most people i know use google by searching whatever question they have and including the word “reddit” at the end to find reddit threads since it currently has the most useful information.
As Lemmy gets more and more filled with useful threads and reviews it would be great if we can collectively improve Lemmy’s SEO so just including the word lemmy in a search will show lemmy threads related to the search.
The obscure tlds used in lemmy servers don’t help and lemmy.com currently redirects to lemm.ee. Is there a way we can improve the SEO of all instances or have lemmy.com be a aggregator of threads from many Lemmy servers?
I think once there is enough info on lemmy it’ll just get to the point where the searches will bring up the information you need.
I feel like if I’m searching on Google I want it to take me to the most relevant source of information even if that still is reddit. It won’t always be. But it is still right now.
Nothing wrong with competition, it’ll give better search results.
Lemmy and other federated systems being spread out across individual instances does make things more difficult. Normally you could just do “site:reddit.com” and automatically filter all results to be from Reddit. But you can’t do that because your result could be on any one of hundreds of instances, many of which do not have “lemmy” in their title.
It appear reasonably large instances like lemmy.one have been indexed, i got results using site:lemmy.one. and for those larger instances, they should still be able to index federated content.
Maybe I’m misinterpreting what you’re saying, but lemmy.one has basically no content on it other than the 8 communities that @jonah has created or allowed there. The whole point of that server is to allow people to simply login and then participate in other instances from there.
That is all to say, lemmy.one would be one of the “smaller” instances from a standpoint of content to be indexed by Google.
The whole point of that server is to allow people to simply login and then participate in other instances from there.
In order for users on lemmy.one to interact with content on other instances, lemmy.one has to import and host that content. So, it has plenty of content on it, just most of it originated elsewhere. That remote content should be just as indexable as local content.
I think it would be best if each post had a canonical tag pointed at the originating server’s version of that post. The lemmy ui generates a canonical tag now but I’m not sure it doesn’t just point to itself.
This sounds like a good idea for an issue on the Lemmy github.
Now that I’m on desktop, I checked and this was just added in 0.18.2: https://github.com/LemmyNet/lemmy-ui/issues/1418
The weird thing (which they’re also discussing elsewhere in this thread: https://upvote.au/comment/148846) is that it links to OP’s instance rather than the community’s instance.
I didn’t quite realize that, figured that when you were viewing another instances’ content it was loading from that instance. I guess that means that Lemmy content across all instances has loads of redundant copies.
Yup. It’s mirrored content all the way down.
Noob question here. Does that mean each instance must hold the totality of the content it knows about?
Yes it does. But only content created after subscribing to that community.
I’m not sure how lemmy or kbin handle instance-hosted media links – whether they import the media and redirect the link, or whether they point to the original media object – but otherwise, yes.
There are ways to access other websites directly from within a given website – iframes and the like – but that’s not what happens here. Each website is independent of each other, and all text is locally hosted in your instance’s database.
There are also (limited) copies of user profiles all over the place – if you click on my username, for instance, you’ll be taken to lemmy.world/u/Kichae@kbin.social. That’s a local lemmy.world user address, even though I’m not on lemmy.world. I can’t login to that account – it’s either credentialless, or has randomized credentials – but it exists. And by going there, you get to see what lemmy.world knows about my activity across the fediverse. Without ever leaving lemmy.world.
Thanks a lot for the explanation!
Probably talking about lemmy.world
Before this is of value we need to get another 100 million users and exist for 20 years
Why does it have to be just like your sister?
Because we will never reach a billion users and exist for 200 years like yo momma.
That’s that og Reddit spirit I crave
No thanks. This adds absolutely no value to the conversation. I get the humor, but why would one want to turn Lemmy into reddit instead of going to reddit to sooth the “cravings” ?
Eh… It has its place (especially if dude is going to set me up like that), and can be ignored if it’s not your cup of tea.
Prurient banter has been a part of online forums since BBS message boards. It’s how most people I know communicate IRL and online: Bullshit, bullshit, nugget of wisdom, picture of cat.
Not everything needs to be THUPER THERIAL, but it can be (and has frequently been) overdone because a lot of people don’t realize that self-censorship is a hammer to be swung heavily.
