Site Speed Optimization Case Study – www.psychologicallyunemployable.net
We are back with our next site speed optimization case study. Our case study site is www.psychologicallyunemployable.net hosted on WP Engine. Instead of making changes directly to the live site, we planned to clone the site on a staging environment before starting our site speed optimization process.
Before moving the site on staging, we took a complete backup of the site to keep it safe and secure in case of any untoward incident. The site can be easily restored back to the point where we started at any time during our site speed optimization process.
Our site speed optimization case studies are developed in a way to help site owners, webmasters and users running their sites in similar environments. They can learn, practice and implement whatever is required to speed up their own websites.
We experiment to discover the best practices for running websites faster before publishing it in an organised way to help site speed optimizers all across the world. Never forget to take a complete backup of your entire site before making any changes to it. Whenever possible, use a testing environment before changing anything on your live site.
Step 1 : WordPress Site Speed Optimization Analysis
Before optimizing websites for speed, we need to spot the bottlenecks responsible for reducing site load speed. Without getting at the root of the problem, it is not possible to speed up any website. WordPress site speed optimization process starts by analyzing site speed test reports.
Many online site speed testing tools allow detailed site speed test reports. These site speed test reports tell us about all the problematic areas of concern. Choosing all the speed hindrance factors one by one and resolving it in a systematic approach can speed up any website.
Different hosting environments impose different speed limitations and what may work in one case might fail in another case. As mentioned in our earlier case studies, no straight formula for site speed optimization exists. All the sites, web hosts as well as hosting environments are entirely different from each other. Every site needs a different site speed optimization approach. Moreover, you must also check if your site is limited by CPU Throttling.
Let us start our site speed optimization process by taking some of the most prominent site speed test reports to understand the existing site speed status of our case study site.
YSlow Grade : 84% (B)
Page load time : 8.53 s
Total page size : 2.15 MB
Total number of requests : 80
Load time : 5.61 s
Requests : 108
Perf. grade : 92/100
Your website is slower than 73% of all tested websites.
Compress Transfer – A
Compress Images – A
Progressive JPEGs – F
Cache static content – C
Effective use of CDN – X
Step 2 : Configuring .htaccess for WordPress site speed optimization
.htaccess is one of the most crucial files for loading websites successfully. Remember that any damage to .htaccess file may leave your site entirely inaccessible. After taking a backup of the .htaccess file, we did some changes in it.
php_value zlib.output_compression on
Header set Connection keep-alive
Header append Vary User-Agent env=!dont-vary
ExpiresDefault “access plus 14 days”
Header set Cache-Control “public”
Header set Cache-Control “max-age=7200, must-revalidate”
AddType text/css .css
SetEnvIfNoCase Request_URI \.(?:rar|zip)$ no-gzip dont-vary
SetEnvIfNoCase Request_URI \.(?:avi|mov|mp4)$ no-gzip dont-vary
SetEnvIfNoCase Request_URI \.mp3$ no-gzip dont-vary
.htaccess file can be easily edited through cPanel using your web host service account.
Step 3 : Configuring WordPress for Site Speed Optimization
Configuring WordPress for site speed optimization is our next step. Looking inside WordPress, we tried to find the troubling plugins which were either not required or the plugins which were not working well with our case study site.
- Removed Better WordPress Minify.
- WP Smush.it was used to optimize images.
- WP HTTP Compression was activated on the site, but we deactivated it after detecting a bunch of plugin errors.
- Entire backup of DB was also taken before moving any further.
- All Wordfence tables were removed.
- All WP Better Security tables were removed.
- We also removed w3tc_cdn_queue
In order to optimize the database, following conversions were done
ALTER TABLE id3dyt_dld_hits ENGINE=InnoDB;
ALTER TABLE id3dyt_formengine ENGINE=InnoDB;
ALTER TABLE id3dyt_formengine_data ENGINE=InnoDB;
ALTER TABLE id3dyt_links ENGINE=InnoDB;
ALTER TABLE id3dyt_posts ENGINE=InnoDB;
ALTER TABLE id3dyt_prli_clicks ENGINE=InnoDB;
ALTER TABLE id3dyt_prli_clicks_rotations ENGINE=InnoDB;
ALTER TABLE id3dyt_prli_groups ENGINE=InnoDB;
ALTER TABLE id3dyt_prli_keywords ENGINE=InnoDB;
ALTER TABLE id3dyt_prli_link_metas ENGINE=InnoDB;
ALTER TABLE id3dyt_prli_link_rotations ENGINE=InnoDB;
ALTER TABLE id3dyt_prli_links ENGINE=InnoDB;
ALTER TABLE id3dyt_prli_report_links ENGINE=InnoDB;
ALTER TABLE id3dyt_prli_reports ENGINE=InnoDB;
ALTER TABLE id3dyt_prli_tweets ENGINE=InnoDB;
ALTER TABLE id3dyt_pty_filters ENGINE=InnoDB;
ALTER TABLE id3dyt_pty_imps ENGINE=InnoDB;
ALTER TABLE id3dyt_pty_popups ENGINE=InnoDB;
ALTER TABLE id3dyt_pty_stats ENGINE=InnoDB;
