> <i>Many site owners find that the titles they carefully craft almost all get rewritten.</i><p>Yeah, I'm with Google on this one. I don't see many reasons why a site owner would spend extraordinary amounts of time to "carefully craft" page titles other than SEO and optimizing for clickbaitness. As a user, I'm fine with Google counteracting this.
Google has a guide on this :)<p><a href="https://developers.google.com/search/docs/advanced/appearance/title-link" rel="nofollow">https://developers.google.com/search/docs/advanced/appearanc...</a>
> Takeaway: to dramatically decrease the chance of Google rewriting your title, matching the H1 to the title tag seems to be an effective strategy.<p>Of course it should be mentioned this wont last if it becomes popular. Historically every time an SEO trick gets popular, the rules are adjusted. Even having this article on the front page of HN might be enough to see Google react by rethinking how (or whether) tags in titles affect the title rewrites.
One think I don't see getting discussed in the pros and cons is the simple fact that you can't even tell what titles have been rewritten. Google gives no information in the search results to tell you what is original and what they've rewritten. This matches other trends like how it's become ever harder to discern sponsored ads from organic search results.<p>I used to love Google for how it presented relevant results and made it easy to discern sponsored ads. Today, I avoid Google products like the plague. (I can't escape all of them, but I'm about 90% off.)
I had this problem recently, was hoping there was a reasonable fix but it appears not... (the H1 already contains the title)<p>I don't think about SEO, and just focus on useful writing / societal impact. However, I recently discovered by accident that I ended up with a top 2 search result for "platform democracy": <a href="https://google.com/search?q=platform+democracy" rel="nofollow">https://google.com/search?q=platform+democracy</a> .<p>But the title is missing the first 3 words—including the key words "Platform Democracy" — so that if I was a random person aiming to learn about the concept, I would likely skip over the result! (I almost did even though I wrote the piece!) This seems not ideal for either users or Google, and also an interesting exploration of AI/NLP impacts, so I tried to dig a bit deeper.<p>I had a brief exchange with Danny Sullivan, Google's public
@searchliaison on it on Twitter (<a href="https://twitter.com/metaviv/status/1484636387366289413" rel="nofollow">https://twitter.com/metaviv/status/1484636387366289413</a>) which linked to two guides from Google on this. Sadly neither were particularly helpful, but will share them here in case they are helpful to others:<p>- <a href="https://developers.google.com/search/docs/advanced/appearance/title-link" rel="nofollow">https://developers.google.com/search/docs/advanced/appearanc...</a><p>- <a href="https://developers.google.com/search/blog/2021/09/more-info-about-titles" rel="nofollow">https://developers.google.com/search/blog/2021/09/more-info-...</a><p>(Also plausibly relevant: I have <a href="http://platformdemocracy.com/" rel="nofollow">http://platformdemocracy.com/</a> redirect to the piece. I imagine this might impact search ranking, but I would be surprised if it impacts the title rewriting.)
Zyppy's content marketing efforts aside, this wouldn't be so much of an issue if Google was any good at it<p>But as with its meta description rewrites, they're often worse than what was there to start with, and in some cases completely change the meaning, to the detriment of searcher experience
Thinly veiled content marketing for Zyppy, complete with CTA at the bottom, and mentions of themselves throughout, including: "Fortunately, here at Zyppy, we have a large database of titles thanks to our title tag analysis tool. Armed with this data, we set out to determine how often Google rewrites titles and the scenarios which trigger this behavior."<p>Furthermore, "HTLM" instead of "HTML"? Needs proofreading. Lol.
It's quite funny to me that people are running these kind of almost scientific experiments on a fully human-generated and in principle knowable system. The reason are understandable of course but it does seem like a waste of human energy.
Google has done this for practically as long as I can remember. If you remember when dmoz was still a thing, Google would favour the title from that, rather than the site's actual title because it perceived it as more useful to the user as it was moderated. By now I would expect that Google has used this and real moderators to train their machine learning model to rewrite titles, perhaps as a way to, you know, hopefully make the product more useful.
