What are BERT and MUM?

What is BERT? (Bidirectional Encoder Representations from Transformers)

BERT was launched in 2019 to better understand context in search queries, especially in natural, conversational language.

BERT SEO

Focuses on understanding intent, not just keywords.

Handles long-tail queries and prepositions better ("for," "to," "from").

Pages that directly answer user questions in natural language rank better. 

SEO Tips for BERT

Write like a human talks. Avoid keyword stuffing.

Use semantic search optimization by addressing topics comprehensively, not just with exact keywords.

Create FAQ sections to address long-form queries naturally.

Improve content clarity. If your answer is vague or off-topic, BERT will deprioritize it.

What is MUM? (Multitask Unified Model)

Introduced in 2021 to understand complex queries, across languages and media types, with a deeper grasp of intent and context.

MUM can analyze text, images, and video, pulling information across multiple languages to understand nuanced, multi-part questions.

MUM SEO Implications
Google is moving toward multimodal understanding, not just text.

MUM can synthesize info from multiple sources to give a rich answer.

SEO Tips for MUM

Use structured data to help Google contextualize your content.

Include visuals, videos, and audio when relevant, especially if you want to rank in featured snippets and rich results.

Create in-depth content that answers not just a question, but also follow-up questions.

Cover topics holistically, like a guide or pillar page would. 

How to Use BERT and MUM for SEO

To match user intent, write naturally and conversationally

Improve topical authority by covering topics in depth, including subtopics.

Be multimodal by using images, video, audio when it adds value.

Help Google understand your content by using structured data and internal linking.

Think beyond keywords by focusing on questions, context, and meaning.

To test the structured data of your site, visit Schema.org