The Interplay Between Browsing, Searching and Recommendation

December 30, 2021

When we look back a generation, people worked really hard to seek information; which consisted of spending hundreds of hours pouring over numerous books, journals or research material in the libraries.

After the invention of the Internet, this process became less tiring and has only become simpler and more intuitive as the years are passing by. Yet when we look back, the style of accessing information on the internet has changed drastically over the past 3 decades which namely are:

  • Browsing
  • Searching
  • Recommendation

We will be covering the first two items in the first part of this article.

1. Browse

Consider this- You are in a library and you want to pick a few books on History to read. So you approach the librarian and he directs you to the History section where there are numerous books to choose from. So you go through all of them to select something closest to what you were looking for. You may or may not find exactly what you were searching. This is approximately how the earliest browsing system worked.

The earliest forms of browsing happened via portal websites such as the now defunct America Online. They were a directory of different URLs and were similar to the earliest forms of newspapers. They had a large centralised collection of content curated by human editors and displayed and organised on the website by an editorial staff. The editors decided which content should be displayed on the top and so forth. Before the advent of the Search Engines, the only way to look for a particular information on the Internet back then was to know the exact URL (Uniform Resource Locator) of the page or the file. No user behaviour was recorded.


Around this time, the first web browser called Mosaic was developed as well. Inspired by Sir Tim Berner-Lee’s WorldWideWeb which worked solely on NeXT computers, Mosaic became the first easy-to-install web browser with graphic capabilities that could be installed on most computers. Essentially it became a window to access the wide web of the Internet. Users now were no longer dependent only on a portal site’s list of URLs; they had the freedom to access any website of their choice via a browser. Post Mosaic, other browsers like Netscape, Internet Explorer and Google Chrome became available for users to download and install easily, widening the accessibility to the Internet.


2. Search

Now you go to the same library with a specific query- to read the book Hitler by Ian Kershaw. The librarian, instead of showing you to the History section, searches through their book-keeping software to find where the book is located (provided they do have it), and then hands it to you.

This is how the search function works which formed the 2nd period of the Internet in the mid-1990s. The web became too big to be indexed manually as more and more people started to set up their own websites and blogs to publish online. The Search Engines solved the issue of locating information across a vast sea of web content.

The user has an idea of what they are looking for and types in the approximate keywords to seek that information. They now are not merely browsing through the vague content that’s readily available on the web.


The advent of Search Engines like Yahoo, Google and Bing saw a huge shift in how information/product/services were sought after. Searching became more efficient and accurate with only one major drawback- the user must know the term they wished to search for. This made search less user friendly for news and entertainment which hold a major discovery element.

3. Recommendation

This problem was solved by the recommendation algorithm which formed the 3rd wave of the internet which is the most recent. This change started with the invention of a new machine learning algorithm called Sibyl by Google in 2011. Sibyl makes predictions and recommendations on the basis of user behavior across Google’s applications namely YouTube and Gmail. Sibyl was a revolutionary change of how content started to get displayed for different users.

Users no longer had to search for information; instead the information came willingly to them. It removed the requirement to think about specific keywords to type onto the search bar or to subscribe to channels. Recommendation infers the user’s preferences based on their previous behaviour. In many instances where users didn’t have a clear cut view of exactly what they wanted, the recommendation algorithm gave them suggestions of what they might like which only got more and more accurate with the corresponding years. It also became more intuitive about predicting the user’s exact query as shown in the image below:


Google further optimised the recommendation system of Sibyl to form Google Brain which further leveraged new advances in deep learning. In the years 2014-2017, it was noted that the average time spent watching videos on YouTube had increased by twenty times due to the placement of ‘Recommended Videos’ on the homepage.


It’s like based on your earlier choice of picking the book Hitler by Ian Kershaw, the librarian now knows your preference and suggests you books like Broken Lives: How Ordinary Germans Experienced the 20th Century and Anatomy of a Genocide whose existance you weren’t even aware of.

The importance of the recommendation was especially noticed in e-commerce where prompts such as- ‘Customers who bought this item also bought…’ made users to purchase more from the website.

This major shift from ‘people looking for information’ to ‘information looking for people’ gave rise to new companies, one of which was TikTok.

TikTok took inspiration from YouTube’s video recommendation and Amazon’s product recommendation to start a platform that generates a continuous stream of short 15 sec completely customized to each viewer’s taste and liking.

For example, if a user A likes watching cat videos, then based on TikTok’s AI’s calculation of other similar users, it’ll recommend funny animal videos to user A as well. The internet is now beginning to flood with applications which instead of proactively pulling search information from users,it pushes information onto them based on similar queries made earlier. The AI is constantly collecting user information in order to make their recommendation more accurate.

4. What’s the future?

The Internet is now advancing rapidly to become more and more intuitive. In the future we might be surprised by the advanced technology where maybe we only have to ‘think’ about something for it to get displayed on our screens. The reality of that happening might be closer than we think it is.

Do follow our LinkedIn page for updates: [ Myraah IO on LinkedIn ]