Google MUM: Is Google Becoming a Purely Semantic Search Engine?

search engine
Search for the right audio, video, images, or written content on google.



Google MUM: Is Google Becoming a Purely Semantic Search Engine?

Innovations regarding Google search have been nothing short of cutting-edge and with the introduction of Multitask Unified Model (MUM), Google is one step closer to achieving the status of a semantic search engine. Similar to its predecessors, MUM represents an important change in how the Google algorithm understands user queries and offers solutions. So, what exactly is the Google MUM model and what does it mean for search engine optimization (SEO)?


What is a Semantic Search Engine?

A semantic search engine is one that attempts to generate relevant search engine result page (SERP) results by understanding searcher intent and the semantic meaning of words. By being able to understand the context, searcher intent, and the relationship between the words of a user’s query, a semantic search engine is able to provide accurate solutions no matter how a user asks a question. 


While there are currently no exclusively semantic search engines, Google has definitely been making advancements that can influence SEO efforts and content creation within the digital landscape in the coming years.

Google’s Pursuit

With the goal of becoming a semantic search engine, Google’s efforts can be traced back to 2010 when the purchase of Freebase, the semantic database, took place. Two years later, Google introduced the Knowledge Graph, a knowledge base that provides relevant information besides search results in the form of a Knowledge Panel or Knowledge Cards.

Google search on the computerYou can find relevant information on google


Then, in 2013, the Hummingbird update was introduced which was not only a rewrite of Google’s core algorithm but enabled the algorithm to become faster and more precise. With Hummingbird, Google no longer matched just keywords in a user query to those in webpages and was designed to better understand the context of conversational search queries. 


2014 was the year that both the Knowledge Vault and the Expertise, Authoritativeness, and Trustworthiness (E-A-T) principle were explained and implemented. The Knowledge Vault paved the way for the “long tail of knowledge” while E-A-T would rank entities.

Afterward, in 2015, RankBrain Machine Learning (ML) algorithm was introduced by Google to better process and understand search queries via vector space analysis. 


Unlike RankBrain, in 2018 Bidirectional Encoder Representations from Transformers or BERT was introduced to understand content as we do without analyzing past user queries by utilizing Natural Language Processing (NLP). This brings us to 2021 when the latest advancement to becoming a semantic search engine was unveiled: the Google MUM model.

What is MUM and How Does It Work?

MUM was introduced in May 2021 as Google’s most recent search technology. The Google MUM model is said to be about 1,000 times more powerful than its predecessor as it is able to generate language or text, is multilingual, and is multimodal or able to understand the information in various forms extensively.

Google MUM: SEO for Text

MUM uses various technologies in tandem with NLP in order to enable semantic search on Google. With the T5 text-to-text framework, MUM is currently trained to understand 75 different languages and carry out multiple tasks simultaneously to gain a deeper, more comprehensive understanding of information. SEO is heavily focused on text content and the Google MUM model considers this.

Google MUM: SEO for Media Formats

However, in order to become a semantic search engine, Google’s MUM will also be able to understand information across text, audio, and visual content. Though text currently reigns supreme in SEO content, this is sure to change as Google better analyses and deeply understands audio and video content for users.

SERP Features MUM Has to Offer

With certain SERP features that users are already accustomed to, the Google MUM model comes with a few new components that we can look forward to:

Things to Know

Similar to People Also Ask, MUM comes with Things to Know for users. 


While similar to the People Also Ask feature, the Things to Know panel is more in-depth and will generate relevant topics that a user is more likely to look at or search for next instead of waiting for a user to jump from one query to another.

“Zoom In” and “Zoom Out” Search

In order to provide an exceptional user experience on the search engine, Google will give users the ability to zoom in and zoom out of a topic in order to explore queries thoroughly. 


Searchers will be able to utilize features that refine or broaden their Google searches and either dive into a certain topic or backtrack and refer to a more general subject matter.

Upgraded Visual Search

While visual search is already available on Google, it is an upgraded experience. Searchers on the hunt for visual inspiration will not be disappointed after inputting a search query and getting a rich SERP with related images, articles, videos, and other content.

Updated Video Search

With its in-depth understanding of searcher intent and information, MUM will be able to process audio and video just like text. Even if a video does not explicitly mention certain information, the algorithm is able to identify the concept and content of a video and generate related topics for users.

How Will MUM Affect SEO

Gone are the days when content was keyword-stuffed in order to catch the algorithm’s eye. With MUM procuring and generating solutions to user queries more accurately and quickly, SEO specialists will surely see a shift in the content users want. 


Instead of focusing solely on keywords and phrases, SEO content will need to be more focused on the user and their needs in order to analyze searcher intent and provide the appropriate answers.

Google on the phoneGoogle can understand various types of content



While SEO content is currently text-heavy, MUM has made it possible for content creators to explore a new facet of the industry. 


Google is now able to deeply understand 


That being said, TF-IDF analysis will still be relevant and SEO best practices will still include leveraging long-tail keyword research, utilizing structured data, and creating an engaging website experience for users


However, in the years to come it will be beneficial to demonstrate expertise, authority, and trustworthiness as well as create content in various formats.

Google MUM: One Step Closer to Semantic Search

MUM is a major advancement in Google search as the search engine continues to prioritize consistent, up-to-date, and relevant answers to queries. 


Being user-centric, MUM does not only take a user’s current interest in a topic in mind but is also able to process future queries. 


Classic SEO efforts will not drastically change but quality content that showcases your expertise, authority, and trustworthiness will take on different formats in the years to come.

We are witnessing a shift in search behavior and the content that businesses and content creators provide in return.


Author:  Gurshinder Singh is a serial entrepreneur and holistic mixed reality specialist who has created record-breaking cross-media solutions for Fortune 500 clients such as Standard Chartered, Red Bull, Ford, IBM, SAP, Johnson & Johnson, and Cisco. He founded GTECH in 2008, a full-service digital agency based in Dubai. and he leads the Immersive Experiences division, offering consulting and development of VR, AR, and MR experiences.


Inveigle Magazine

Inveigle Magazine is a lifestyle, fashion, and beauty magazine. We love sharing informative articles with our readers.

Previous Post Next Post