WordPress performance work does not always produce an immediate improvement in Largest Contentful Paint. In this case, the first optimization phase focused on reducing image weight with Imagify and WebP delivery, limiting reCAPTCHA to pages that actually needed it, and auditing autoloaded database options. The changes lowered Total Blocking Time, reduced unnecessary page weight, and made the site more responsive, while render-blocking CSS remained the main opportunity for improving LCP in a later phase.
Issue background
A WooCommerce site needed an initial round of speed improvements. The work was deliberately scoped to three broadly useful areas:
- Compressing and modernizing images
- Preventing reCAPTCHA from loading site-wide
- Checking the WordPress database for oversized autoloaded options
The site used both WooCommerce reCAPTCHA and Gravity Forms-related verification. Loading those assets on pages without a checkout or form added JavaScript, network requests, and processing that visitors did not need.
The goal of this first phase was to improve responsiveness and reduce transfer size without making high-risk changes to the theme’s rendering pipeline.
Diagnosis
The performance review identified several practical opportunities.
First, the Media Library contained images that could benefit from compression and next-generation delivery. Large image files increase bandwidth usage and are especially noticeable on slower mobile connections and image-heavy product pages.
Second, reCAPTCHA assets were loading more broadly than required. Scoping the integration was expected to remove approximately 727 KB from pages that did not need the verification scripts.
Third, the database needed an autoload audit. WordPress loads options marked for autoload on every request, so one plugin storing a large amount of data there can slow down the entire site.
The autoload review did not find any unusually heavy entries that required cleanup. This was an important result because it ruled out the options table as the main performance bottleneck.
After the first-phase changes, a lab test still showed Largest Contentful Paint at approximately 4.8 seconds. However, Total Blocking Time improved and moved into the green range, indicating that the page became responsive to clicks and taps more quickly.
Resolution steps
- Back up the site before bulk optimization. Create a current restore point before processing the Media Library or changing performance-related settings.
- Install and configure Imagify. Activate Imagify on production and review its compression and next-generation image-delivery settings.
- Start bulk image optimization. Process existing Media Library images rather than optimizing only future uploads.
- Exclude unsupported files. Add an exclusion for PDFs or other file types that should not pass through the image-compression process.
- Enable WebP or next-generation delivery. Confirm that supported browsers receive the optimized format and that original files remain available as fallbacks.
- Inspect important image-heavy pages. Test the homepage, product detail pages, category pages, and any templates with large hero images.
- Scope WooCommerce reCAPTCHA. Configure the integration so its scripts load on checkout or other protected WooCommerce screens rather than across the entire site.
- Scope Gravity Forms verification assets. Prevent reCAPTCHA-related scripts from loading on pages that do not contain a Gravity Form.
- Verify protected forms after scoping. Test checkout and every affected Gravity Form to confirm that validation and bot protection still work.
- Audit autoloaded options. Query the options table for the largest autoloaded records and map those records back to their plugins or themes.
- Change autoload behavior only when justified. Do not disable entries simply because they appear in the query. Confirm that an option is both large and unnecessary on most requests before changing it.
- Compare more than one performance metric. Review Total Blocking Time, transfer size, responsiveness, and field data in addition to LCP.
- Monitor the changes over time. Real-world performance data can take longer to reflect improvements than a single lab test.
- Plan the next optimization phase. When LCP remains high, audit render-blocking CSS and JavaScript, remove unused assets, and prioritize only the code required for above-the-fold content.
Do not evaluate an optimization project using one Lighthouse number alone. Image weight, interactivity, mobile bandwidth, and field data can improve even when a single LCP test remains similar.
Final outcome
Imagify was installed and its bulk optimization process was completed with PDFs excluded. WebP delivery reduced image transfer weight across the site, providing the greatest benefit on image-heavy pages and slower mobile connections.
reCAPTCHA was limited to the pages and forms that required it. This reduced unnecessary frontend work and improved Total Blocking Time, which moved into the green range.
The autoloaded-options audit found no unusually heavy data requiring cleanup. The first-phase changes remained stable during the monitoring period.
Largest Contentful Paint still measured approximately 4.8 seconds in a later lab test, so the next major opportunity was identified as render-blocking CSS and JavaScript. That work would involve removing unused assets and loading only what the visible portion of the page needed.
The key lesson is that WordPress performance improvements should be evaluated across several layers. Image optimization, selective script loading, and database audits can make a site lighter and more responsive even before deeper rendering work moves the headline LCP score.
For help optimizing WordPress images, reducing reCAPTCHA overhead, auditing autoloaded options, or improving Core Web Vitals, contact Freshy.
