As content marketers, we know that staying up to date with Google updates and algorithm changes is crucial to the success of our blogs. With updates like BERT making quality content more rewarding, we have faith that Google updates will continue to evolve to fit consumer needs.
We look forward to the MUM update for this reason. MUM, which stands for Multitask Unified Model, is one of these exciting updates, as this makes search queries easier, more understandable, and improves upon past keyword searches.
Read on to learn more about MUM, this new AI milestone, and how content marketers can adapt to it:
So what is MUM – in Non-technical terms?
While Google search is used by billions of users every day, and roughly 3.5 billion Google search queries are made every day, not many understand technically how this process works. The good news is that we don’t need to understand the science behind it, and we simply need a general understanding of it.
Google search queries work because they read the data of automated web programs like crawlers. Crawlers scour the web looking for new and updated web pages, and then Google stores this information in their databases. In the database, the data is indexed. This means that the pages have been analyzed and organized in computer language. This allows the Google search results to pull the data from the indexing to deliver a result. The search query will try to determine the highest quality results and present them in order.
This is the tricky part, and where content marketers and SEO experts come into play. There are a number of factors that determine what is “best,” and these also include location data points, language, the device that the user is using, and the user’s previous queries.
As you can see there is a lot that goes into this process, but Google always wants to improve. And that’s where MUM comes in. MUM is advanced AI research and it aims to improve upon these search queries.
With previous search queries, users with complex questions would have to conduct multiple searches to find their answers. Google Fellow and VP Pandu Nayak gives a really great example of climbing Mt. Adams. If you’ve hiked Mt. Adams and you want to hike Mt. Fuji, you are looking for different ways to prepare. You’ve hiked the mountain before, and you just want to know the changes that need to be made.
Without MUM, you would have to search for the height differences, differences in weather patterns, and gear changes, and go through multiple rounds of research.
But MUM aims to simplify this. Instead of issuing eight queries on a complex task, MUM would understand that you’re comparing the two mountain treks, and it would be smart enough to provide elevation and trail information. It could also go into things like training recommendations and gear.
Key Improvements With MUM
As you can see, MUM is taking the learning power of Google and BERT and amplifying it. Compared to BERT, MUM is 1,000 times more powerful. Across 75 different languages, it can read and understand different tasks at once and it can also generate language. MUM, in essence, has more reading comprehensive.
It’s as if BERT graduated from a 3rd grader to a 4th grader (but with the ability to understand 75 languages).
The ability to translate language is also key, especially in our hiking comparison. Before, a lot of the information about Mt. Fiji might have been in Japanese. But if you’re an English speaker, this information would not have pulled up, even with translation options. Now, MUM can search through different languages and offer new, more applicable results.
In addition to this, MUM might also be able to “read” images and provide results based on these images. This is key when it comes to gear, for example. Not only will it look for the text about gear, but it can also pull up that blog on recommended gear that you’ve been looking for.
Does MUM Change Any of the Search Quality Guidelines?
According to Google, MUM does not change any of the current search quality guidelines, but I gather this will change based on how the Human Raters perceive MUM’s effectiveness. MUM will still be looking at on-page and technical SEO, and continuing to source the best, most reliable web pages out there.
If you’ve put off any SEO best practices for another time, well now’s that time. Addressing critical SEO infractions now will set you up for success when MUM is fully active. And this timing also gives you a chance to optimize your blogs for updated keywords, adding images, and reformatting any headers.
Removing Barriers With Advanced AI
It’s clear that MUM has the advantage of removing significant language barriers and providing a deeper topic exploration for search queries. And just like with BERT, MUM will be tested and trained to improve its search results based on what we’re looking for.
What I love about Google is that its search improvements are always human-centric. This means that we don’t have to change really any of our barriers when we’re typing in a search query; the program is modified so that it makes search queries easier. All we have to do is go along for the ride.
For content marketers, a similar notion applies. We should see that more relevant blogs, even within niche topics, are being rewarded. Instead of having to optimize a blog post to a strongly competitive keyword, the topic may be rewarded on its own to the niche, long-tail keywords that we know and love.
It’s not exactly sure what MUM will be released in all its glory, but for clients, we recommend staying patient. For smaller firms who are targeting competitive keywords, it might useful to bring in more long-tail keywords. Additionally, those targeting only long-tail keywords could afford to bring in some high-powered targets as well to create a balance.
Other thoughts come up around images and other languages. It’s clear that blogs need more images these days, but these will only work when the alt text is included. And it might become more useful to start guest blogging or backlinking with blogs written in another language to boost authority.
