Speed optimization case studies
Before and after speed results for 24 WordPress sites we optimized, with the Google PageSpeed, GTmetrix and Pingdom screenshots for each.
Before and after speed results for 24 WordPress sites Codiffy optimized between December 2019 and September 2022. Every test is of the site’s home page, with Google PageSpeed Insights (a score from 0 to 100), GTmetrix (a grade with its performance percentage) and Pingdom, and each case study shows the screenshots side by side.
The four biggest improvements
Ranked by the rise in the Google PageSpeed mobile score.
- reflectionswatergardens.com, March 202113 → 100
Google PageSpeed mobile. Desktop 48 to 100, GTmetrix E (51%) to A (100%).
- whitewolfrafting.com, April 202113 → 99
Google PageSpeed mobile. Desktop 48 to 95, GTmetrix D (60%) to A (97%).
- lifesecure.com, May 202114 → 98
Google PageSpeed mobile. Desktop 47 to 100, GTmetrix E (41%) to B (81%).
- randyclark.com, September 202113 → 97
Google PageSpeed mobile. Desktop 71 to 100, GTmetrix B (82%) to A (100%).
All 24 sites
“Fully loaded” is the GTmetrix fully loaded time. The three sites tested before late 2020, on GTmetrix’s old scale, are in their own table below.
| Site | Business | When | PageSpeed mobile | PageSpeed desktop | GTmetrix | Fully loaded |
|---|---|---|---|---|---|---|
| thatscoffee.com | Online coffee store | November 2020 | 22 to 85 | 55 to 100 | D (61%) to A (100%) | 6.1 s to 3.4 s |
| 1stonthelist.ca | SEO agency | November 2020 | 48 to 84 | 85 to 98 | B (81%) to A (100%) | 3.6 s to 2.1 s |
| snowyowltours.com | Dog sled tour company | December 2020 | 12 to 68 | 42 to 95 | F (6%) to C (63%) | |
| liftylife.ca | Vacation rentals business | January 2021 | 9 to 92 | 38 to 100 | F (41%) to A (99%) | 8.2 s to 2.0 s |
| keezingkreations.com | Jewellery business | January 2021 | 27 to 94 | 73 to 100 | C (70%) to A (96%) | 6.5 s to 1.8 s |
| prigraphics.com | Commercial printer | February 2021 | 30 to 90 | 62 to 100 | F (25%) to A (96%) | 13.5 s to 1.8 s |
| reflectionswatergardens.com | Water gardens and pond builder | March 2021 | 13 to 100 | 48 to 100 | E (51%) to A (100%) | 10.6 s to 0.9 s |
| arcadiasourcing.net | Manufacturing sourcing company | March 2021 | 49 to 95 | 79 to 100 | D (60%) to A (100%) | 6.6 s to 1.8 s |
| whitewolfrafting.com | Whitewater rafting company | April 2021 | 13 to 99 | 48 to 95 | D (60%) to A (97%) | |
| lifesecure.com | Emergency kits store | May 2021 | 14 to 98 | 47 to 100 | E (41%) to B (81%) | 10.3 s to 2.1 s |
| neheliskiing.com | Heli-skiing operator | June 2021 | 19 to 96 | 65 to 100 | D (57%) to A (97%) | 9.1 s to 1.5 s |
| miragescreensystems.com | Retractable screens company | June 2021 | 16 to 80 | 59 to 95 | B (75%) to A (95%) | 3.4 s to 2.5 s |
| premium-electric.ca | Electrician | July 2021 | 38 to 100 | 84 to 100 | D (58%) to A (100%) | 4.0 s to 1.9 s |
| enengineering.com | Engineering firm | August 2021 | 42 to 95 | 86 to 100 | C (72%) before, no after image | |
| flatoutauto.com | Auto accessories business | August 2021 | 21 to 95 | 62 to 100 | D (60%) to A (95%) | |
| bestautopartsonline.com | Auto parts store | August 2021 | 32 to 84 | 72 to 100 | D (60%) to A (98%) | 8.1 s to 1.3 s |
| randyclark.com | Hair replacement studio | September 2021 | 13 to 97 | 71 to 100 | B (82%) to A (100%) | |
| summer.neheliskiing.com | Summer lodge site of Northern Escape Heli Skiing | May 2022 | 38 to 96 | 65 to 100 | D (65%) to A (100%) | 8.8 s to 2.4 s |
| unionvilleoptometry.ca | Optometry clinic in Markham, Ontario | July 2022 | 50 to 92 | 83 to 99 | D (57%) to A (100%) | 11.0 s to 1.9 s |
| killerproductions.net | Video production company in Dallas | July 2022 | 26 to 80 | 72 to 99 | C (80%) to A (100%) | |
| uniparts.net | Industrial electronics repair company | September 2022 | after: 97 | after: 100 | after: A (100%) | after: 2.4 s |
Earlier tests on the old GTmetrix scale
Until late 2020 GTmetrix scored pages with PageSpeed and YSlow percentages instead of today’s grade, so these results are not comparable with the table above. Google PageSpeed was not tested on these sites.
| Site | Business | When | GTmetrix PageSpeed score | Fully loaded |
|---|---|---|---|---|
| keswickkidsdental.ca | Dental clinic | December 2019 to January 2020 | 40% on the old host, 57% after the hosting move to 95% after optimization | 11.7 s to 3.9 s |
| countrydental.ca | Dental group | December 2019 to January 2020 | 61% on the old host, 87% after the hosting move to 94% after optimization | 5.3 s to 1.7 s |
| choquer.store | Small WooCommerce store | January 2020 | 83% to 97% | 6.9 s to 2.9 s |
How we make a site faster
Most of these jobs, from late 2020 onward, were NitroPack installs with manual tuning: plugin conflicts removed, old cache and image plugins switched off, and exclusions set so chat widgets, sliders and forms kept working. The three earliest jobs, from December 2019 to January 2020, were optimized by hand without a speed service: scripts minified and deferred, images scaled and compressed, caching, and fewer requests.
About the numbers
-
As tested Every figure is quoted from the test screenshot taken before and after the work, on the date shown.
-
Initial page weight Very low after figures for page size and requests reflect lazy loading on first load: they are the initial page weight, not everything the page can load.
-
One side only Uniparts has no test from before the work, and the GTmetrix test of EN Engineering after the work has no image, so its result is shown with PageSpeed and Pingdom.
Want this for your site? Read how our performance optimization service works, or book a free call.
Questions.
Which tools do you measure speed with?
Google PageSpeed Insights, GTmetrix and Pingdom, and on one site WebPageTest as well. We test the home page before and after the work and keep the screenshots.
What does a speed optimization job involve?
On most of these sites, a NitroPack install with manual tuning: plugin conflicts removed, old cache and image plugins switched off, and exclusions set so chat widgets, sliders and forms kept working. The earliest jobs were optimized by hand: scripts minified and deferred, images scaled and compressed, caching and fewer requests.
Why are some page sizes so small after the work?
Very low after figures for page size and requests reflect lazy loading on first load. They are the initial page weight: the page loads the rest as the visitor needs it.
Tell us your goal. We'll show you the plan.
Book a free 30-minute call. You'll leave with a clear plan and a fixed quote, whether you hire us or not.
Book my free call
