What is Google MUM?
In May 2021, Google announced their newest search engine technology, with MUM being the word and replacing their previous software, BERT.
Multitask Unified Model, or MUM, has been developed to break down complex queries and deliver comprehensive search results. On average, it takes people up to eight searches to complete complicated tasks, and Google’s goal is for users to do fewer searches to find the information they want.
In this week’s blog post, we will be looking in more detail at how this new AI Search Engine technology works, as well as how it can benefit your website’s SEO.
How does MUM work?
Before 2019, traditional Google search queries were restricted by the limits of their language understanding capabilities, especially with more complex questions. Google analysed the language by highlighting keywords and results directly related to the keywords, regardless of whether they were relevant to your search.
MUM brings a whole new dimension to how it analyses the language in the query by using NLP (Natural language Processing), a new form of AI language analysis software that Google introduced into their search engine in 2019. It analyses the language on either side of keywords in searches, breaking down complex queries to deliver a more tailored result.
So, for example, you can type in ‘I want to plan a wine trail from Melbourne to Adelaide in August, and Google will now have the intelligence to show you the best wineries along your route, culling out any results that aren’t relevant. Google MUM may even further use its insights to understand the term ‘road trip and provide information recommending car hire services, petrol stations or the best time of year to visit these wineries on this specific route.
So, Google is now essentially able to do all of the research and put together an answer for you. Google may only have given you directions from Melbourne to Adelaide or even wineries in your local area with the previous search engine technology. But now, it can provide you with relevant information much closer to your actual search request.
How does it compare to BERT?
MUM is 1000 times more capable than their most recent Search Technology, BERT.
BERT was rolled out in October 2019 and stands for Bidirectional Encoder Representations from Transformers. BERT was a crucial development and a step towards developing MUM. This new technology helped our search engines understand the meaning of keywords from the context by analysing the words before and after them, rather than just focussing on individual keywords.
Words can have many different meanings depending on their context so that this new technology can give much more accurate search results because of that added information from the context. This is particularly useful for understanding the aim behind general search queries with not much detail.
But this is the limitation of BERT. Rather than processing multiple keywords and phrases simultaneously, Google’s algorithms picked out the ones that stood out the most. So although you can find search results that match your queries, they may not be entirely relevant as the search engine can’t look up information for more than one keyword at a time.
And this is where MUM takes it to the next level. MUM has been built on the back of BERT and so is similar by using Bidirectional training for search queries. The main difference between the two types of software is that MUM can multitask better and generate multi-layered content that answers complex questions, resulting in more intuitive search engine software.
In addition, MUM Is Multimodal, meaning that it can understand and analyse information from different media formats, including videos, images and webpages at the same time. So as the software develops in the future, you may eventually be able to take a photo of something and ask a question. For example, you could take a picture of your running trainers, upload the image and ask, ‘are these suitable to run a marathon?’. MUM would understand the question, connect it to your picture and provide an answer or recommendations on better shoe types or brands to use.
So, in contrast to BERT, MUM can read text, understand the meaning, form deep knowledge about the topic, use video and audio to enrich that, and get insight from over 75 languages to translate that information into an answer from your search queries. Pretty cool, right?
How will this affect my SEO?
For now, nothing much should change for content SEOs. SEO’s should continue as usual for the time being – creating high-quality content over a range of media formats.
The significant advantage Google MUM will provide will be increasing a website’s potential target group, as long as that website has high-quality content that Google MUM will include in its search and analysis. But Google has also stated that they will roll out this technology over the next few years, so more noticeable changes will be over the longer term, and we may need to wait a while to see how it impacts search behaviour (if it does at all).
How to make the most of Google MUM.
So, how can you tweak your content to make the most of Google MUM?
1. Geotag your images
Geotagging is the process of adding a precise global location to your images. By embedding this information in your website, it means that it is more likely to get picked up by Google’s crawlers when looking for data related to this location.
2. Ensure you use Alt Text appropriately
Also known as Alt Tags, alt text is the text displayed in place of an image on a webpage if the media fails to load on the user’s screen. It’s essential for SEO ranking as it increases the user experience on your web page and allows search engines to crawl better and rank your site’s content. As MUM will look up information from multiple sources simultaneously, including video and images, it will need to verify that it is relevant, meaning alt tag will be more important than ever.
3. Create content that is more relevant to your niche
Content is crucial when it comes to improving your SEO. Search engines won’t have enough information to index your pages correctly and rank your page without relevant and high-quality content. This is even more important with Google MUM, as when a user searches for that topic or term, and will only select relevant results.
4. Make full use of text, images, audio and video on your site
Because Google MUM is multimodal, it will be vital to utilise various media types on your website. Although this is something you should already be using as it gives variety to your content, adding additional media types will ensure that Google notices your page and includes it in search results through Google MUM.
For more information on how we can help you optimise your website for MUM search technology, contact Design Grid today!