Hacker News new | ask | show | jobs
by rodonn 983 days ago
I'm pretty sure the author is misinterpreting what she saw.

Here is a thing that Google does do: The user searches for "children's clothing". An advertiser has created a search ad for the keyword "NIKOLAI-brand kidswear". If the advertiser has enabled "broad match" for their ad, then their ad will be eligible to be shown on the user's page in one of the advertising slots, even though there are no words in common between the user's query and the advertiser's keyword.

Does anyone have evidence of Google manipulating the user's query to affect the organic search results shown to the user?

4 comments

its usually from user behaviour and how the advertiser setup the keywords for the product. If users searching for "children's clothing" start having a high conversion for "NIKOLAI-brand kidswear" then more and more of their ads will start appearing when searching for children's clothing.

Google (and also the advertiser) have list of keywords that work well for products and then keep testing new keywords

This actually makes sense to me - conversion is possibly the most clear signal that the query intent was answered. While there is a financial incentive for Google to make this happen, it also results in improved user experience - in the less cynical days I might call this a “win win”.
Query re-write is probably one of the top keys to search quality, at any search engine.
Is it, though? I reckon it’s commonly accepted. I reckon there are cases that can be highlighted. As I’m sure it’s possible to make the case that some trickle-down policies work. But in the end, plural forms or synonyms do not equate to the original word.
You can just surround a word with double quotes if you need exact matches. It's nice having synonyms otherwise.
Haven't worked reliably in Google since somewhere around 2012 AFAIK.

Edit: Downvote all you want but this is documented.

I use it all the time. My biggest issue is I frequently get 0 search results with it on specific searches. It may not always have a result, but I don't get returned results that don't include the quoted portion.
As I write above it doesn't work reliably.

And as I think everyone on HN knows "my Google" isn't "your Google".

Even if it works reliably for you doesn't mean it does for me and many others.

Looks like her claim was in fact false: https://news.ycombinator.com/item?id=37802116
Searching for "children's clothing" gives 100% spam (== store front landing blurbs or things that are labeled "sponsored"). Searching "NIKOLAI-brand kidswear" gives some organic results and some spam.

So, it's now worse than the previous alleged behavior.

Great.

Exactly. “Children’s clothing” is a completely useless search until the search engine joins it with a million signals you failed to type. I can’t even begin to imagine how people believe every page in the internet could possibly be ranked in a useful way without those extra implicit terms.
I find it interesting how you manage to draw such weird conclusions. "Children's clothing" is not a useless search query. It is very generic or wide and THEREFORE useful.

Running that query through a GPT to transform it into "Popular brand among this bucket of users that spends a lot on ads" + "storefront" is completely useless, because now you have "enhanced" the query itself to a narrow subset of indices. All the fashion, sustainability, diy blogs are immediately filtered out in the query itself. The "frontend" now does not even have a chance at drawing from different buckets to form SERPs - every result is storefront now.

This is the issue discussed in OP. There is one implicit question here. Not only how much we need to tweak input queries to get the search engine to actually cast a wide net and allow discoverability, but whether that is possible at all.

If it's a useless search, follow the GIGO principle. The user will realise they made a garbage query and will make the keywords more specific.