google changed search again in 2024: what happened, what i noticed, what i did about it

seo · 2024-11-08 · 7 min

google changed search again in 2024. this is not a surprise; google changes search most years. what made 2024 different was that the change was visible to people who do not read seo blogs: ai overviews at the top, more ads before the first organic result, and a results page that increasingly looked like a conversation rather than a list.

i watched my own searches for a month and wrote down what changed. not the announcements. the searches.

what actually changed

ai overviews. for a lot of informational queries, google now answers first and links second. the answer is usually a summary of a few pages, with citations that are easy to miss. for "what is a variable in css" it is fine. for anything with a disputed answer it is a confident summary of a disagreement.

more ads. the first screen is often an ad, a shopping block, or a "people also ask" cluster before the first blue link. on mobile it is worse. the organic result is still there; it is just below the part of the page that used to be the answer.

forums and reddit. google leaned harder into "discussions and forums" for queries where people want experience rather than documentation. that is a real improvement for some searches and a disaster for others, because a reddit thread from 2019 is not always better than a product page.

site reputation abuse. google started enforcing against big sites that rented their authority to low-quality subdomains. a lot of affiliate content disappeared. a lot of it moved.

core updates. the usual. some sites won, some lost, and the people who lost wrote the usual posts about how google is broken.

what i noticed in my own log

i kept a note for thirty days: every search where the result page felt worse than it would have a year earlier, and why.

the pattern was not "google is bad". it was "google is optimizing for a different job". for a quick fact, the answer box is faster than a list. for a purchase, the shopping block is faster than a review. for a technical problem, the forum result is often better than the vendor's documentation. the cases where it felt worse were the ones in the middle: a specific question with a small number of good answers, where the answer box was wrong and the good page was below three ads.

the second pattern was that the good page was often still there. it had just moved down. that matters, because it changes the response. the response is not "write more content". the response is "be the page that deserves the click when someone scrolls past the answer box".

what i did about it

i wrote answer-first pages. the first paragraph answers the question in the title. not a story, not a preamble. if google is going to summarize someone, it should be able to summarize me correctly.

i used structure. one h1, clear h2s, short paragraphs, a faq where the question is real. not for the crawler, for the reader who is scanning. the crawler reads the same thing.

i added schema where it was honest. article, faq, breadcrumb. not a markup carpet. the schema describes the page that exists.

i linked internally. the cheapest ranking signal is a link from a page that already has authority to a page that answers the next question. i added those links by hand, because a related-posts widget is not the same thing.

i checked the pages that lost. search console, queries in positions 20 to 50. those are the pages where a refresh, a better title, or three internal links can move something. new posts are more fun. old posts are where the return is.

what i ignored

every "seo is dead" post. every "one weird trick" thread. every prediction that ai overviews would end the web. the web is still there, the clicks are still there, and the job is still the same: be the best answer to a question someone typed.

i also ignored the temptation to chase every update. google ships changes constantly. a site cannot respond to all of them. it can respond to the ones that show up in its own search console, which is a much smaller and more useful list.

where google still wins

local. maps. shopping. fresh news. "did you mean". the things where the index and the data are the product, not the ranking. for those, i still open google first, and i expect to for a while.

the lesson of 2024 is not that google got worse. it is that google got more opinionated about what a results page is for. the sites that adapted answered the question. the sites that complained wrote a post about it.

for the checklist i use before launching a page, see the seo checklist. for the search engine i use when i am not using google, see why i stopped using google.