The Web is a vast expanse, and so it has a number of mechanisms to discover content, some designed and built in from the outset, and some of which have emerged throughout its history.
Users have three main methods to open a desired webpage — by typing the URL in the address bar; through bookmarks they have curated; or through a search engine, which forms a large majority of web access and is a core activity for web browsing.
URLs provide unique identifiers for webpages and as such should be well placed to get the user to the website they want. They were invented as part of HTTP, the core protocol that turns URLs and requests into webpages and responses, so they are critical to the Web. The URL has problems in that they’re not human-centric, can be long and hard to remember and there are serious issues with trust.
Enter bookmarks, where a user saves a collection of links to their computer or device for later use. It has a critical issue in that bookmarks must first be saved before they can be used — you have to know what you want in advance — so the value only comes on revisiting the site on multiple occasions after the first visit. With shared bookmarks and bookmark sharing services, users have shared links of interest in various verticals, but have not fundamentally indexed or organised the whole Web.
Search engines soon emerged and have turned out to be indispensable for web navigation, and yet with all the advances in Web technology, search engines haven’t developed at a similar pace. All major search engines take a query and return a sorted list of results, often running into the millions and billions of results, each with a title, description and a URL. This turns out to be noisy and information overload, especially so when the user types a simple query, and has to deal with such a heavy cognitive load of parsing results and navigating to their intended destination.
We want an answer to web discovery: to take the user’s query and produce one or a small set of results — to return what the user wants. In a more concrete example, a user wants to get to a particular website, like BBC News, quickly (within 1 to 3 seconds) and easily, without relying on URLs or bookmarks. There is no current solution available that has an answer for this problem.
I’ve been searching for the answer since 2013, being tired of collecting bookmarks for sites I visited and being ever more uneasy with verifying the identity of websites to ensure they were legitimate and trustworthy. In 2014 I had an aha moment and since then I’ve been working on concrete solutions for web discovery, focusing on the key tenets of speed and trust. Today we’re ready to talk more about how we can upend web discovery and change the dynamic of the Web for good.