Welcome to the Q&A Forum

Mobilio

So far we have Microdata, RDFa and JSON-LD.

I'll cover Microdata and RDFa in same because they're similar. So they both are addition to HTML attributes indicating what Schema.org field names correspond with what user-visible text on the page. Works perfect but need lot of developer work and designer changes. Because both backend (admin interface) and frontend (HTML) must be changed. And there are many issues that can be messed - incorrect implementation, "rich snippet spam", software bugs, etc. Look easy as 1-2-3 but in reality it's pain (PIA) for implementation and support. Example - only product implementation require in backend least 10 edit boxes if they're manual filled.

JSON-LD - it's relative new protocol based on Schema.org. The main benefit is that you split representation layer (HTML) from semantic layer (JSON-LD). In prev. formats - they're same and linked each other. Now they're split. This give you much more freedom than before. You can place HTML data whatever you wish and just add hidden JSON in head or in content that will add semantic marking. This is future (for now).

If you wish to read more about creation of JSON-LD then this article is for you:
http://manu.sporny.org/2014/json-ld-origins-2/

Mobilio

As long as you return 404 this is good. But the question is "why search results in CMS are indexed?". And answer of this is "Joomla". I will keep them as 404.

Few domains that didn't exist anymore:

file:///Z:/Clients/Current%20Clients/Ethicare%20Dental/Articles/JULY%20ARTICLES/Is%20the%20new%20best%20toothpaste%20in%20the%20world%20all%20its%20cracked%20up%20to%20be.docx <- http://www.ethicaredental.co.uk/is-the-new-worlds-best-toothpaste-all-its-cracked-up-to-be/
http://www.hypnotherapydunstable.co.uk/img/dentist.jpg <- http://www.ethicaredental.co.uk/taking-the-fear-out-of-a-trip-to-the-dentists-chair

Mobilio

Yes - there is bug in your robots.txt. You should wrote some as:
Disallow: /?display=table
or:
Disallow: /?display=*

Mobilio

This is pretty awesome! And i like it!

Would you like to extend this in Jan 31 this year for all members that we can see? I can make importer of their table to XML so later we can use =ImportXML. And one bad news "Only members who have logged in within the last 60 days and have 25 or more MozPoints will be shown." So there can be other "top valued" members but we can't see it because this limitation.

And also special users like "ghost in the shell" -> https://moz.com/community/users/2380

Mobilio

Ok, here is document:
https://docs.google.com/spreadsheets/d/1YGkY2Yz1RefwcG4nqJ1lbnEfTghn7pRtb9di2HICiaA/edit?usp=sharing
captured before hour.
I leave it as-is just extract username but later make regex to extract their posts and ugc posts.

This is Top 4291 as far as i see.

Here is procedure how to get it:

wget https://moz.com/community/users - i love *nix command line tools but you can use web browser and save as HTML.
Open HTML for editing.
Find " var top_users =" and delete everything BEFORE it. Now document should start as "[[1,"
Find "]];" and delete semicolon and everything AFTER it. Now documents should end with "]]"
As you can see we have well formated JSON. Bad news - i can't import it in Sheets due bug in =ImportJSON function.
Let's go in PlanB - https://json-csv.com/
Copy all JSON into clipboard
Paste it on json-csv.com
Push button, get CSV
Import CSV into Google Docs
Enjoy!

Just use some of mine formulas to extract number of posts or UGC. I can make formula for extracting username too.

Mobilio

TL;DR - It dependents.

Main answer is - how you get traffic? If this is traditional way - "write great content and they will follow" this also bring comments. And comments create new content with new traffic. Because Google/Bing/Yahoo can index content and comments too. Moz uses this way in their blog section for example. This is organic way.

But in new way - comments in social networks bring traffic. Because more people seeing that person comment/share/like something and they do "curiosity driven research" what is this content. But this bring only social network traffic since not all of Facebook comments can be indexed from search machines. This is social way.

This is theory about traffic and i describe only two of them (they're much many but i simplify little bit things). Now turning things into practical. Disqus can have some issues with crawling comments from Google. This is well known and described many times: https://help.disqus.com/customer/portal/articles/762307-why-isn-t-google-indexing-my-comments-
Facebook comments can't be indexed at all too from Google. There are some articles from 2011 that they can but i can't find working example.

That's why it's best if you can check - are your existing comments indexed? If they are you can leave both commenting systems. It's much better. If they wasn't indexed you can switch to Facebook commenting system. Also check on other site can you find some Facebook comment in Google.

And one suggestion "Don't Build Your House on Rented Land". If you can use your own commenting system this will be much better. Just like Moz.

Mobilio

"1. Focus Your META Description
Let's say that, for some reason, we really wanted that SEOmoz blog post to rank for "January 19". One solution is to make sure that phrase appears in our META description for the relevant page. If Google can find the matching copy in your description, they're more likely to use the tag as is. It's also just a good exercise – figuring out what your core target keywords are and targeting them naturally in your META description (don't just make it a list of keywords, of course) will help you focus your overall on-page SEO efforts."
https://moz.com/blog/why-wont-google-use-my-meta-description

And i didn't see GlobeCar in your meta description. That's why on one keyword Google didn't display your meta description and show something from code that contain GlobeCar. Also in title you have "Globe Auto", domain is different and this will confuse user so they didn't display original description.

