Not a bug Redirect likes

rfc0001

Well-known member
Affected version
2.1.3
Since XF has implemented reactions, crawlers constantly spew 404s for "/likes" URLS, e.g.:

/forum/media/media-xyz.123/likes
/forum/posts/123/likes

Can you automatically redirect /likes to /reactions?
 
The likes pages always were and the reactions pages still are marked as noindex so I'm not really sure where or why you're seeing them anywhere.

As they are already noindex there is little benefit in redirecting them (and no harm in not). Any crawlers ignoring it, should eventually drop them off the index.
 
I think some crawlers are optimized to look at Xenforo likes for factoring in PageRank. Their code is apparently unaware of the emotions. I haven't narrowed down who is doing it, since the referrer information isn't shown based on how I'm looking at the 404s. It's one of the major/reputable crawlers since I block all unknown ones.
 
Just read the description on the add-on, looks like it is bingbot:


Fixes bingbot issue where Likes are being searched instead of Reactions. This add-on will redirect Likes to Reactions.
 
I think some crawlers are optimized to look at Xenforo likes for factoring in PageRank.
Or is it just trying to crawl links it previously knew about to check if they are still noindex? That seems more plausible. I guess they can't assume that an instruction not to index a page is permanent. Even so, I wouldn't recommend trying to redirect them (by any means). They'll drop off eventually.
 
The've been removed since 2.1, which I installed 6 months ago, and I just started redirecting them after installing 2.1.3, so I don't think it's "memory" of the indexed pages that they keep checking. I was assuming they were using the likes to help with page rank. I'm also assuming they have this hard coded for xenforo sites but are unaware of reactions, thus the redirect would allow the ranking to still work.
 
Last edited:
As mentioned here, I also agree that this has to be a bug.

All forums that used to run on the 2.0 version and upgraded to 2.1 have these errors. Sure, one need an addon to identify it, but it is based on the core, not on the addon.
 
The pages no longer exist, therefore they are now 404. It's that simple. If they are indexed (despite us stating they shouldn't be) they will at some point stop being crawled. That might take a few weeks after upgrading to a few months or even a few years.

If it matters to you for some reason (it won't affect any SEO signals) then there are measures to rectify it - we won't be.
 
If you say that it won't affect anything, then all is good. I just thought 404s are bad in general, but you are the expert, alright.
 
it won't affect any SEO signals
This might be a safe assumption, but it's an assumption nonetheless since search engines are black box, so really no way to tell if they are "smart" enough to detect and weigh likes into page rank. Google certainly weighs google analytics traffic into page rank so would make sense they would also weigh sentiment for social media, which forums essentially are. Of course, we are talking Bing not Google here, so clearly Google is not doing this, but Bing might be. I'll try to dig into this more.
 
Last edited:
I think some crawlers are optimized to look at Xenforo likes for factoring in PageRank. Their code is apparently unaware of the emotions. I haven't narrowed down who is doing it, since the referrer information isn't shown based on how I'm looking at the 404s. It's one of the major/reputable crawlers since I block all unknown ones.
Highly unlikely.
 
This might be a safe assumption, but it's an assumption nonetheless since search engines are black box, so really no way to tell if they are "smart" enough to detect and weigh likes into pagerank. Google certainly weighs google analytics traffic into pagerank so would make sense they would also weigh sentiment for social media, which forums essentially are. Of course, we are talking Bing not Google here, so clearly Google is not doing this, but Bing might be. I'll try to dig into this more.
It really isn't just an assumption. We are talking about 404 errors (Page Not Found ). These have never impacted Google rankings or any other search engine:


This is just Google informing you that they have found a 404 error for a link in case that link was supposed to lead to a page you actually care about. As @ChrisD said, these can safely be ignored and will eventually drop out,
 
That's not what I'm saying at all. I'm saying Bing is obviously hard coding crawling the /likes nodes on threads/posts, since they fell off many months ago but are still be crawled. This implies they are aware of the site being a Xenforo site, and hard coding the likes navigation. The reason they might do this is to include user sentiment into their page ranking (e.g. ranking more popular threads higher than less popular threads). Assuming they do this (which it appears they do, although their algorithm is black box so we really don't know what they do with that information), while the 404s may not hurt your pages rank on average, it certainly could make an otherwise more popular thread rank lower and less popular thread rank higher in their results. Also, we are talking about Bing, not Google. I just used Google as an example since they certainly use all the signals at their disposal including Google Ads/analytics signals. Bing is somewhat handicapped in this regard, since Microsoft has no Google Ads/Analytics equivalent, therefore cannot weight pages within a site based on relative traffic.
 