ALTER TABLE id3dyt_seopressor_automatic_internal_links ENGINE=InnoDB;
ALTER TABLE id3dyt_seopressor_external_cloacked_links ENGINE=InnoDB;
ALTER TABLE id3dyt_seopressor_logs ENGINE=InnoDB;
ALTER TABLE id3dyt_seopressor_lsi ENGINE=InnoDB;
ALTER TABLE id3dyt_seopressor_roles_capabilities ENGINE=InnoDB;
ALTER TABLE id3dyt_seopressor_users_custom_roles ENGINE=InnoDB;
ALTER TABLE id3dyt_seopressor_visits ENGINE=InnoDB;
ALTER TABLE id3dyt_sharebar ENGINE=InnoDB;
ALTER TABLE id3dyt_so_tracking ENGINE=InnoDB;
ALTER TABLE id3dyt_spyders_live ENGINE=InnoDB;
ALTER TABLE id3dyt_term_relationships ENGINE=InnoDB;
ALTER TABLE id3dyt_term_taxonomy ENGINE=InnoDB;
ALTER TABLE id3dyt_terms ENGINE=InnoDB;
ALTER TABLE id3dyt_thesis_backups ENGINE=InnoDB;
ALTER TABLE id3dyt_thesis_terms ENGINE=InnoDB;
ALTER TABLE id3dyt_wp_rp_tags ENGINE=InnoDB;
ALTER TABLE id3dyt_zamlb_autoresponders ENGINE=InnoDB;
ALTER TABLE id3dyt_zamlb_creatives ENGINE=InnoDB;
ALTER TABLE id3dyt_zamlb_generalSettings ENGINE=InnoDB;
ALTER TABLE id3dyt_zamlb_keywords ENGINE=InnoDB;
ALTER TABLE id3dyt_zamlb_lists ENGINE=InnoDB;
The staging site was configured to display only a single post so that we can test its speed performance. We did some site speed tests once again and got the following site speed test report at GTmetrix.
Looking at the above GTmetrix speed test report, we can see some major problems:
- Specify image dimensions F (0)
- Serve scaled images F (14)
Images must be scaled before serving so that it can be served in the same size as uploaded.
We also realized that the landing page was over populated with the content. A simple solution was the use of “Read More” tag. It solved our purpose.
Once again, we did site speed optimization tests to analyse the site speed improvement results.
YSlow Grade : 87% (B) [Earlier: 84% (B)]
Page load time : 1.96 s [Earlier: 8.53 s]
Total page size : 747 kB [Earlier: 2.15 MB]
Total number of requests : 37 [Earlier: 80]
Page size : 1.1 MB [Earlier: 5.5 MB]
Load time : 1.91 s [Earlier: 5.61 s]
Requests : 43 [Earlier: 108]
Perf. grade : 90/100 [Earlier: 92/100]
Your website is faster than 72% of all tested websites. [Earlier: Your website is slower than 73% of all tested websites.]
First Byte Time – D [Earlier: A] It will improve when the site moves from staging to main server.
Keep-alive Enabled – A [Earlier: A]
Compress Transfer – A [Earlier: A]
Compress Images – A [Earlier: A]
Progressive JPEGs – F [Earlier: F]
Cache static content – B [Earlier: C]
Effective use of CDN – X [Earlier: X] It will also improve when the site moves from staging to main server.
Site speed test reports are fine, but something seems to be messing the site load speed. Doing an in-depth analysis of the site, we noticed the presence of a plugin Pippity.
Pippity plugin helps a site to get pleasing popups to increase conversion rate. We had our own doubts about Pippity pulling down the site load speed, but we wanted to confirm before making any statement.
We performed two set of site speed tests. One with Pippity Active, while another with Pippity Inactive.
Let us have a look at the site speed test reports to understand the difference it makes to the site load speed.
Considering the above site speed optimization results, it is quite clear that the site is performing better without Pippity. Since Pippity is an integral part of our case study site, removing it would compromise functionality of the site. Even with Pippity Active, site speed optimization results are not bad, but without Pippity, it is better.
|Factor||Before Optimization||After Optimization||Result||Gain|
|Page Speed Grade||75% (C)||96% (A)||Improved||28%|
|YSlow Grade||84% (B)||96% (A)||Improved||14%|
|Page load time||8.53 s||1.57 s||Improved||82%|
|Total page size||2.15 MB||692 kB||Improved||69%|
|Total number of requests||80||28||Improved||65%|
Values of Page Speed Grade, YSlow Grade, Page load time, Total page size as well as the Total number of requests improved.
|Factor||Before Optimization||After Optimization||Result||Gain|
|Page Size||5.5 MB||1.0 MB||Improved||82%|
|Load time||5.61 s||571 ms||Improved||90%|
Your website is faster than 96% of all tested websites. [Earlier: Your website is slower than 73% of all tested websites.]
Values of Page size, Load time, Requests and Perf. Grade improved.
|Factor||Before Optimization||After Optimization||Result|
|First Byte Time||A||A||Neutral|
|Cache static content||C||A||Improved|
|Effective use of CDN||No||Yes||Improved|
Values of Cache static content and Effective use of CDN improved.
Now, we have finally tuned our case study site for achieving best site speed optimization results under the given limitations. The only problem with the site is Pippity, but being an integral part of the site, it cannot be removed.
Mark de Scande optimized one more website for speed. Now, it is running at full throttle, without making any compromise to page load time and/or site load speed time. If you are running a site under similar conditions, you can implement the same site speed optimization strategy on your own site. In case of any question, please feel free to get in touch with us. We would be happy to help. Mark de Scande also offers his professional site speed optimization services through oDesk.