Web Optimization techniques: Enhancing Website Load Performance through Combined Caching, Compression, CDN Routing, and Lazy-Loading Strategies

Abstract

Website speed is one of the strongest predictors of user retention, conversion, and search visibility, yet a large share of production websites still take more than four seconds to become interactive on an average mobile connection. This paper investigates how caching, compression, content-delivery-network (CDN) routing, code minification, and lazy loading can be organized into a single adaptive pipeline rather than applied as isolated, one-off fixes. We propose the Adaptive Multi-Layer Web Optimization (AMWO) framework, which selects and sequences optimization modules according to the type of website and the composition of its assets, and evaluate it on forty publicly accessible websites spanning five categories: news, e-commerce, blogging, portfolio, and SaaS dashboards. Baseline and optimized versions were benchmarked with Google Lighthouse and WebPageTest under matched, throttled network conditions. The proposed framework lowered average page load time by roughly 46 percent and raised the mean Lighthouse performance score from 54 to 86, while Cumulative Layout Shift and Total Blocking Time both fell substantially across every category tested. These results indicate that layering optimization techniques according to page type, instead of applying a single generic technique uniformly, produces measurably better outcomes than the single-technique approaches reported in most of the prior literature reviewed here.

Introduction

The proposed research focuses on improving website performance by reducing page loading time, improving user experience, and optimizing resource usage. As websites become heavier due to JavaScript frameworks, images, videos, and third-party scripts, slow loading negatively affects user satisfaction, business performance, and search engine rankings.

The research proposes an Adaptive Multi-Layer Web Optimization (AMWO) Framework, which automatically selects and combines suitable website optimization techniques based on the type and requirements of each website.

1. Importance of Web Performance

Website performance is important for three main reasons:

• User experience: Faster websites improve user satisfaction and reduce waiting time.

• Search engine ranking: Google evaluates website performance using Core Web Vitals: LCP, INP, and CLS.

• Business performance: Faster websites can reduce bounce rates, improve sales, minimize cart abandonment, and lower hosting costs.

2. Challenges in Modern Web Applications

Modern websites face several performance challenges:

1. Large JavaScript files increase downloading, processing, and execution time.

2. Third-party scripts for advertisements, analytics, and support tools increase page loading time.

3. Slow or unstable internet connections affect website performance.

4. Unoptimized images and videos increase page size.

5. Existing optimization tools identify individual problems but do not always recommend the best combination of solutions.

3. Related Works

Previous research has explored different website optimization techniques:

• Caching Mechanisms: Store frequently accessed data to reduce repeated processing and server response time.

• Content Delivery Networks (CDNs): Deliver website content from servers closer to users, reducing network latency.

• Minification and Code Splitting: Reduce JavaScript and CSS file sizes and load only the necessary code.

• Image and Media Optimization: Use WebP, AVIF, responsive images, and compression to reduce file sizes.

• Lazy Loading: Load images and videos only when they are required.

• HTTP/2 and HTTP/3: Improve communication speed and reduce connection overhead.

• Server-Side Optimization: Improve database queries, server configuration, and caching to reduce response time.

Previous studies show that combining multiple optimization techniques can produce better results than applying individual techniques independently.

4. Proposed Methodology

The research introduces the Adaptive Multi-Layer Web Optimization (AMWO) Framework to improve website performance through intelligent selection of optimization techniques.

The methodology consists of three main stages:

A. Data Collection

A dataset of 40 publicly accessible websites is selected, divided into five categories:

• News websites.

• E-commerce websites.

• Blogging websites.

• Portfolio websites.

• SaaS dashboards.

Each category contains eight websites.

B. Data Preprocessing

Each website is tested five times using Google Lighthouse and WebPageTest. The average results are calculated to reduce measurement variations and improve reliability.

C. Proposed AMWO Framework

The framework analyzes website characteristics, including image size, JavaScript size, third-party requests, and framework usage.

Based on this analysis, it selects appropriate optimization techniques, such as:

• Caching.

• Compression.

• CDN routing.

• Minification.

• Lazy loading.

The framework avoids applying unnecessary optimizations. For example, important images visible immediately when a page opens are not unnecessarily lazy-loaded because doing so could increase Largest Contentful Paint (LCP).

5. Proposed Algorithm

The proposed algorithm follows these steps:

1. Collect website URLs and asset information.

2. Measure baseline website performance.

3. Classify websites according to their categories.

4. Analyze images, scripts, third-party requests, and other assets.

5. Select suitable optimization techniques.

6. Apply the selected optimization modules.

7. Measure the optimized website performance.

8. Compare the original and optimized results.

9. Generate an optimization report.

Performance is evaluated using metrics such as:

• Load Time.

• Time to First Byte (TTFB).

• Largest Contentful Paint (LCP).

• Total Blocking Time (TBT).

• Cumulative Layout Shift (CLS).

• Speed Index.

Conclusion

This paper set out to test whether combining web-performance techniques adaptively, rather than applying a fixed technique uniformly across every page, produces measurably better results. Using the proposed Adaptive Multi-Layer Web Optimization (AMWO) framework on a 40-site test set spanning five categories, we observed an average page load time reduction of about 46 percent and a mean Lighthouse performance score increase from 54 to 86, alongside meaningful drops in Total Blocking Time and Cumulative Layout Shift. These findings support the view that module selection should be driven by a site\'s asset profile rather than treated as a one-size-fits-all checklist, which is consistent with, and extends, the interdependence-coefficient idea proposed in prior work on combined optimization strategies [8]. Future work should address the limitations noted in Section 4.5 by validating the framework against field data from the Chrome User Experience Report rather than lab-only measurements, and by testing additional site categories such as media-streaming and government portals. A further direction is to replace the current rule- based Adaptive Optimization Engine with a learned model that predicts the best module combination directly from a site\'s asset manifest, which could allow the framework to improve automatically as it is applied to more sites. Finally, as HTTP/3 and edge-computing adoption continues to grow, it will be worth re-testing whether techniques that mattered most under HTTP/1.1 and HTTP/2, such as manual request bundling, still provide any benefit at all, or whether they can be safely retired from optimization checklists.

References

[1] Souders, S. High Performance Web Sites: Essential Knowledge for Front-End Engineers. O\'Reilly Media, 2007. [2] Grigorik, I. High Performance Browser Networking. O\'Reilly Media, 2013. [3] Google & Deloitte. “Milliseconds Make Millions: The Business Impact of Mobile Site Speed.” Think with Google, 2020. [4] Google Web Fundamentals Team. “Web Vitals: Essential Metrics for a Healthy Site.” web.dev, 2020. [5] HTTP Archive. “The Web Almanac: Page Weight and Performance Chapter.” httparchive.org, 2024. [6] Tripathi, B., & Umar, M. “Experimental Analysis of Server-Side Caching for Web Performance.” arXiv preprint, 2025. [7] “Caching and Content Delivery Networks (CDNs) for Performance Optimization.” ResearchGate, 2025. [8] “Overview of Web Application Performance Optimization Techniques.” ResearchGate, 2024. [9] “Performance Optimization Techniques.” ResearchGate, 2024. [10] “A View on Edge Caching Applications.” arXiv preprint arXiv:1907.12359, 2019. [11] “A Distributed Real-Time Transcoding CDN System Design and Performance Simulation Study.” PMC, 2022. [12] “Optimizing Digital Experiences with Content Delivery Networks.” arXiv preprint arXiv:2501.06428, 2025. [13] Salesforce Trailhead. “Optimal CDN & Caching Strategies for E-commerce.” trailhead.salesforce.com, 2024. [14] SitePoint. “5 Techniques for Lazy Loading Images to Boost Website Performance.” sitepoint.com, 2024. [15] HackerNoon. “Web Image Optimization: Best Practices for Speed and SEO.” hackernoon.com, 2024. [16] Request Metrics. “How to Optimize Website Images: The Complete Guide.” requestmetrics.com, 2026. [17] FooPlugins. “Boost Your Website Speed With Lazy Loading: A Beginner\'s Guide.” fooplugins.com, 2026. [18] Website Speedy. “Lazy Loading vs. Preloading: Which Improves Website Speed?” websitespeedy.com, 2026. [19] Belshe, M., Peon, R., & Thomson, M. “Hypertext Transfer Protocol Version 2 (HTTP/2).” RFC 7540, IETF, 2015. [20] Iyengar, J., & Thomson, M. “QUIC: A UDP-Based Multiplexed and Secure Transport.” RFC 9000, IETF, 2021. [21] Google Chrome Developers. “Lighthouse Performance Scoring.” developer.chrome.com, 2023. [22] WebPageTest. “WebPageTest Documentation.” webpagetest.org, 2024. [23] Mozilla Developer Network. “Populating the Page: How Browsers Work — Critical Rendering Path.” developer.mozilla.org, 2023. [24] Nielsen, J. “Website Response Times.” Nielsen Norman Group, 2010. [25] Akamai Technologies. “State of the Internet: Performance and Security Report.” 2023.

Copyright

Copyright © 2026 Mohit Ameta, Kapil Kumar Sharma, Dr. Vishal Shrivastava, Dr. Akhil Pandey. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.