If you are concerned about how your community is performing in search engine results (SEO), there are a number of questions to keep in mind when looking to optimize your community software of choice. We frequently get asked questions in our support inboxes and in our public community about how Discourse handles SEO. In this article, we’ll aim to demystify the most common Discourse SEO questions, including:
Thanks for the blog. We have been using Discourse for two years now as successor to Confluence and it definitely attracts users for the content they find relevant in an easier way. On the technical SEO-side, Discourse works great in general. It performs well, and is easy to have content correctly indexed.
I will share some finding that are still relevant. Most challenges were already known and most have been included in a release. The biggest challenge seems to be better multi-language support, which is probably only relevant for a small part of the installations that target users from multiple countries or countries such as Belgium with three official languages (French, German and Dutch).
Multi-language may seem like a small thing, but only by providing translated content in multiple languages instead of just Dutch the traffic has approximately grown by 400%, significantly better distributed from more countries in our commercial target area.
Keyword Cannibalization
Biggest problem now seems keyword cannibalization when publishing a wiki in multiple languages. Discourse currently out-of-the-box does not provide relationships between topics that contain identical content in different languages. For now, we hope that the ping-back links in posts help Google consider it as the same pages, but we don’t know for sure. More information on Link related topics using link-tag for better helping search engines
Category per language
For multi-language, we now use multiply the real categories with the two languages mostly used (English and Dutch). It works to dedicate a category to a specific language, and Google seems even without hints to make reasonable choices when detecting actual language of a topic, even when the language differs from the meta-data (see Language support on category level for more relevant search results).
I never heard of keyword cannibalization. At the moment I am studying keyword clustering by means of Rush Analytics. Should I somehow consider keyword cannibalization while clustering? I am relatively new to SEO and would highly appreciate your assistance.
Sitemaps
I read the section on your SEO post about sitemaps, but have some concerns.
We have over 9,000 orphaned pages showing up in our SEO monitoring tools from our forum.
Looking into this further, they’re all being access via the sitemap. They have no other internal or external links pointing to them.
My view is this is not a good thing for SEO and can impact our crawl budget. Google is highly unlikely to index orphaned pages.
Meta data
The second most common SEO issue we have with Discourse is that title and description meta data is too long. What are the default character limits for this? It feels like this needs to be reviewed in line with the changes Google have made to allowed characters/pixel width over the last few years.
That’s not true. Those are included in th budget. From theirs point of view it is just your headache if you don’t lead users thru content. Most of users are bouncers, though.