Test your site on https://developers.google.com/speed/pagespeed/insights/ Free trial seo report rank https://www.woorank.com/ Also on https://www.webpagetest.org/


Videos https://www.youtube.com/watch?v=6UeRMMI_IzI&index=7&list=PLWa0Ky8nXQTaFXpT_YNvLElTEpHUyaZi4 Github https://github.com/WPO-Foundation/webpagetest Docs https://sites.google.com/a/webpagetest.org/docs/advanced-features/webpagetest-restful-apis

General Tips

  • Be sure that each page has a unique title on each page <title>Nice page</title>, 10-70 chars, with keywords
  • also <meta name="description" content="It's really nice page <%= @user.name if @user.present? %>"> should be 70-160 chars, uniq per page
  • use one <h1> per page with keywords but do not duplicate title, use h2-h6 for other keywords.
  • use alt atribute for description of images, but also image source link should we unique and self explanatory
  • Links within your content tend to carry more weight than links within a sidebar or footer.
  • try to have more and more inbound links from other sites. Increase link count using: social media, directories, some sites links to competition, try to write them explaining why your site is better. To find inbound links use google webmaster tools, https://moz.com/researchtools/ose/
  • if your content expires (like job posts, wheater data) than link old pages to category pages (group of similar results)
  • outbound links can be no-follow, especially outgoing links that are not relevant (do not have quality content). For example, links to a Feedburner page.
  • anchor text plays the most important role in link building. If you want to rank for ‘blue widget’ then you want the anchor text of the link to be <a>blue widget</a>
  • 25-70% should be only text (not html markup tags)
  • add XML sitemap (with correct protocol http/s, subdomain, trailing slash) if more than 100 than only most popular
  • use 301 redirect from root to www domain (or vice versa) do not split value to two different domains (also from IP address)
  • replace undrescore _ with dash (hyphens) - in urls
  • flash and frames are not indexed
  • nice mobile rendering (buttons at least 48px width/height, 32px padding around tap targets)
  • add viewport meta tag <meta name="viewport" content="initial-scale=1.0, width=device-width"> and use CSS media queries to apply different style depending of screen size
  • enable gzip compresion, optimize image, eliminate render-blocking javascript and css, leverage browser caching, minify css, to have small page size (less than 320Kb) and load under one second
    • google web fonts are cached only for one day (each day, it is downloaded again). For different browsers it uses diffent files: woff2, svg … Download fonts using this heroku app. Change folder prefix to poppins/, and extract downloaded fonts (without folder) to app/assets/fonts/poppins so you can see the file app/assets/fonts/poppins/poppins-v1-latin-regular.svg. Copy/paste the content to some app/assets/stylesheets/popins.scss file and include in applications.scss with @import 'popins';
  • add favicon. It is fine just to add to the root of web site (for rails just copy your png to public/favicon.ico) Size should be 32x32. Some of the best favicons is https://cdn.shopify.com/shopify-marketing_assets/static/shopify-favicon.png More info on https://en.wikipedia.org/wiki/Favicon#How_to_use
  • custom 404 error page
  • add language <html lang="en">
  • SSL secure (http should redirect to https), add STS in header, xml sitemap, link to css files to use https
  • social media
  • optimize css https://csswizardry.com/2018/11/css-and-network-performance/

Rails gzip

It should show in Response Headers:


# Gemfile
gem 'heroku-deflater', :group => :production

# config/environments/production.rb
  config.assets.compress = true



/robots.txt is used to tell some good crawlers wheter they should read or not. Bad crawlers will read the site anyway.

# public/robots.txt
# See http://www.robotstxt.org/robotstxt.html for documentation on how to use the robots.txt file
# To ban all spiders from the entire site uncomment the next two lines:
# User-agent: *
# Disallow: /
# Sitemap: http://www.example.com/sitemap.xml.gz

You can use meta tags to disable some well known robots. Add <meta name="robots" content="noodp,noydir"> to disable Open Directory Project and Yahoo Directory to provide description for your site


sitemap_generator gem is excellent tool. This gem follows protocol

echo "gem 'sitemap_generator'" >> Gemfile
rake sitemap:install # this will generate config/sitemap.rb
vi config/sitemap.rb # put some `add path, options`
rake sitemap:refresh:no_ping # to generate sitemap to public folder without ping
less public/sitemap.xml.gz
gunzip public/sitemap.xml.gz && chromium-browser public/sitemap.xml

Use Post.find_each instead of Post.all.each since we do not want to load all posts in memory.

To run every day, you can use whenever to call rake sitemap:refresh -s

SitemapGenerator::Sitemap options

  • Sitemap.default_host = 'http://example.com' is your default url
  • Sitemap.create_index = true always create index file <sitemapindex><sitemap><loc>http://www.example.com/sitemap1.xml.gz</loc></sitemap></sitemapindex> that will link to other sitemap files definition
  • Sitemap.sitemaps_host = 'http://s3.amazonaws.com/my-site/' where we will deploy sitemaps. User with Sitemap.adapter =
    • Here include_index (include <url><loc>http://www.example.com/sitemap.xml.gz</loc></url>) be off since sitemaps_host is different than default_host and we can not have different domains in same sitemap
  • Sitemap.public_path directory used to write localy before eventual upload (default is ‘public/’).
  • Sitemap.sitemaps_path = 'my-folder/' is relative folder where to store sitemaps, it will be used to determine url in index file.