Highly unlikely.
Why else would they navigate the /likes node when it doesn't exist many many months after it's been removed from my site (and the site is crawled regularly)? There seems to be some hard coded logic that is aware of the likes on XF sites, which implies they are using these signals for page rank.
 
It really isn't just an assumption. We are talking about 404 errors (Page Not Found ). These have never impacted Google rankings or any other search
As always: It depends.

We've had cases where 404's definity did impact site performance on Google search:
When we intentionally removed rather large amounts of content (and thus google got a lot of 404's), Googlebot signficantly reduced its crawling rate/budget down to ~ 0.
This of course meant that other content on the site did not get crawled and indexed/updated which did have an impact on SERP rankings.

Google does this as a protection against temporary site outage/misconfiguration to avoid accidently removing large amounts of content from the index.

There seems to be some hard coded logic that is aware of the likes on XF sites
I doubt that as well. Google just does recrawl URLs from time to time even if they gave 404/410 in the past.
We are still seeing spikes of crawling activity for vB 3 vBSEO URLs every now and then, those URLs have been gone since many years.
IIRC there is even a post on their webmaster blog somewhere acknowledging this.
 
Last edited:
There's quite literally no point in redirecting these (in regards to Bing). The Bing crawler won't magically know what to do with a completely differently structured page with completely different data just because you redirect it there. If they would accommodate for this new data, they certainly would handle the transition from /likes to /reactions as well.
 
There's quite literally no point in redirecting these (in regards to Bing). The Bing crawler won't magically know what to do with a completely differently structured page with completely different data just because you redirect it there. If they would accommodate for this new data, they certainly would handle the transition from /likes to /reactions as well.
One reason would be if you track valid 404s (e.g. threads merged/deleted, posts/attachments deleted, etc.) which you actually want to fix and having lots of noise (e.g. /search, /likes, etc.) makes the signal to noise ratio high. There are a lot of things that could be built into XF to make managing 404s not necessary, but unfortunately XF doesn't do this. E.g. URLS to deleted posts could still redirect to the next post in the thread or at least the thread. Merging two threads could automatically redirect the old thread to the new thread (without the option for the old thread to still be displayed which si the only option now), etc.
 
E.g. URLS to deleted posts could still redirect to the next post in the thread or at least the thread.
That would be semantically incorrect and should not be done.
The next post ist not the new URL of the deleted post - the deleted post is gone for good and can't be redirected anywhere.

Merging two threads could automatically redirect the old thread to the new thread (without the option for the old thread to still be displayed which si the only option now), etc.
Yep, that makes sense.
 
That would be semantically incorrect and should not be done.
The next post ist not the new URL of the deleted post - the deleted post is gone for good and can't be redirected anywhere.
I'll give you navigating to the next post is bad, but navigating to the parent thread would be better than nothing.

Consider this:

Someone clicks a post # and copies the resulting URL into another post (or bookmarks it, posts it on Facebook, etc.):

https://xenforo.com/community/threads/redirect-likes.168068/#post-1364888

vs. someone right-clicks and copies the post # on a post and copies the link into another post (or bookmarks it, posts it on Facebook, etc.):

https://xenforo.com/community/threads/redirect-likes.168068/post-1364888

The first one will still resolve to the thread if the post is deleted. The second one will not, which is a shame, since it's actually more durable (e.g. will survive repagination)

If you bookmark/post/link to a post within a thread, you certainly would still want to get to the thread rather than a 404 with no clue what thread it was contained in, even though the thread still exists.
 
Back
Top Bottom