Lmao well said
Gosh
It’ll happen if Lemmy gets big enough. I only worry about search engines getting tangled in the natural duplication of Lemmy posts.
Like, if a web crawler sees a Beehaw post, and then seees Lemmy.ml’s mirrored page of that same post, could it just show up as two different results? Could it work against the SEO in that it gets marked as “duplicate” or “spam” content in some way?
Like, if a web crawler sees a Beehaw post, and then seees Lemmy.ml’s mirrored page of that same post, could it just show up as two different results? Could it work against the SEO in that it gets marked as “duplicate” or “spam” content in some way?
The ideal solution is that the page has a canonical tag, telling search engines what the main URL for the content is: https://ahrefs.com/blog/canonical-tags/. I don’t know if Lemmy already does this, nor do I know how well canonical tags work cross-domain as I’ve only ever used them for content on the same domain.
The ideal solution is that the page has a canonical tag, telling search engines what the main URL for the content is: https://ahrefs.com/blog/canonical-tags/. I don’t know if Lemmy already does this […]
I checked and it does, this post’s canonical is:
<link data-inferno-helmet="true" rel="canonical" href="https://merv.news/post/26663">Weirdly it uses OP’s instance, in this case merv.news. Shouldn’t it be the instance where it was posted?
Canonical tags were added in 0.18.2.
I would think it’s because users only interact with their own instance. They would need to post it to their instance first before it can be forwarded to the appropriate community’s instance.
If/When Lemmy and other federated services grow to the point that’s an issue in major search engines, said search engines should be smart enough to group and/or suppress mirrored results.
You can see that sort of thing in Google now for major sites like Reddit and StackOverflow, though it’s more along the lines of “the same question in a different post”.
You can also, in the interim, just pick an instance and add,
site:lemm.worldor whatever instead of just “lemmy”.It might help it, as well. I believe in the Yandex source code leak they detail their algorithms SEO techniques. Might be a good lead
We just need years of community created content. EzPz
Unpopular opinion, but I don’t want that. I don’t want to start adding SEO stuff. If we have good content Google can figure out how to index it better themselves.
Just curious, why not? Everything on the Fediverse is already public, by nature of federation. I think making the information shared here more easily discoverable is always a good thing.
Not OP but I think the point is not to force it but just let it happen organically.
SEO really doesn’t work organically, though. That’s why there’s a whole industry devoted to it.
SEO is an industry devoted to undermining search engines’ ability to organically surface good content. Good content will still be surfaced on its own, just maybe not quite as quickly.
Good content will still be surfaced on its own
Will it, though? This all seems like untested theory, to be honest.
While SEO may have started as a means of manipulating search engines, search engines have grown to adopt to new SEO techniques and now use those techniques as part of their built-in ranking systems. Outside of content that goes truly “viral”, I think it’s pretty difficult to get anything new to the top of a Google search without some massive SEO these days. Especially considering the head start that bigger players have already gotten on their SEO game, and the sheer wealth of content that search engines have to parse through.
I think maybe if we were still in 2010’s internet, that could be true. But search engines aren’t the same as they were in the past. SEO is the new norm.
What if we wanted to reject and change the new norm, just as we rejected Reddit and Twitter and started the Fediverse?
The silent majority are never going to have the time to blow on shit that people who get so mentally invested in this stuff do.
That’s what people don’t seem to understand. The silent majority is not mindless by choice. They’re mindless by design.
I used to do a lot of SEO and run AdWords campaigns for smaller businesses back in the early days and they were always the norm. If anything Google has been constantly tweaking its algorithm to make it harder for non-organic SEO.
Something as huge as Lemmy that grows organically doesn’t even have to worry about SEO. The problem is that the Fediverse being so spread out is a nightmare for Google spiders to crawl and rank.
The problem is that the Fediverse being so spread out is a nightmare for Google spiders to crawl and rank.
Isn’t that exactly why we should put in some basic SEO?
Good content will still be surfaced on its own
Not if you don’t do basic SEO at least… Things like ensuring that pages have:
- Good title tags, with the most important parts (post topic) at the start of the title
- Meta tags - description, keywords, canonical URL, etc
- Open Graph tags - for when links are shared on social media sites
- been optimized to load quickly - Google prioritises faster sites above slower ones
Plus the site should have a robots.txt and sitemap XML that’s been submitted to Google Webmaster Tools.
there’s a lot of “exclusivity” behaviour on this community that I’ve noticed (and a loooot of us vs the mentality too, in regards to twitter/reddit/bluesky/mastodon)
Yeah, I think a lot of the new Reddit refugees are failing to realize that Lemmy and the Fediverse existed before the recent migration and expect everything to just work the way they want, instead of how it’s been working just fine without them for years. The Fediverse doesn’t “belong” to them, and open connections are what this platform is built on.
When I read the posy what I heard was: how do we ensure high quality content? If we have that, SE won’t need the O.
And I’m down for high quality.
If and when Lemmy proves it has staying power then search engines will adapt. It’s only been a month or so since it blew up.
SEO takes loads of time. And since every instance is it’s own thing it takes more times to propogate this. I can only see a future where the most popular instances and communities are searchable in search engines. But I hope I’m wrong.
You normaly say that you have to wait about six months until google sees you.
I think to some extent this is starting to work. I googled a software question the other day and found a lemmy answer within the first couple of results! Definitely better than a couple months ago where even searching for “lemmy” didn’t bring up lemmy among the top few.
Keep in mind how Google works. It’s taking your metrics into consideration. So someone who has never Googled lemmy likely wouldn’t have seen thst same result.
That’s a great point, I don’t know much at all about SEO, but I do know how much Google tailors everything to your profile.
I’d rather find another solution. Imagine a fediverse search? Fuck google and their shitty SEO. It makes everything suck.
Eventually.
The future is here: https://www.search-lemmy.com/
including the world “Reddit” at the end
I found that adding
"join Lemmy"works for Lemmy.Also https://fedi-search.com has a few more ideas
There is also https://www.search-lemmy.com/ which seems to work pretty well. It’s also open source
Part of the problem is that there aren’t really a lot of questions/answers yet. You should try asking your questions here and hopefully over time we’ll have a big backlog of content.
I don’t think iw oild be unreasonable for someone to make a search engine that is specifically for indexing lemmy posts. Seems like it would be a good addition to the lemmyverse site
It already exists https://search-lemmy.com
But you have to choose the instance you’re searching, that’s the opposite of what I want, I want to search all instances, or at least the top 100 or something, all at once
It does search all instances, the instance choosing is so links open in your preferred instance. Try it, you’ll see
Sadly, it doesn’t :(
That difference is probably because it respects which instances your chosen instance federates with. (It has to, to open it there)
I imagine that’s easy to change for people who really want to search everything
I agree
Oh man, it’s like when the search engines were battling back in the day. We need a Lemmy version of MetaCrawler.
Dogpile!
While being decentralized certainly creates a barrier, most of the details behind PageRank (and the other algorithms in use by Google) are pretty well documented. If it doesn’t already, throwing in Lemmy as a keyword should soon bring up a Lemmy intense (probably Lemmy.ml or Lemmy.World) as a top result. As people click those links, the results will go higher.
The bigger challenge is that the content you are trying to find isn’t here yet. Those results on the old site were built over years of massive user engagement. Lemmy has barely had a month since people started joining en masse, and it’s still a fraction of what we lost.
TL;DR: Just keep using it and spread the word. The rest will happen naturally
This is going to present a formidable challenge because Lemmy is so decentralized. I mean I am all for it but wouldn’t even begin to conceptualize how to get started.
Not really… Nearly every phpBB, SMF, Invision forum etc is a separate self hosted community and they get indexed. Hell even small ones like one of my gaming guilds ranked high for specific key words.
For SEO to work well you need clear links to clear topics, and it helps to have automatic metada also to assist indexing.
Having google compatible site maps assists in crawling the site.
They can all get crawled, but you can’t just add lemmy and force results from the fediverse to the top.
You can still do SEO (search engine optimization) but making sure all the prefered metadata is there, sitemaps etc… It improves your rank.
Isn’t going to put you at the top, that is the job of the content but it can be the difference between being on page one or not.