Options for add base on Sitemaps.org protocol are:

  • changefreq default is weekly
  • lastmod default is Time.now
  • priority default is 0.5
  • expires request removal expires: Time.now + 1.day

You can add some existing sitemaps with add_to_index '/my-old-sitemap.xml.gz'


For heroku you need to move it on S3. It is straighforward configuration, AWS S3 does not need to be public.

SitemapGenerator::Sitemap.sitemaps_host = "http://"\
SitemapGenerator::Sitemap.sitemaps_path = 'my-sitemaps/'
SitemapGenerator::Sitemap.adapter = SitemapGenerator::S3Adapter.new(
  fog_provider: 'AWS',
  aws_access_key_id: Rails.application.secrets.aws_access_key_id,
  aws_secret_access_key: Rails.application.secrets.aws_secret_access_key,
  fog_directory: Rails.application.secrets.aws_bucket_name,

If your bucket is not on us-east-1 region you need to add fog_region option to adapter

  fog_region: Rails.application.secrets.aws_region

Than you need robots txt to point to that sitemap url ( bucket-name.s3.amazonaws.com/filename)

# public/robots.txt
Sitemap: http://s3.amazonaws.com/my-site/my-folder/sitemap_index.xml.gz

Test with heroku run sitemap:refresh.

Webmaster tools

Check your sitemap on google webmaster tools. Just click Add property and and verify your domain. (For AWS S3 you need to enable website hosting). Refresh robots.txt to point on new sitemap location.

From sitemaps.org you can see that one sitemap should not have more than 50.000 links and uncompressed size 10MB. Url should be less than 2048 chars.

You can use search to see current index status Just type info:www.my-site.com and you can find links there. Or you can add to https://www.google.rs/search?q=site:www.my-site.com

  • site: search only for this domain or even subdirectory. You can use this to see if some page is indexed, just add uniq part of page url
  • link: find links for specific domain or subdirectory or pages (this is no longer supported as of Jan 2017)
  • cache: see current archived copy of domain or page. It is better to use https://www.google.com/webmasters/tools/googlebot-fetch since google crawler can execute javascript. You can submit to index on the page, but even google renders it properly (with ajax/angular data) it will not add to index for some other reasons.

Some reasons could be post

  • server can not respond fast enough to all crawler requests, maybe it will help to reoder sitemap

All links in sitemap should return 200 (not redirection 3xx). Also you need to check this answer for duplicates (this is common for angular since data is fetched after page load), canonical url (this is common when you have search queries that return same results).

If you have more domains on same server than you need to iterate for each default_host and to call SitemapGenerator::Sitemap.ping_search_engines (since by default is to ping when whole sitemap process finishes, so than it will ping only last one).

Other tools

Chrome plugins


You can follow http://schema.org/docs/gs.html to add microdata in your markup like: <div itemscope><h1 itemprop="name">Duke</h1></div>

You can add head tags for facebook open graph protocol. In Rails it will be like:

/ haml
- content_for :head do
  %meta{property: "og:type",        content: "article"}
  %meta{property: "og:title",       content: "MySite | #{@post.title}"}
  %meta{property: "og:image",       content: share_image_url(@post)}
  %meta{property: "og:description", content: @post.description}
  %meta{property: "og:url",         content: post_url(@post.id)}
  %meta{property: "og:site_name",   content: "MySite"}
  %meta{name: "twitter:card",       content: "photo"}
  %meta{name: "twitter:site",       content:"@mysite"}
  %meta{name: "twitter:title",      content: "MySite | #{@post.title}"}
  %meta{name: "twitter:image",      content: share_image_url(@post)}
  %meta{name: "twitter:url",        content: post_url(@post.id)}


You can enable Destination URL Auto-tagging. Automatically tag my ad destination URLs in Account Settings -> Preferences

User mailing list

If someone is linking your site or your competitor site, you can email the site owner (need a script that will search for email address on that site) saying that you have some new content, video or other material that will be interesting for his audience.

Customer Analitics

You can use some of the services: Google Analytics, Firebase anal, Amazon anal, Fabric Answers, Mixpanel, Keen, Segment, Amplitude, Localytics, Ionic Analytics

SEO search enginge optimization

hangouts tutorials https://plus.google.com/collection/oSHR9

  • subdomain is not bad for rank, since google can treat as two different sites


  • url is most important and should point to single state (not multiple states). After hash # usually everythings is ignored so that should not be a router for your app. Do not use secret stuff in url like /session=1234/...?user=john
  • canonical url is the main key for this content, and can be defined using link rel canonical, redirects, sitemaps
  • imporant content that needs to be indexed should be on visible part of the page, not on ajax call since crawl will be able to index that

TODO https://www.youtube.com/watch?v=Afy7H04X9Us