation engines, shaping future customer interactions.
Engineering an Enhanced Customer Experience: Utilise insights from the iDarpan system to amplify the customer experience, foster personalisation, bolster marketing effectiveness, and boost sales. The goal is to create a dynamic, responsive environment that adapts to each customer’s evolving preferences.
Continuous Monitoring & System Refinement: Maintain an active feedback loop to continuously fine-tune the iDarpan system based on actual customer behaviours. Consider integrating external data, such as weather conditions or festive occasions, that can potentially influence customer behaviour, ensuring the digital twins stay relevant and accurately represent their real-world counterparts.

iDarpan is thus about constructing the equivalent of a “Virtual You” for every customer and situating it in the context of a software-powered replica of their digital (and perhaps even physical) storefront.

Large Customer Models

One concept that needs some elaboration is that of the relationship between Large Language Models (LLMs) and Large Customer Models (LCMs). The key to this analogy lies in the predictive capacity of these models, where LLMs anticipate the next words in a text sequence and LCMs predict the next actions in a customer journey.

The transformative potential of Large Language Models (LLMs) and Large Customer Models (LCMs) pivots around their predictive capacity. LLMs like GPT-4 excel in generating human-like text based on prior sequences of words. By analysing extensive text data, these models learn the probability of particular words following a given string of text, effectively predicting what comes next. Companies like OpenAI leverage LLMs for various purposes, including drafting emails, writing code, creating written content, and more. The power of these models stems from their ability to capture context, recognise patterns, and make educated predictions about subsequent textual sequences.

LCMs, in parallel, function on similar principles but operate in a different ‘language’—the language of customer behaviour. They predict the next action in a customer’s journey by analysing extensive datasets of customer interactions and behaviours. The aim is to recognise patterns and sequences leading to specific outcomes. A prime example would be a good recommendation engine, which predicts what product a customer is likely to be interested in next, based on their browsing history and past purchases.

However, LCMs surpass LLMs in their ‘vocabulary’ breadth. They not only consider individual actions, like browsing a product or making a purchase, but also can consider a wide array of external factors. For instance, factors like the time of day, weather conditions, or significant events like holidays and festivals can influence customer behaviours.

What makes LCMs truly powerful is their ability to fuse learnings from vast datasets with real-time behavioural information. They capture the ‘customer language’ at an individual level, learning from each interaction to refine the predictive models. They anticipate behavioural changes, offering marketers the opportunity to adjust strategies proactively.

LCMs are inherently designed for continuous learning and improvement. They allow eCommerce businesses to dynamically optimise customer interactions based on real-time behavioural changes. As the LCM evolves and adapts over time, it refines its predictions, making every subsequent customer interaction more personalised and relevant.

In the fiercely competitive eCommerce arena, where customer loyalty is a prize to be won, LCMs can be game changers. When integrated within systems like iDarpan, they facilitate true one-to-one personalisation. By accurately predicting the next steps in a customer’s journey, brands can create highly customised experiences that resonate with customers, fostering loyalty and driving repeat purchases. In the long run, LCMs can enable brands to carve out a ‘profits monopoly’ (profipoly). As they mature, brands can maximise profitability by effectively nurturing their customer base.

With iDarpan, LCMs can power an eCommerce revolution, driving customer engagement at an individual level and shaping marketing strategies through predictive customer behaviour. This transformative approach can transform the future of eCommerce, mirroring the customer to reflect their needs, preferences, and behaviours, crafting a shopping experience that feels uniquely their own.

Martech 2.0

iDarpan is the latest in a series of innovative marketing concepts that I’ve discussed over recent years. The overarching theme, which I refer to as Martech 2.0, challenges the current imbalance in ad budgets, which are predominantly allocated towards new customer acquisition. This practice often results in AdWaste and subsequent strain on brand profits. Martech 2.0 seeks to answer a constant query among marketers: “How can I reduce the Customer Acquisition Cost (CAC)?” The solution isn’t merely to optimise the ROAS (Return On Ad Spend) on prominent adtech platforms. Instead, it advocates a different approach – —concentrating on harnessing the spending power of existing customers.

Martech 2.0 revolves around increasing sales while reducing marketing costs, thereby enhancing profitability. It is about bringing back existing customers for more, and ensuring they get their friends. It is about laying the groundwork for exponential forever profitable growth, and eventually, a profipoly, by building remarkable products that function as profit generators, enabling brands to establish a competitive edge and ultimately, a profit monopoly (or ‘profipoly’).

Here is a summary of these concepts.

ProfitXL (PxL): The ultimate goal is to revolutionise the Profit & Loss statement by substantially increasing profits. This transformation utilises the SHUVAM framework, comprising Storytelling, Hotlines, Unistack, Velvet Rope Marketing, Acquisitions, and Metrics. It includes advancements like Email 2.0 (utilising AMP for interactive emails) and Loyalty 2.0 (providing Atomic Rewards – micro-incentives – for customer attention and data). Velvet Rope Marketing prioritises the top 20% of customers, who contribute to 60% of the revenue and can yield 200% of profits, acknowledging the long tail phenomenon’s inefficiency.

Inbox Commerce: Minimising the ‘funnel friction’ can be achieved by bringing conversion actions closer to the customer—directly in their inbox. Tactics such as Email Shops, Reactivation Sequences, and Engaging Footers can eradicate three profit pitfalls: a vast majority of email recipients not clicking through to the website, a large fraction of clickers not making a purchase, and a significant portion of email IDs in the database remaining disengaged.

Adtech-Style Martech: By adopting a performance pricing model, martech companies can rectify their past mistakes and benefit from unlimited budgets. This adjustment necessitates a ‘progency’ mindset, merging the product’s power with the creative approach of an agency. This strategy could significantly mitigate the $200 billion AdWaste faced by B2C/D2C companies, providing a sustainable profitable growth model for brands.

iDarpan: The focal point of this series, iDarpan leverages the concepts of mirror worlds, digital twins, and the technologies driving the Metaverse and Generative AI. It helps brands beautify every profit killing customer experience—delighting both the customer and the eCommerce manager. Central to iDarpan is the Large Customer Model (LCM), which predicts and navigates customers on their journey to satisfaction.

Martech 2.0 is about bringing these new ideas to fruition. As Peter Drucker once said, “There are only two things in a business that make money – innovation and marketing, everything else is cost.” Here is my adaptation: “There is only one thing in a business that makes a money machine – innovation in marketing.”

Share this:

Like this: