Engineering2 min read
The AEM Performance Audit Checklist (Before You Blame the CDN)
akenside.ai team·
Most AEM slowness is self-inflicted
When an AEM site feels slow, the CDN gets blamed first, but the real culprits are usually upstream: uncached dispatcher responses, heavy client libraries, and components doing too much work per request. Here's the checklist we run before recommending a single infrastructure change.
1. Dispatcher & caching
- Are HTML responses being cached, or is something (a cookie, a query param, a
Cache-Controlheader) busting the cache on every hit? - Is the dispatcher cache invalidation surgical, or does every publish flush everything?
- Are static assets fingerprinted and far-future cached?
2. Client-side weight
- Total client-library payload, how much JS/CSS ships on first load, and how much is used?
- Are libraries concatenated and minified, or shipping as dozens of requests?
- Lazy-load below-the-fold components and images.
3. Server-side rendering cost
- Which components are expensive to render? (Sling Models doing synchronous service calls, deep resource resolution, uncached queries.)
- Are JCR queries indexed, or are they traversing?
4. Core Web Vitals that matter
- LCP, usually a hero image or web font. Preload the LCP image; subset and preload fonts.
- CLS, reserve space for images/embeds; avoid late-injected banners.
- INP, keep main-thread work small; defer non-critical JS.
5. The measurement loop
Fix nothing until you can measure it. Capture real-user metrics (field data), not just lab scores, and re-measure after each change so you know what helped.
Want this done for you?
Our AEM Health & QA Audit runs this checklist against your site and returns a prioritised remediation backlog, performance, dispatcher config, accessibility and content-model smells, so your team fixes the highest-impact issues first.