It was only a matter of time until Google released an updated piece of algorithm tech. Since the latest adaptation, BERT, launched in 2019, it seems the artificial intelligence team has been working on something new: Google MUM.
What is Google MUM, and how will it impact the world of search engine optimization?
What We’ll Cover:
- What is Google MUM?
- Why it’s Important for Users
- How MUM Could Impact SEO
- An Example of How MUM Could Work
- MUM vs. BERT
- How Google Plans to Roll Out MUM
- FAQs About Google MUM
What Is Google MUM?
In May, Google threw down with their annual I/O conference and divulged programs like new schema markup and even a new piece of search engine technology.
Google MUM, or Multitask Unified Model, is the latest algorithmic successor for the search engine giant. Google has called MUM “a new AI milestone inside of Google search.”
The tech is still in beta mode, but it shows a lot of potential.
Here’s a rundown of what is unique about MUM:
MUM is a new way for Google to understand information. By developing a more comprehensive understanding of data throughout the web, Google can provide answers to more nuanced questions. It basically gathers subcategories for the query and delivers a more holistic picture for the benefit of the end user. MUM is particularly attuned to comparisons for an umbrella of related queries.
One thing that’s interesting about MUM is that it understands things across text and images. In the future, Google expects to be able to incorporate audio and video in the omnichannel mix, too.
That’s not even the end of it. MUM can also find information that’s written in another language and bring that info together with a different language. For example, it can find information written in Japanese and translate it to English to combine it with other information. Google MUM’s ability to pull information from different languages gives you the best results possible.
Why Google MUM Matters for Users
The quantitative results of MUM have yet to be understood. However, Google believes the technology could cut down on the number of searches required to gather the same amount of information. This could make searching more efficient and help people understand thoughtful subjects more holistically.
It will probably take time for users to understand this new technology once it rolls out beyond beta mode. Right now, we know how to search given the current norms, but those norms are changing.
Google’s algorithm update combines “entities, sentiments and intent” all for the sake of the user experience. Do Google developers ever sleep?
How Google MUM Could Impact SEO (& Why That Matters for Marketers)
Whatever the impact on search engine optimization, MUM will definitely change how SEO works. It’s too early to say what changes digital marketers will have to make to continue providing the most useful information for users.
Some people wonder if SEO is even relevant in a MUM-driven environment. Apparently, it is—and will always be. Google’s Senior Webmaster Trends Analyst John Mueller says, “I don’t really see how this would reduce the need for SEO. Things always evolve.”
One thing is for certain, though. The content of the future will shift and SEOs will have to respond accordingly. It will be amazing to see what creative solutions marketing experts of the world come up with—and see how content fuses together in a MUM-driven SERP.
An Example of How MUM Could Work In The Wild
We know that Google considers MUM to be a new AI milestone for understanding information, but what does that actually look like?
Let’s look at the example they provided for their new search engine technology:
If you’ve already hiked one mountain (say, Mount Adams) and you’d like to hike another mountain next (say, Mount Fuji), you may be looking for some comparison information so you know what you should do differently to prepare for your upcoming trip.
Today (without MUM), Google can do this, but it’s tedious. The user must search different questions separately while being thoughtful to consider all the searches necessary to come up with the full picture. This might mean searching for the elevation for each mountain, average temperatures in a particular season, the number of trails and the difficulty of those trails, and even the proper hiking gear to bring for adequate preparation. This would take a lot of searching, but it’s a common type of query.
With MUM in place, users could theoretically ask a single question: What should I do differently to prepare? It’s like they’re talking to a trail expert with international knowledge. Users can include the keywords Mount Adams and Mount Fuji, and Google will do the legwork of gathering information to form a well-thought-out answer.
Basically, MUM understands the complex fact that you’re comparing two mountains. That means details like elevation and trail info are likely relevant. But through this new technology, the search engine could also understand that, since you’re asking about hiking, you may want to know how to prepare through fitness training and gear packing.
This example shows how there are a lot of pieces to the puzzle, and Google wants to cater to the complex environment. Today’s searches haven’t been really able to answer these types of questions in one fell swoop, but google says that could change with MUM.
MUM vs. BERT: What’s the Difference?
BERT and MUM are both built on something called a Transformer Architecture. However, MUM has more multitasking capabilities than BERT. Because of this, Google reports that MUM is 1,000 times stronger than BERT at providing nuanced results.
MUM is able to function in 75 different languages as well as a whole range of tasks, all at the same time. Its ability to comprehend multiple forms of media at once and tie them in with one another really sets MUM apart from BERT.
Don’t get me wrong…BERT is strong, and it’s been wonderful to watch SEO evolve since its 2019 launch. To think that MUM could be even stronger is somethin’ else.
Google’s Roll-Out Plan for MUM
Google aficionados admit that MUM is a major leap forward.
While all this is great, and this is an awesome leap forward with AI, they want to do it responsibly. Google says just as they’ve carefully tested the many applications of BERT which launched in 2019, MUM will undergo the same process as we apply these models in search.
They’re specifically looking for patterns that may indicate bias in machine learning to avoid biases in the system.
Right now, Google is only pilot testing MUM, it’s not fully live everywhere yet, but Google SEO will probably be impacted by MUM and probably in a good way.
FAQ About Google MUM
1. When will Google MUM expand from beta mode?
Google’s been mum on when MUM will expand from beta mode. It didn’t take an excessive amount of time for BERT, so the outlook seems promising.
2. What should SEOs do to prepare for MUM?
Continue optimizing your content with multimedia in mind. Keep the user at the forefront of your strategy, since that’s exactly what Google MUM is doing.
3. How does MUM differ from BERT?
While both MUM and BERT are built on the Transformer Architecture, MUM takes things up a notch. It can combine information from 75 different languages and across media types.
The Bottom Line on Google MUM
MUM isn’t a fully-fledged solution to the average searcher’s problems, but it is a sprawling leap forward in machine learning. Incorporating this new level of AI could change the SERP for good.
It could reduce the number of searches a user needs to make for thoughtful, multifaceted queries. And it could change how marketers address SEO to cater to new searching trends. To answer “What is Google MUM?”…it’s a lot of things. But above all, it’s a peek into the future for Google and the larger digital ecosystem.
Want more information on Google’s latest updates? Start here: