this post was submitted on 06 Aug 2023
798 points (96.6% liked)
Fediverse
28520 readers
295 users here now
A community to talk about the Fediverse and all it's related services using ActivityPub (Mastodon, Lemmy, KBin, etc).
If you wanted to get help with moderating your own community then head over to !moderators@lemmy.world!
Rules
- Posts must be on topic.
- Be respectful of others.
- Cite the sources used for graphs and other statistics.
- Follow the general Lemmy.world rules.
Learn more at these websites: Join The Fediverse Wiki, Fediverse.info, Wikipedia Page, The Federation Info (Stats), FediDB (Stats), Sub Rehab (Reddit Migration), Search Lemmy
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
One problem I see:
You can google
site:reddit.com whatever
But if you googlesite:lemmy.world whatever
then you're losing a significant amount of results. To get good results, you need to know which Lemmy instances is likely to have your answer, and with communities duplicated over different servers, that can be tough.In the end I find I prefer this federation model, although I'm not sure although I'm a bit concerned about funding it if it scales up to the size of Reddit (same with Mastodon vs twitter).
Google should be finding searches with "lemmy" keyword, but it isn't at the moment.
Lemmy needs some SEO people.
I don't think lack of SEO is the issue. There's just not enough content and brand/domain authority to get results from here high in SERPS.
There might be something fediverse related that would affect performance in search, but I'm not knowledgeable enough about this setup to speak to it.
I think it's just lack of content, general awareness/interest, and longevity that's keeping Lemmy low in search
Lemmy contents are replicated by federated servers, so you might find what you want by using
site:lemmy.world
or other big instances because they might also has replicated contents from other smaller instances.This has more to do with how bad Google has gotten, such that you're forced to add restrictions like Reddit to get rid of SEO sites and get useful answers. A proper working search engine would show these (and any that are found in Lemmy) high up by default.
I'm sure the search problem will be solved somehow. Like all the content is on each instance so its just a case of it being accessible and indexed by google I guess?
I'm sure it's already being indexed by Google. But people like to add site filters like
site:Reddit.com
orsite:stackoverflow.com
to prevent google from barfing up a bunch of garbage results on the front page, when they know that's probably where the results they want will be. There is no way to add a Lemmy-wide filter to a Google search, because Lemmy instances are all different sitesDoes it actually matter though because Lemmy contents are replicated by federated servers, thus big Lemmy instances such as lemmy.world might have contents from smaller federated instances as well. Try using
site:lemmy.world
next time and see if it'll improve the search result, though Lemmy.world is just 2 months old so maybe Google hasn't indexed it allThat's a good point. If you filter by a major site, then it'll have content from all the major communities.
That won't help if you're looking for niche content, but that's not as important.
I wonder how replicated data shows up to the indexer. I don't know enough about search engine indexing or SEO. Will google index replicated data? Presumably it won't index feeds or searches, it'll index the actual posts, and I wonder if replicated posts are considered posts for the purposes of indexing or if the indexer will only look at local posts.
Google isn't thrilled with duplicate content. Following this thread here, it sounds like identical content might be hosted on multiple servers? If that is so, it's not going to be high value in Google's eyes.
If it's indexed, you'll be able to search it with Boolean modifiers, but it might not get priority in organic searches.
Yes, contents are replicated across federated instances. For example, here is the link to this thread on my instance: https://lemmy.institute/post/49173
If you check the html source there, there is a canonical link in the header that points to https://sh.itjust.works/post/2334723 , which is in the OP's instance. I think google will respect canonical links when indexing duplicated contents, so maybe the SEO aren't affected too much?
Presumably how it should work is that that even if content is duplicated, the crawlers would only index the "local" for Mastodon/Lemmy/etc servers, so they wouldn't see the duplication.
But idk how it actually works, and we're right back with my original concern of
site
filtersIdeally it would be popular enough that you wouldn't need the site modifier. Google would see that Lemmy has the most seen and perpetuated answer just like it sometimes does with Reddit now, whatever the instance.
People still often out the site modifier on just to prevent google from barfing up a bunch of crap they don't care about, even if they know that Reddit results will be near the top.
In the eyes of a search engine, yes.
But once a site is popular enough for traffic and engagement to influence it's position in search, it's def going to be popular enough for bots, trolls, bad faith actors, grifters, etc.
Welcome to the old Internet. Decentralization is good in a way, people will have to try harder instead of having everything spoon fed to them by Google.
I'm not personally a fan of that brand of elitist gatekeeping. Having it be harder to keep out the plebs is not a look I think we wanna get behind.
Decentralization is important, but the goal isn't to keep people out.
I guess I didn't exactly mean it as elitist gatekeeping, I see it more like people are being abandoned by major websites and this is the result.
People having to work harder is good? No I disagree with that entirely.
Part of what makes reddit so amazing is the amount of amazing knowledge and answers you can find from google.