Memory and Personalization; notes on the evolution of the web
Exploring the web's transformation from a forgetful, static platform to a dynamic, memory-rich ecosystem shaping modern advertising and user experiences
In its infancy, the web lacked memory, memory of its users, making every interaction feel like a first-time visit. This stateless nature of the web was impersonal, with every interaction starting from scratch. Compared to the experience of internet today, it was rather amnesiac. Every transaction was individual and forgotten in its context.
But the invention of cookies changed this forever. Cookies were created in 1994 by Lou Montulli, an engineer at Netscape, while working on the first widely-used web browser, Netscape Navigator. A cookie is a small piece of data sent from a website and stored on a user's computer by the web browser while the user is browsing. Cookies were designed to be a reliable mechanism for websites to remember stateful information and to record the user's browsing activity. It provided the web an ability to recall, to remember its users. This was foundational, giving birth to persistent web browsing. This eventually allowed services online to deliver new experiences, ones that could build on the past activities of the user such as personal site preferences for instance, if you set the language preference on a site, the next time you visited, it would remember and display content in your chosen language. Beyond simple preferences, online retailers were among the earliest beneficiaries. Sites remembered what items a user placed in their shopping cart, even if they left without making a purchase.
As the web began to recall user details through cookies, a sense of continuity emerged. Sites could remember users, bridging the context between interactions across sessions. Businesses online capitalized on this newfound memory, targeting benefits and personalization for the user based on their past behaviors. Instead of casting a wide net with generic ads, businesses could hone in on specific interests, making advertising more efficient. The personalization wave began, with the web monetizing individual user interactions. This further created an incentive to strengthen and expand its memory.
The Web grows more connected and enhanced in its memory
While the cookies initially (first-party cookies) were tied to the website you were visiting, soon, third-party cookies emerged. These were set by entities other than the website you were visiting, often advertisers. This allowed for tracking of user behavior across multiple websites. In 1995 DoubleClick, one of the world’s first adtech firms, was founded. Their engineers began to tweak cookies to track users across the web.
As third-party cookies tracked users across different sites, they created a profile of user habits, preferences, and interests. This allowed for ads to be specifically targeted to individual users based on this accumulated data across various advertisers and publishers. The ability to provide more targeted ads meant that advertisers could get better returns on their ads, driving up the value of ad space on popular websites. This led to the Rise of Ad Exchanges and Real-time Bidding (RTB), websites could now auction an ad space in real-time, based on the profile of the visiting user. Within milliseconds, as a page loaded, an auction would decide which ad to show based on this profile.
This drove companies to start aggregating user data from various sources, creating more comprehensive user profiles. These could include online behaviors, but also offline data like purchasing habits, all tied to a digital identity. Giving rise to Data Brokers and Aggregation companies.
The Web at a Crossroad
As the web's memory became more sophisticated, it also became more intrusive. Users began questioning the web’s recall abilities, leading to a push for more transparency and control. While many users appreciated the convenience brought about by personalized experiences, others felt uneasy. The idea that websites 'remembered' them although novel, felt intrusive. This period was marked by debates on how detailed users should be remembered and how it should be used. Terms and Conditions checkboxes became more prominent, and companies started giving users the option to 'opt-out' of certain tracking. An event that underscored the inflection point in awareness and dilemma of user data practices was the The Cambridge Analytica scandal.
Due to privacy concerns and technological advances, there's been a move toward more decentralized, user-centric data models. Concepts like data ownership, privacy-preserving algorithms, and federated learning are becoming more prevalent. Infact it was for this very reason that my friends and I started building Eder Labs, Inc.
Envisioning a future for the web
The future web must be designed to deliver on some fundamental values. Users must have the power to influence and control the web's picture of who they are, creating a personalized yet more transparent and ethical version than the one we experience today. The advent of LLMs, federated learning, secure computation, and web3 protocols could deliver on the promise. A web where users own and control their digital representation online is the central pillar upon these technologies be moulded and introduced into our lives.
Challenges like monetization and potential misuse of technology loom, but the overarching theme is clear: a web that clearly remembers, understands, and enables its users to meaningfully expand their consciousness.