I forgot also this one:
https://support.google.com/webmasters/answer/35624?hl=en
"Google will sometimes use the meta description of a page in search results snippets, if we think it gives users a more accurate description than would be possible purely from the on-page content."
seems that they think that your description is different than on-page content.

Mobilio

If you just will display icons then it's OK. But it's dangerous to display text with display: none:
https://www.seroundtable.com/google-display-none-20626.html
https://support.google.com/webmasters/answer/66353?hl=en
https://youtu.be/7y-m_jiayLQ
https://youtu.be/B9BWbruCiDc
https://www.seroundtable.com/google-hiding-content-17136.html
https://www.seroundtable.com/google-hidden-tab-content-seo-19489.html

Because hiding text give "less weight" and you may get into trouble.

Mobilio

There are many plugins for schema.org markup in WordPress directory. But this one:
https://wordpress.org/plugins/schema-creator/
is extremely useful.

As alternative - you can use JSON-LD and put it anywhere in HTML.

Mobilio

There are few bugs:

1. You have 2 items for same language. 2 pages are for EN and 2 for DE. You should have only one item per language. This mean 1 page for EN and 1 page for DE.

2. You need x-default item. X-default item should be for worldwide users - so will be EN for your case. Please note - you must have x-default and EN.

So here is example how it should be in your case:

And this snippet should be placed on all that 3 pages.

I hope that this helps. If you have other questions first read Moz articles about hreflang:
https://moz.com/learn/seo/hreflang-tag
https://moz.com/blog/hreflang-behaviour-insights
https://moz.com/blog/open-source-library-tool-check-hreflang
https://moz.com/blog/using-the-correct-hreflang-tag-a-new-generator-tool

And read some Google information about it:
https://support.google.com/webmasters/answer/189077?hl=en
https://support.google.com/webmasters/answer/2620865?hl=en

So - hreflang are trick but isn't something impossible.

Mobilio

Clean (w/o parameters) canonical URL helps Google to understand better your url structure and avoid several mistakes:
https://googlewebmastercentral.blogspot.bg/2013/04/5-common-mistakes-with-relcanonical.html <- mistake N:1
http://www.hmtweb.com/marketing-blog/dangerous-rel-canonical-problems/ <- mistake N:4

So - your company that giving this advise is CORRECT! You should provide naked URLs everywhere when it's possible.

Mobilio

This is normal. First you need to evaluate this guide: http://www.elite-strategies.com/flavors-types-google-penalties/ and see what kind of penalty you have. Then you need to get all data from Analytics, Search Console both Links and Analytics there.

And here is huge analysis...

You need to check timeframe of your sudden drop. Are there some turbulence in SERP? New algo released? Anything weird in your own castle (server, CMS, hosting provider, etc)? If there is something - fix it. Also read about algos/turbulences in industry news sites.
Check for messages in SearchConsole about "Manual Actions" and fix them if it's any.
Check for sudden change of your links and create linking profile. If there are new and weird links from spammy sources - disavow them ASAP. If there is overoptimization for keywords - change anchor or disavow. This can be something Penguin algo related.
Check for sudden change of your site - thin content, aggregated content, duplicate, irrevelant, etc. Check also user signals - low CTR, high bounce rate, low time on site and low % of returning users. Also check bad UI - design, navigation, mobile redirects, site speed, ads. Check for low trustworthiness - in Google Raters Handbook that was recently released. This can be - providing contact information, ToS, PP, etc. And last - overoptimization. This can be something to Panda algo related.
Check for other algorithms - like TopHeavy aka Page Layout Algorithms that may affect your site.

So after you know what can be your drop related then check and recheck. Isn't first nor last time when you can think about Panda but in reality was Penguin or vice versa. This can save you few months. After you're 100% sure what was then try to make changes to escape from algo penalty.

Same happens to me almost year and half. We're running brand new site and faced sandbox algo. But we make few links more than we should and from low quality sites. So site was shown in SERP for month in first or second page and then suddenly drop to 10 page. Our organic traffic was fell to 1/10 of before almost overnight. But we make new traffic sources - social media and referral traffic so we could live without organic. We change links, make diversify linking profile changing links, anchors or even both. We also delete some links. But nothing happen in SERP position. Then i use disavow tool and 2 weeks ago there was algo refresh (Penguin). Almost overnight search traffic jump more than 10x and recovering previous positions and getting new and ever better positions. Today - this site make traffic as 10x than after recovering.

The problem is that you need to make changes in site against penalty. And wait for algo refresh. You also need huge patience because some drop can happen anytime, but for recovering you need algo refreshes. And refreshes take months or even years.

