“Link rel=preload” is a browser hint to provide resource loading prioritization to improve the web page loading performance and speed. Preload is a value for the “rel” attribute of the HTML “link” tag to fetch a web page resource earlier than others. To improve the initial contact with users by improving the user experience, preload can be used for certain types of web page elements such as fonts, images, or CSS and JS files.
To improve the “preload” command’s efficiency for browsers, preconnect, DNS-prefetch, prerender can be used within a harmony. Using preload more than 4 different resources can cause First Input Delay increase, and Total Blocking Time since preloading everything will block the main thread of the browser for rendering. Thus, the resources for preloading should be chosen with manual web page speed tests, and the field-lab data should be collected for evaluation.
“rel” is an attribute for the HTML “link” tag that can signal the type of resources that will be fetched and the priority of the resources for downloading. Thus, using the value “preload” for the “rel” attribute of the “link” tag can be useful, but using it with “onClick” events, or different types of methodologies such as the “as” attribute can block some browsers from using the specific preloaded resource. To provide the best possible performance for the users, and clarity for the search engine crawlers, in the “preload browser hint” guideline, the methods, and working principle of “link rel=preload” will be explained.
How does Preloading Work?
The Preloading works based on the critical request chain by changing the order of fetched and downloaded resources. A key request from the critical request chain can improve the users’ interaction with the web page’s initial phase of the lifecycle. Based on the user-centric web page loading performance metrics such as “Visual Stability”, or “Loading Responsiveness”, preloading can improve the quality of user experience and quality assignment of the specific web page. Preload browser hint and “preload key requests” warning from the lab-data, or page speed test tools (page loading simulators for page speed metrics) are connected to each other. To preload, a key request, the key requests, and the critical request chain should be known.
What is a Critical Request Chain’s relation to Preloading?
Preloading and Critical Request Chain’s relation is to provide a correct resource loading order to fasten the initial contact of the user by making the web page responsible. Critical Request Chain contains the web page resources that are significant for making the web page meaningful, usable, and responsive to the user activity. Preloading these specific resources in the correct order can improve the crawl efficiency of a web page by improving the core web vitals. A critical request chain can look as below.
index.html |--style.css |--open-sans.woff2 |--nav-style.css
The example of the critical request chain above shows that the “index.html” is the “initiator” for the resources “style.css”, “open-sans.woff2”, and “nav-style.css”. Since, the “navigation styles”, and the font styles can be loaded after the “style.css”, preloading the “style.css” can help for a faster web page loading experience by prioritizing also calling the “open-sans.woff2” and “nav-style.css”. To make this critical request chain more efficient for web page loading performance with “preloading”, the “open-sans.woff2” can be used within a “link” element, and the possible FOUT and FOIT problems can be prevented by letting the web page shows its font styles earlier. An example of preloading an element from a critical request chain for the font files is below.
<link rel="preload" href="example.com/open-sans.woff2" as=woff2/font crossorigin> <link rel="woff2/font" href="example.com/open-sans.woff2">
In this example, the critical request chain has been changed by the “preload” command for the WOFF2 file. And, it can be downloaded and used even before the “style.css”. If loading the font file before the main style CSS file has benefits such as showing the styled text features as soon as possible, this method can be used. Or, the above-the-fold section of the web page can be used with inlined Critical CSS, and the font file can be requested with a preload command. In this methodology, the critical request chain would be organized better for better first contentful paint, and largest contentful paint. All these examples show the value of “preload”, and understanding the “critical request chain” effect for the web page loading experience of users. To learn more about font file loading improvement methods, you can read the related guideline.
Note: The “crossorigin” attribute for the preload request is to prevent the two sequential requests for the specific font file. Using the cross-origin attribute for preloaded font files is a must to keep a more healthy request count, and size.
The resource types that can be preloaded are listed below.
- CSS Files
- Font Files such as TTF, EOT, WOFF, WOFF2
- Track Files such as WebVTT
- Video Files such as MP4, MP3, WebM
- Images such as AVIF, WebP, JPG, and JPEG, or PNG
- Object Files such as Embedded element within <object> tag
- Embed Files within the <embed> tag
What are the Basics of Preload?
The basics of preload involve the syntax of preload and attributes related to using it. The basics of preload and related attributes are listed below.
- “href” attribute for specifying the resource that will be preloaded
- “as” attribute to specify the type of the resource
The as attribute is to provide the type of the preloaded resource. The type of information will help the browser to prioritize the resource with better precision and cache it within the browser cache. The values that can be used for the “as” attribute are “script”, “font”, “image”, “style”, “object”, “worker” and others. Example usage of the “href” and “as” attributes with the “preload” is below.
