SEO: structured data - making elgg sites discoverable by google etc.

i recently found that my site is scoring very poorly in google. this appears to have mainly started after i changed the domain name a while ago.. however, in the process of fixing this i have found that there are many possibilities that are available to improve elgg from an SEO perspective and some of them would definitely be best added to core.

for example, google uses structured data to 'know' what is on webpages and to support structured data is to get a better presence in google. this is the main list of data object types: http://schema.org/docs/full.html

so every elgg object would need to have tags added to allow google to know what is the name, date, description and so on. i have done this for videolist on my site as a test and it appears to be working ok. however, i needed to over-ride several core views to do this and it would definitely make sense that the tags just be added at the core level. i am about to go through and update blogs, pages and other objects on my site too...

does anyone have any comment on this?

  • I am about to start work on a SEO plugin for a project. Part of the problem is knowing if a page is actually about an entity. URL sniffing can be useful at times, but is also misleading. I am currently inclined to build upon this: https://github.com/Elgg/Elgg/pull/9102, and push entity information to elgg_view_page(), so that the page shell can then use an entity in the tags hook.

    Another consideration is entity privacy. Structured data would only make sense for publicly accessible content, however you loose elements of the graph if you drop all the entities with logged in and other access. That was something I have tried to solve here: https://bitbucket.org/hypejunction/hypediscovery. 

    I will know more and have more thoughts once I dive into it. Will keep you posted.

  • On a related note, elgg.org does pretty poorly at it as well. I have opened an issue about it recently: https://github.com/Elgg/www.elgg.org/issues/88

  • i have updated my interconnected plugin quite a lot recently, with a lot of fixes and some seo improvements. i haven't uploaded the latest version to github or the community though as i am still working on it and it is not yet 100% separated from my theme plugin here.

    i have used steve's sniffer code in the plugin and also needed to handle some cases where the sniffer wasn't completely effective. the net result is that my plugin is quite large and replaces several aspects of the core code, such as title handling for pages.

    the privacy aspect is something i have thought about, but i figured that if an entity is set to be non public, then it will not be visible to search engines anyway, so there is no need to remove tags from those entitites.

    the way i am coding the structured data here is to just over-ride the object files for each object type - to add specific tags and then also over-ride image-block and some of the page views to insert tags for the data container elements.

    so far videos and blogs look fine.

  • btw, your link to hypeDiscovery on bitbucket is broken

  • thanks. i am not 100% clear on how this would help SEO. for example, it appears that the main problem i am having with google is due to their 'panda' updates that attempt to rank sites in terms of 'quality' and 'authority'. since my site has a lot of pages which are part of videolist and the bookmarks plugin, there are many pages which have little content on them and the content is essentially duplication of content from other sites (such as youtube). this has apparently caused google to basically completely bin my website! i get almost no traffic from google at all.

    the advice i have been given to resolve this is to delete all the pages that are of this subset and start again. however, i am not willing to do that as i collect the videos for a purpose. so... i could either set all the videos to be for logged in users only or something of that nature. unfortunately though, that does not prevent other users from just uploading items and setting them to be public.

    using the walled garden with the type of idea offered by hype discovery could be of some use, but even then it is no guarantee that this problem won't continue.

    is there any newer version of hype discovery that i can use with 1.12.x?

    do you have any comment on it's effects on seo Ismayil? @ihayredinov

  • so having done further research into google's algorithms, it appears that my site may be being penalised due to it's use of the elgg core bookmark plugin, plus videolist as these produce pages which often have little unique content on them and for other reasons relating to external linking. i have changed my site now so that the robots link in the head of the videolist pages includes the directive 'noindex' and also so that bookmark links have the nofollow directive applied to them. this should mean that google doesn't penalise the site due to these items. unfortunately i don't know how long it will take to learn the results of these changes. this appears to be an issue for all elgg sites that use these and similar plugins.

  • My first stab: https://github.com/hypeJunction/Elgg-seo

    This is a URL rewriter and metatag editor. Haven't looked into schema.org yet.

  • Good luck on this, guys. It appears that elgg users, so far as this discussion is representative, and from my own experience, aren't particularly in Google's ken. I've been attempting to build a beckoning site on the subject of fairness, SimplyFair.net. But when I query Google on "fairness" it's nowhere to be found, at least I haven't yet dug that deeply. Without decent Google listings, we're going to be hampered in seeking members for elgg social networking sites.   

  • google's updates in recent years have solved some problems and caused some others. in my case, my site is literally not ranked for any relevant keywords at all! i have found numerous tools that you can use to help get to the root cause of the problem for any particular site - though even then it requires a lot of exploration to find issues to fix.

    i think that some of the changes that are needed are best applied to the core code, so it's tough to offer anyone any advice that will be guaranteed to work and that can be easily applied to a standard-ish elgg installation.

    http://www.ranktrackr.com - lets you keep up with how your site is performing for keywords

    http://www.alexa.com - gives a variety of estimated tracking data points and allows comparisons with other sites

    https://www.google.com/webmasters/tools - lets you engage with google to improve your ranking and perform various action that may help, such as providing a sitemap.

    http://www.semrush.com - lets you run detailed tests and processes related to seo

    https://toolbox.seositecheckup.com - more tests for seo