Mobilio

Well they probably explain this better:
https://support.google.com/webmasters/answer/66356?hl=en
http://searchengineland.com/google-links-in-a-press-release-should-be-nofollowed-like-advertisements-168339
http://searchengineland.com/google-adds-large-scale-guest-posting-advertorials-optimized-anchor-text-to-list-of-link-schemes-168082
http://www.wordstream.com/blog/ws/2013/07/24/follow-nofollow-links

Please note two things. First they talk about "optimized anchor text". This mean that if you link your domain with keyword anchors. There are many example like "wedding dresses", "wedding photographer", "wedding cakes", etc. If you link only "naked url" (url w/o keywords in anchors) then it's OK. And second - press releases can mess your linking profile to your site. So there are many articles all over the internet about something holy grail of linkbuilding - "perfect link profile". I won't share links because this is some myth and in reality this is hard to be proof. But why this is important? Because let's says that you have 70/30 or 60/40 external links to home page vs. internal pages. So passing press releases (most of them are to your home page) will change this to 90/10 or 85/15. And this could raise a red flag for Penguin.

As you can see it's tricky for you.

Now it's also tricky for sites sharing press releases. Because general each PR is text + link. And building too much content with outgoing links isn't good. Could trigger Panda. Now add that same content appear on many sites at once. Could trigger Panda too. Now also add "linking schemes" in 1st link. In result this is headache for them. That's why they use "nofollow" to minimize risks for them.

As you can see now - it's tricky for everyone. And safe way for linking is only "nofollow".

Mobilio

Just as Dirk says you can use Analytics. But... dependent on GA version (classic/universal) you can add this in tracking code:
_gaq.push(['_setSampleRate', '100']);
_gaq.push(['_setSiteSpeedSampleRate', 100]);

and for UA code is:
ga('create', 'UA-XXXX-Y', {'sampleRate': 100});

Please note - for classic you can put code after initialization and before track page view. For UA you need to change way of initialization to include this sampleRate. Then you can go in GA -> Behavior -> SiteSpeed and will get speed data for about 100% of visits.

Later - use PageSpeed Insights to track speed of pages. If you're in WordPress you can use plugin:
http://mattkeys.me/products/google-pagespeed-insights/
and check score of all pages/posts. There is free version, just scroll to bottom.

After you can use Pingdom Tools, GTMetrix and WebPageTest to see how your site is loading on different geolocations, browsers and connection speed. All of tools give some scores and suggest improvements, just follow their suggestions for better performance.

And if you are willing to get even higher speed improvements you need to read least two books:

Ilya Grigorik - High Performance Browser Network; http://shop.oreilly.com/product/0636920028048.do you can read it for free http://chimera.labs.oreilly.com/books/1230000000545
Lara Callender Hogan - Designing for Performance;http://shop.oreilly.com/product/0636920033578.do you can read it for free here http://larahogan.me/design/

So both books are amazing and completing each other. First explaining server performance, network performance, some specific improvements about browser core, etc. Second book is explaining how to clean up actual site for better performance - minimizing and cleaning CSS, selecting best images, JS improvements, HTML tricks, etc.

After you reading that books and using tools you can optimize site slashing minimum 50% of it's footprint.

Mobilio

Sure:
http://www.wordstream.com/learn
http://www.annielytics.com/annielytics-dashboard-course/

Best of the best of the best!

Mobilio

You can't use ALT text to background images.

If your images are loaded as img tag then you have title, alt. Example: some alt

But background images isn't regular images loaded with img tag. They can be loaded only with CSS styling. And therefore they doesn't comes with title and alt. Example:

As you can see you can't use title or alt for image there. Only way is to define title as this:

or better this one:
some alt
I know that isn't 100% background but also helps.

Alternative way can be ARIA. If you love background images but really wish alt text you can use WAI-ARIA and define them as:

...
But GoogleBot doesn't support aria-label for now.

Mobilio

Hotlinking can't be stopped with .htaccess. Because users will see "hotling warning" but bots will still see link.

Since i have exactly same issue with mine images - i disavow them w/o any doubt. This is low quality directory with all images based on some keywords like "aztec". They scrape SERP for images, get top 100 and make "website". No, no please, no.

That's why disavow them ASAP.

Mobilio

You can use Moz Rank Tracker!

OR SemRush, Google SearchConsole, Advanced Web Ranking, RankWatch. And many, many more. They all work almost perfect and have almost all features that you need. But you need to evaluate them to see best solution for your needs. For example - SearchConsole is 100% accurate and amazing (and free!!!) but can't produce reports automated. You need to login, extract data, save to CSV, import in Excel, do some changes and save as PDF.

Mobilio

Link here is nofollow FYI.

Mobilio

Hey - i just found that we can "scrape" historical MozPoints using their profile pages:
https://moz.com/community/users/397332
if you view source this you can see "var mozpoints_data" with JSON inside.

I can scrape all historical points for Top1000 for example starting from 2015-11-01 so you can built even better sheet.

Welcome to the Q&A Forum

Browse the forum for helpful insights and fresh discussions about all things SEO.

Mobilio

@Mobilio

Best posts made by Mobilio

Products

Moz Solutions

Free SEO Tools

Resources

About Moz

Why Moz

Get Involved