What is Google’s MUM Update?

Google’s MUM update is a huge technological leap forward. But what tangible changes will we see, and what will it mean for marketers?

Here you’ll find:

• An explanation of this new Google technology
• The changes you can expect to see
• How marketers will be affected by MUM
• How to keep your site compatible

These days, when you type an inquiry into Google, chances are high that it’ll understand what you’re looking for and serve up the proper results. But as far as the search engine has come, it’s still not perfect.

That means queries still sometimes get misinterpreted, leading to unrelated results on the search engine results page, or SERP. Maybe you wanted facts about the song “Hotel California” but got results for the best hotels to stay at in California instead.

Thanks to Google’s new MUM update, these instances of the search engine not understanding search intent may continue to become even more rare. However, that’s not the only thing that will change, and marketers should know what to expect. 

Let’s go over this pending update, including what it is, what it does, and how your marketing efforts like search engine optimization (SEO) may be affected.  

What is the Google MUM update?

Google has been working on a new update to implement an artificial intelligence-based machine learning model that they call “Multitask Unified Model,” or MUM.

It’s such a huge advancement in search, AI, and machine learning technologies that it’s reportedly one thousand times more powerful than their previous BERT update released just two years prior. 

The goal is to make it easier for users to find answers to complex questions that can’t be resolved with Google’s standard couple-sentence summary snippet.

MUM will allow Google to process your query, identify the full search intent, and source answers to each step of the problem. (Image via Pexels)

How will MUM affect Google Search?

According to the search engine itself, users currently need an average of eight search queries to accurately address complex, multi-step problems. MUM is meant to tackle them faster, easier, and more naturally. 

Responses will more closely resemble the answer a seasoned expert would give that addresses the intent of the question rather than just the individual words used. It does this by implementing a few key changes.

Fewer questions, better answers

Google’s new MUM model uses advanced language processing technology called a T5 text-to-text framework along with other state-of-the-art machine learning methods to better comprehend language. This allows it to understand complex questions more fully and provide nuanced, multi-step answers.  

They use the following scenario as an example:  

“You’ve hiked Mt. Adams. Now you want to hike Mt. Fuji next fall, and you want to know what to do differently to prepare.”

While this would typically require multiple searches for each difference (e.g., weather, elevation, difficulty, gear, etc.) MUM will allow Google to process your query, identify the full search intent, and source answers to each step of the problem. 

This way, they’ll be able to provide a comprehensive and useful response that includes all the info you’ll need.

A global source of information

It stands to reason that the most accurate and specific information about Mt. Fuji would likely be written in Japanese. However, Google traditionally only provides results written in the language used to perform the search. 

As you can imagine, this severely limits the results.  

MUM, on the other hand, can understand 75 different languages, allowing it to respond to a query in one language using information written in another. This basically gives you access to 75 times as much information as before.

Thinking outside the text box

Another big MUM advancement: being able to understand information presented in different forms. This includes images, videos, and audio. 

If you ask a question that was recently addressed in a podcast or YouTube video, you’d usually be out of luck. With MUM’s new capabilities, it can better understand the audio and provide that information in its response.  

Answering questions instead of matching keywords

As marketers are well aware, search relies on keywords to identify the most relevant results to a query. 

But just because a result contains the same keywords, that doesn’t necessarily guarantee it’s relevant to the question asked (as the scenario at the beginning of this article illustrates).

A better understanding of language and meaning allows Google to identify answers that genuinely match the searcher’s intent, even if none of the keywords used are present in the results. 

It will present information with a meaning that answers the question, not just ones containing the same phrases. It could even present subtopics for deeper research.  

Google presenting information from sites in 75 different languages and translating them for users means we’ll all have more competition than ever before. (Image via Pexels)

What does the Google MUM update mean for marketers?

The answer to this question is complex. The nature of machine learning or AI technology means that it will, by definition, evolve over time. 

However, it’s possible that Google’s ability to aggregate information from multiple sources and provide a response will mean less traffic. If Google answers the query fully on the SERP, there’s less incentive to click, though users may still visit the page for context.

Many have speculated that this could eliminate the need for SEO altogether. But John Mueller, Google’s senior webmaster trends analyst, doesn’t see this as a realistic possibility. He believes that SEO will evolve, as it has for years, but it will always be needed. 

Marketers will likely adapt just fine, as we always have.  

Here’s what we know for sure: Google presenting information from sites in 75 different languages and translating them for users means we’ll all have more competition than ever before. On the plus side, it also means a wider potential audience.

Pro tip: The MUM update means you’ll be able to use non-text media like videos more often without a penalty. Once the changes take effect, you can slowly start implementing more types of media that have been traditionally barred from SEO benefits.  

Keeping your site compatible

There’s good news: You don’t need to do anything different right now to account for MUM.

As long as you create content with the intent of providing users the best information and most accurate answers possible, your materials are already optimized for the way MUM works. For now, especially since the rollout will be lengthy and could take years, nothing will change.   

The takeaway

This is yet another new and exciting shift in the world of search. It can also be scary, as the unknown often is. 

But you should take comfort in a few things. First, if you were optimizing correctly for the old system, your work is already done. Also, your reach and the kinds of content that can be useful for attracting organic traffic are about to grow significantly.

With an AI that understands your content as well as a human and without a need for keywords, there are fewer tricks others can use to gain favor. That means we’re all on a level playing field. 

As always, it’s best practice to create the most high-quality content you can. Beyond that, all our fates are in the hands of a higher power now — a thousand times higher, to be exact.

This is a complex topic. If you’d like help making sure your site is as optimized as possible before the changes take effect, let’s talk!