Since inception, the web has seen three major waves in the evolution of relevance: portal, search, and social. These waves manifest typically by the examples of Yahoo, Google, and Facebook/Twitter, and otherwise known as Web 1.0, Web 2.0 and the real-time/social/next/3.0/whatever web.
Jeff Weiner, CEO of LinkedIn, did a great job of framing that history in his talk at TechCrunch Disrupt in September. I’d like to borrow his framework but take a slightly different approach to the question. I’m probably short-changing this topic, so feel free to continue the conversation in the comments.
The goal of any system of content distribution is to present relevance as early as possible in the process of intent, search, discovery, and consumption. The earlier in this process that a system can present relevance, the greater the opportunity to provide value to the user, and, in theory, the greater the opportunity to monetize that value. The history of relevance on the web is therefore the history of a long steady march towards the holy grail of discovery – consumption without intent: content that you don’t even know that you want.
Each system described below uses the same essential building blocks to produce relevance:
- Human intelligence
Web 1.0 – The Portals – In the early web, the mechanism for discovering content was based on static human-built information hierarchies (Puppies is a category in Dogs which is a category in Pets). The opportunity to present content that you don’t even know that you want was confined to the moment after discovery, and so the value was minimal. Think: recommendations and advertising alongside content in the Teen Portal in AOL – you would have already reached your destination by the time you were presented with recommendations. Relevance was determined by assumed predispositions based on the demographic that had in the past consumed that content (in this case, teenagers). Intent as a predictor was not truly captured.
Web 2.0 – Search – The search paradigm introduced a major step forward for relevance, where the mechanism for discovery was based on dynamic and implicitly social intent-optimization. Whereas in the age of portals, the best you could hope for were recommendations alongside the content you were already in process of consuming, by basing relevance on user intent and an informed human curation via PageRank, Google was able to present relevance prior to the process of consumption.
PageRank, Google’s system for ranking relevance depending on inbound links to a particular piece of content from other publishers, is fundamentally human (it takes people to link to other pages/people), and its emergence foreshadowed the coming influence of the social web on discovery.
The Social Web – Facebook/Twitter – In the emergent social web, discovery has taken on passive characteristics, resulting in a model of consumption that occurs without search, and often without intent.
Consider the patterns of content discovery and consumption prevalent on Facebook or Twitter. Discovery is based on the stream paradigm, where relevance is determined implicitly along three dimensions:
- Time (real-time = conversational)
- Social group (work or friends or school, etc.)
- Social proximity (1st, 2nd, 3rd degrees)
Networks will expand along these dimensions, as users tend to follow or focus more of their attention to those streams that present the greatest relevance (think Twitter and Facebook lists). The social web enables users to iterate through a group of curators who provide relevant content by way of social proximity and temporality. It has replaced intent with context, and so while wading through the stream, we are left with a feeling of serendipitous discovery, as we stumble blindly into content that we don’t even know that we want.