Interview with Mert Erkal: What is Google MUM, and how does it affect SEO

MUM is a multi-modal algorithm designed to answer complex queries by simultaneously evaluating the information on the multilingual text, images, video, and audio.

The algorithm is quite complex, so that will be implemented in several steps. Pandu Nayak presented this way of MUM development:

• Short-term development: Knowledge transfer between languages. The first example of using MUM was in the case of improving searches about vaccination. The new algorithm has identified more than 800 variants of Covid-19 vaccine names in 50 languages.
• Medium-term development: Multi-modal search with text, image, and video. It means an interface for simultaneous query processing with photos and text.
• Long-term development: Understanding complex questions and offering greater depth of relevant knowledge. Content engagement: Google spokespersons insist that MUM will not turn Google into a question-answering system.

SEOs may have to look at search data this year. You should notice the changes in search engine results quickly and take action.

We see that the concept of User Intent plays a decisive role in Google results. Google can change the page titles and descriptions that appear in the search results when necessary to achieve this.

Technical SEO will remain important, but with MUM, technical improvements, structured data, etc., implementations will lose their importance. Much more important will be the quality of content, the reputation and credibility of the site, and meeting user expectations.

In 2022, I think that the concept of producing quality content that best meets user expectations will be adopted instead of the idea of "SEO-friendly content." With the increasing importance of BERT and MUM algorithms, many SEO professionals will focus on the quality of the content and whether it meets the reader's expectations instead of producing SEO friendliness.

SEOs focus primarily on text content. But with MUM, Google makes SERPs significantly more diverse in terms of media formats, as it gets better at understanding and contextualizing video, images, audio, and text. For example, you can already see this when looking at the classification of images in image search or the automatic marking of timestamps in YouTube videos.

With MUM becoming more active, SEOs will have to do SEO work for different media types, not just text-based.

I think that with MUM, Google is starting to answer more search queries in Google results, this will increase, and sites may be at risk of getting less and less organic traffic.

It also means that only sites that create the best content and that can best meet their users' design and semantic expectations will survive.

BERT is Google's neural network-based natural language processing (NLP) pre-training technique. BERT stands for Bidirectional Encoder Representations from Transformers.

Google said BERT helps better understand the nuances and context of words in searches and better match those queries with more relevant results. It is also used for featured snippets, as described above.

Google provided this example to explain BERT: A search for "2019 brazil traveler to the USA need a visa," the word "to" and its relationship to the other words in the query are essential for understanding the meaning.

Before BERT, Google wouldn't understand the importance of this connection and would return results about U.S. citizens traveling to Brazil.

With BERT, Search became able to grasp this nuance and know that the widespread word "to" actually matters a lot here, and Google provided a much more relevant result for this query," Google explained.

MUM has the potential to transform how Google helps you with complex tasks. MUM uses the T5 text-to-text framework and is 1,000 times more powerful than BERT.

MUM not only understands language but also generates it. It's trained across 75 different languages and multiple tasks simultaneously, allowing it to develop a more comprehensive understanding of information and world knowledge than previous models. And MUM is multimodal, so it understands information across text and images and, in the future, can expand to more modalities like video and audio.

We see that AI-assisted content production is gaining momentum, and Google spokespersons have declared that such content is against Google's guidelines and will be considered SPAM. They recently updated this type of content in the Google Webmaster Guidelines; They started to state that they do not find AI-supported content that will benefit users against their guidelines, provided that it does not provide a ranking advantage. Over time, as the use of such tools increases, Google will probably change its stance on this issue.

With our contentevergreen.com brand, we are also in the content production business. We realized that we could use such tools much more efficiently than human writers. We think such tools can provide efficiency in producing quality content, provided that the finishing touch is done by a human.