Speaking of rewriting titles... I noticed that HN reworte a self post title a few days ago. [0]<p>Why is HN editing self post titles?<p>[0] <a href="https://news.ycombinator.com/item?id=30053890" rel="nofollow">https://news.ycombinator.com/item?id=30053890</a>
The issue isn't that Google search results exclude stuff that was in the page title.<p>The issue is that Google search results insert stuff into the page title that wasn't there.<p>So the issue isn't that the overly long + pipe<p><pre><code> <title>Which programming language is fastest? | Computer Language Benchmarks Game</title>
</code></pre>
is abbreviated. The issue is that the domain of the hosting service is inserted, which gives the misleading impression that this is a project in-some-way approved and promoted by the Debian organisation:<p><pre><code> "Which programming language is fastest? - Debian"
</code></pre>
:when it would be better just to snip:<p><pre><code> "Which programming language is fastest?"</code></pre>
20+ years ago also used the meta description tag from a page instead of page text snippets. We're many decades past blindly accepting page author provided content as being the most useful thing to display. People keep thinking of Google as a search engine greps pages to find matching text. That is old/obsolete thinking, any Google-like services has evolved how directly it can return the information/answer you seek instead of returning a page that may contain the information/answer you seek.
What I find fascinating is I’ve seen a small, but increasing, subset of results where:<p>- the result title is clearly not original, usually derived from content on the page<p>- the original title is known to be generated<p>- the original generated title is as close to harmless as any web content could be<p>- the result title is actively harmful and misleading<p>- the original title is demonstrably better<p>- this idiosyncrasy is applied to very high trust hosts (eg GitHub)<p>- it’s <i>not</i> applied to <i>the same content</i> from obvious scraped content/spam/scam sites with obvious tells
Good. Google's interests more closely align with my own than page authors. I'm glad to have Google as an agent working for me to make more useful page titles.
Surely the misconception is the belief that what google displays is the page title. Google displays a link to a page, with a short description of what you will find there. Likewise, when I link to a page from my page, I don't use the title of the page: I use some text that I chose. This a non-story, as far as "rewriting titles" goes. What is interesting is that Google has an automated way to briefly summarize a page.
Every day my desire to be able to rate sites relevance after a search increases.
And I'd love to be able to choose if the original or Google generated title was the most relevant. (Cmon there is some machine learning training potential in that).<p>Rather that than ditching google search completely which is getting closer every day.
On modern browser, the page title is almost completely obscured. It’s not a thing users generally see, and the few views where users have access to the page titles that are not some sort of developer tool, the title is more often than not cut short.<p>I don’t see why google has to use the page title as a headline for a link result.
Something has to be fishy with this because I get tons of "Untitled" results now which directly lead to spam. This sucks big time because I usually got really good results since I search a lot coding related things and now I cannot use this account anymore for searching
<a href="https://web.archive.org/web/20220126145329/https://zyppy.com/blog/google-title-rewrite-study/" rel="nofollow">https://web.archive.org/web/20220126145329/https://zyppy.com...</a><p>Since people are reporting failure to load.
This data is based on what's seen in the wild, right? So if they see text in brackets removed more often than text in parentheses, that could reflect what sort of text people tend to put in brackets vs parentheses rather than (or in addition to) how google treats those characters.
I’m no search engine expert. Is this standard practice at some level across other search engines? Is “retitler” just part of every search engine stack (e.g. DDG, Bing, BraveSearch, etc)?<p>Or is this unique to the “I’m Feeling Lucky” folks?<p>Honestly curios.
Google literally turns the internet into a garbage dump. There are so many spam news sites that can come to the fore thanks to their seo nonsense that the sites that provide real news are not even seen lol
I hate those dropdown things that say "How to change gamma values in gimp" and they lead to a YouTube tutorial.<p>Please stop serving me YouTube tutorials; they all suck.
I totally get this. Back in the day when I was a kid, we went to the local library and read about the world. When the librarians weren’t serving me by “checking out books” to me, they were busily putting new and improved titles on the books in receiving.<p>/s<p>Seriously. Google is starting to feel less like the librarian of the net (we index the world) and more like the Truman show: we craft your reality.
Advertising corrupts. Ad-tech corrupts absolutely.<p>The reason is the cumulative impact of two things:<p>1. Algorithmic optimization of results for ad click-through rates, and<p>2. The scarcity of space for organic results on results pages for queries with commercial intent (because of the large amount of space given to ads). The high value of the clicks on those pages (sometimes $100+) drives marketers to focus disproportionate resources on SEO tricks and gaming to show in one of the few spaces left on the front page.<p>A search engine with ads and ad-tech tracking cannot work well for consumers in the long term. Google is now an ad-tech company, not a search company. It employs 3x as many people on advertising as search.<p>[Edit for clarity] It makes sense in this context to programmatically re-write titles to optimize conversion, rather than consumer experience.
Why was the title edited here on HN? The original title is much better, and had more information. A bit ironic given the subject matter.<p>Dang, was this your doing? If so, can we please have an open discussion on this? It's happened a few times and it's annoying and seemingly randomly enforced. The guidelines state not to editorialize headers but this rule gets ignored a lot. What was deficient about the original title?