Stream: fmg
Topic: Old HL7 Wiki dominates search results
Josh Mandel (Aug 31 2021 at 15:22):
For this week's meeting: can we discuss tactics for getting search results to lead people to Confluence rather than wiki.hl7.org ?
Lloyd McKenzie (Aug 31 2021 at 15:26):
We could rename wiki.hl7.org to wiki-old.hl7.org. Those in the know could still find stuff if needed, but the spiders would presumably see that everything that linked to it is now pointing to something that no longer exists and quit sending people there.
Josh Mandel (Aug 31 2021 at 15:26):
(One trick we might want to try: https://moz.com/learn/seo/canonicalization -- but this is definitely not my area of expertise, and if others have some SEO experience, it'd be great to brainstorm.)
Josh Mandel (Aug 31 2021 at 15:27):
We don't just want to take away content, I think -- right? Like, it'd be good to leverage our existing strong placement.
Josh Mandel (Aug 31 2021 at 15:28):
(And possibly HL7 has some organizational capability or external support in place for SEO? It'd be worth understanding this.)
Lloyd McKenzie (Aug 31 2021 at 15:33):
In general, wiki content is only intended to be there for HL7 insiders. External folks shouldn't be going there ever. Renaming wouldn't take it away, it'd just break all existing links - and make the search engines not want to point to it anymore.
Josh Mandel (Aug 31 2021 at 18:57):
Sure; I'm just saying that ideally we'd leverage our current (very strong!) search result positioning rather than just throwing it in the trash. I think things like "canonical" tags (and maybe optimized robots.txt, etc) could help with this.
John Moehrke (Aug 31 2021 at 18:59):
that can prevent search results.. but doesn't address where someone has saved a link that they should no-longer be using. Are we going to try to solve both?
Josh Mandel (Aug 31 2021 at 19:08):
I'd hope so! We should
1) keep old links working (or at the very least, keep them in a state where following them provides a good trail of breadcrumbs to the latest/greatest), and
2) convince search engines to index our newest/best content in favor of the old stuff
Lloyd McKenzie (Aug 31 2021 at 19:22):
There shouldn't be anything on the wiki that's relevant anymore. If there is, it should be moved. Eventually, we want the wiki to go completely away. The only reason it's still available is so we can propagate stuff that we missed or mine things we didn't realize we'd need and then discover we do.
Lloyd McKenzie (Aug 31 2021 at 19:23):
So I'm not super concerned with the search engines not finding anything on the old wiki anymore...
Lloyd McKenzie (Aug 31 2021 at 19:23):
If you need stuff from the old wiki, it should already be a last resort and you should already have some idea that it's there and search on the wiki itself is sufficient.
Josh Mandel (Aug 31 2021 at 21:00):
My point throughout this thread is about search engine optimization -- the old wiki has great positioning today. We should leverage that to enable a smooth transition if we can (rather than throwing it away).
John Moehrke (Aug 31 2021 at 21:07):
I suspect that the search engines will find what everyone needs to find on the confluence. I think that we should definitely tell the search engines to not index the wiki anymore.
I don't think the canonical tag is as helpful as just stopping all search discovery of the wiki content.
Josh Mandel (Aug 31 2021 at 21:13):
I don't think the canonical tag is as helpful as just stopping all search discovery of the wiki content.
Is this based on experience working with Advanced SEO guidance like https://developers.google.com/search/docs/advanced/crawling/consolidate-duplicate-urls?
John Moehrke (Aug 31 2021 at 21:15):
not directly. More based on the fact that the wiki should just disappear, and that the spirit of canonical is not what we have with wiki-->confluence.
Josh Mandel (Aug 31 2021 at 21:18):
OK. Today our old wiki pages have notes and links like "Content on this page has been migrated to Confluence here: https://confluence.hl7.org/display/FHIR/IG+Publisher+Documentation" at the top, which feels like "here's where you should go for canonical content" to me.
Anyway, I really don't feel that strongly about this; just want to make sure we're setting ourselves up for a staged transition process guided by informed decisions. If we decide to "just pull the plug," so be it.
Grahame Grieve (Aug 31 2021 at 21:18):
I have had to help people fooled by this, and partly because that warning is nowhere near in-your-face enough
Josh Mandel (Aug 31 2021 at 21:25):
(I am one of those people every time I do a Google search; using "canonical" links or removing the wiki would both solve this problem. In theory, making the warning more in-your-face would help too, but it's hard to get people to focus where you want them to :-))
John Moehrke (Aug 31 2021 at 21:25):
in theory we could add to each wiki page that has a "Content on this page has been migrated..." with also the canonical html header decoration. I can guess that this is possible to do. Not sure how to do this given wiki. I don't find quickly a how-to do this with wikimedia. Surely it could be done
Lloyd McKenzie (Aug 31 2021 at 22:04):
All sorts of things 'could' be done - but most would require a lot of bandwidth and we don't have a lot. I'm not sure there's an SEO way to migrate goodwill from our wiki to Confluence. It's mainly driven by volume and quality of inbound links, as well as what people have clicked on. But I'm not an expect - do we know anyone who is?
Melva Peters (Sep 01 2021 at 21:49):
@Joshua Procious @Eric Schmitt
Joshua Procious (Sep 02 2021 at 12:16):
I think we can use a noindex tag on each page of the wiki and shrink it's presence on returned results by Google very quickly. I'll look into it today...
Josh Mandel (Sep 02 2021 at 15:31):
Joshua Procious: I think we can use a noindex tag on each page of the wiki and shrink it's presence on returned results by Google very quickly. I'll look into it today...
This is true, but if you have the ability to add tags, it may be good to see if you can add canonical tags on a few of the most popular pages (do we have analytics set up to know what the most popular pages are?)
John Moehrke (Sep 02 2021 at 17:33):
agree. If you are updating each page, the canonical would be good.. but to set a robots.txt at the root telling all search engines to ignore wiki.hl7.org would be a nice first step
Last updated: Apr 12 2022 at 19:14 UTC
