The move by Google last year to focus more on topic relevance than exact match keywords forced me to re-engineer my query to content mapping. I used to be a little aggressive with individual keywords, but now I am groping for intent-based terms into themes. I then build content based on those themes, and I back that up with smaller pieces that give topical depth to clusters. This means that there will be less of standalone posts and more connected pages that point to expertise in an organized manner.
I do not miss observing the changes in Google algorithm since it directly affects how the business addresses its audience. Over the last one year, Google has given further emphasis on the user intent and the content quality and this has changed my key word research strategy. Nowadays, one is not satisfied with using general keywords or the number of searches. I have started giving much time towards the background of the search requests in a manner that the keywords that I would be focusing on would be based on the need of the audience to obtain answers to the issues. One of them was the case when I had a client in the field of e-commerce. We also did not use long-tail keywords that are too broad such as buying a pair of shoes but rather we used long-tail keywords that attracted a specific need of the customer such as the best running shoes when you have flat feet. This change of keywords to user friendly resulted not only in ranking them higher in terms of SEO but also in customer participation and conversion.
Generative search features changed everything when AI Overviews hit every US user in May 2024. According to Semrush data (https://www.semrush.com/blog/semrush-ai-overviews-study/) as of March 2025, the feature is now on 13% of queries and more prominent on informational content, how-to guides and question based searches. Zero-click risk has gone up big time and SEOs need to analyze SERP layouts, entity context and search intent gaps before trusting traditional volume metrics. Keyword research in 2025 starts with cartography not mining. Plot the SERP landscape first, see where AI Overviews, local packs and featured snippets dominate, then pivot to question based queries and mid-tail phrases these features miss. Build topic clusters that connect to each page with unique evidence and original insights that AI can't replicate.