How ecommerce backend optimization increased revenue by 66%

Polar Sport is one of Poland's largest outdoor retail chains and has been an Alpacode client for ages. Last year we optimized the frontend and backend, which resulted in passing the Core Web Vitals Assessment and better and faster metrics. You can read about our work here.

 

However, optimization is a never-ending process. After a few months, the time comes to provide the next improvements, mostly on the backend.

 

 

Results

 

An increase of up to 66% in revenue.
6,53 % in conversion rate
64,57% in transactions
0,91% in avg order value

 

Comparision y/y before (1-31.10.2021) and after optimization 1-31.10.2022

 

Avg Page Load Time
Decrease 41%

 

 

Page Views
Increase 42,65%

 


 

Users
● All users increase 46,09%
● New users increase 46,06%
● Sessions increase 41,01%

 

 

What we have done:

 

1. Slim down and cleaning of the database (from almost 12GB to 5GB), cleaning
customer logs and unnecessary report tables. Database cleaning takes place
cyclically on an ongoing basis, which strongly prevents it from swelling. It
wasn't just a one-time cleaning.

2. Improvement of the mechanism for cleaning outdated quotes (carts data) in
the database.

3. Introduction of a mechanism for removing temp tables after reindexing.

4. Server changes, PHP upgrades, Percona and Varnish.

5. Removal of unused plugins and modules: 24 unnecessary modules have
been removed.

6. PDP profiling: disabling the wishlist and unused reports, code optimization.

7. Profiling custom listings, code optimization.

8. Blocking unwanted indexing robots and specifying the subpages they have
access to (through the robots.txt file and appropriate SEO changes).

 

Results in NewRelic, avg:

 

A decrease in web transaction time from 2000-3000ms to 1000-1500ms,
including:
● MySQL times down from 1000-1500ms to 500-600ms
● Drop in PHP times from 1000-1200ms to 600-700ms
● Increasing and maintaining Apdex score for the End User from around 0.91
to around 0.97

 

NewRelic - results in detailed pages:

 

 

Business impact: a better user experience. Before the optimization there were
issues with crashes on databases, indexes were jamming during high traffic. After
the database was cleared, these incidents stopped. Also, excessive traffic of
crawlers overloaded the website and database, after appropriate SEO changes the
problem stopped.

 

The decrease in MySQL times is especially visible, - related to the cleaning and
unloading of the database. The increase in the Apdex score for the App user also
looks good.

 

Apdex score - after optimization

 

 

Apdex Score

 

 

We have done more work but we are still waiting to implement the following:
● Removing unused photos from the media folder.
● A module for removing unused products and their url rewrites.
● JS Bundling module.
● A module for removing unused attributes and attribute options.
● A module responsible for the automatic optimization of photos uploaded by
the administration panel and generated by the Magento engine.

 

Optimization is a process. We are still working on improvements and implementing
new, modern solutions.