<head> <meta charset="utf-8"> <title>Link Rel Preload Guideline</title> <link rel="preload" href="style.css" as="style"> <link rel="stylesheet" href="style.css"> <link rel="preload" href="main.js" as="script"> </head> <body> <h1>Link Rel Preload Guideline</h1> <p>Preload can be used with different methods such as onload events, or click eventListeners.</p> <script src="main.js" defer></script> </body>
In the example above, the preloaded resources which are “style.css”, and “main.js” are called separately with the “script” and “link” tags after the “preload” link elements. Because, the preloaded resource can be fetched and downloaded earlier, but to use it earlier, the browser will need a second declaration by specifying its type within the attribute “as”. At the second call with the related type information of the preloaded resource, the browser won’t download the resource a second time, but it will use it directly with high priority. Thus, preloading and declaring the same resource with the related type are connected to each other for prioritizing the usage of the related resource.
What is Cache Preloaded?
Preloaded resources of a web page have a place within the HTTP Cache Hierarchy. A resource that is cached with the “preload” will be kept in the “session-based” cache of the browser. When the session is stopped, the preload cache will be purged. Since the preloaded resources are a high priority, most of the time they are used also for other web pages of the same website. In this context, a preloaded resource can be cached to eliminate extra requests, downloading and compiling for the next web pages. To use the “preload cache”, the “as” attribute should be used with the correct format. If the “as” attribute is omitted, or if the value for the “as” attribute is wrong, browser will handle the preloaded request as an XHR request. And, the browser won’t be able to prioritize the preloaded resource correctly. In the same context, the “crossorigin” attribute should be used with preload browser command for the font files. If the “crossorigin” is not used, the specific font file will be requested two times. To cache the font files in a correct format with preload cache, the “crossorigin” attribute will provide the correct insights for the browser.
To understand the HTTP Cache Hierarchy along with cache preloaded, the guidelines below are related.
- Last Modified Response Header
- Expires Response Header
- eTag Response Header
- Cache-control Response Header
Understanding the effects of “onLoad Event” for a web page, and its connection to the Rendering Tree requires knowing the things below.
What is the Relation between Subresource and Preload?
“link rel=subresource” is an old resource loading hint from Google Chrome that does the same thing with preload. “Subresource” resource loading hint (browser hint) didn’t have the attribute “as”, thus prioritizing was not perfect for the resources. Link prefetching with subresource didn’t have a priority hierarchy, anything that has been put into the link element with “subresource” would be requested and used immediately. On the other hand, the preload can provide build custom resource loading and execution unlike subresource by letting the browser download a resource earlier, and use it when the time comes. Below, you will see the replacement announce from Google Chrome for “subresource”.
Below, you can see a request from Google Chrome’s Google Group which is “blink-dev”.
Below, you will see another image that explains the removal of subresource by saying it is problematic and buggy, unlike Preloading API.
What is the relation of Modulepreload and Preload?
<head> <link rel="modulepreload" href="module-1.mjs"> <link rel="modulepreload" href="module-2.mjs"> <link rel="modulepreload" href="nodule-3.mjs"> <link rel="modulepreload" href="module-4.mjs"> </head>
To learn more about “modulepreload”, you can read the related guideline.
import(_/* webpackPreload: true */_ "CriticalChunk")
What are the other Browser Hints besides Preload?
The other browser hints for better page loading performance besides preload are listed below.
- Preconnect: to perform DNS-Resolution, TLS Handshake, and TCP Connection.
- DNS-Prefetch: to perform DNS-Resolution.
- Prerender: To prerender a web page at the background after loading the current web page.
- Prefetch: Pre-fetching a web page while the user stays on a web page.
Which User-centric Metrics can be improved with Preload?
The user-centric metrics below can be improved with preload value for link elements.
- Perceived Load Speed: Since, it can improve the Largest Contentful Paint, and First Contentful Paint, the perceived load speed can be better.
- Load Responsiveness: Preloaded resources can provide a better main thread organization, it can help a browser to improve the iddle time for user interactivity with web page.
- Runtime Responsiveness: Preloading key request can help a browser to give a feedback to the users’ inputs. If a user tries to click while a browser loads and renders resources by blocking the main thread of the browser, it won’t be possible. But, preloading the key requests can help a browser to delay the some heavy non-content relevant resources to give a user chance to perform an input event such as clicking.
To learn more about user-centric page speed metrics, you can read the related guideline.
Which Core Web Vitals Metrics can be improved with Preload?
The Core Web Vital metrics that can be improved with preload resource hints for key requests can be found below.
- Largest Contentful Paint: Prioritizing the loading of the LCP Element can improve the LCP Score, since the related resource will be loaded faster. LCP element prioritization is related to the Percived Load Speed.
- First Input Delay: Prioritizing the relevant resources for interactivity with user will help users to perform an input event earlier. Prioritization of the iddle time of CPU of browser with “link rel=preload” is related to the Load Responsiveness.
Preloading key requests and Cumulative Layout Shifts are not connected to each other, since preloading key requests from the critical request chain is about prioritization of the requests, not the visual stability of the web page components.
What are the relevant concepts for browser hints like Preload?
To understand the working principle of the “preload” browser hint at the working principles of the things below should be understood.
A web browser sends the resources to the web browser when a user agent requests them from the webserver. A website and a web page can be loaded via a web browser with different resource loading prioritization. A web browser might not support a web browser loading hint or feature, thus the web browser’s version and the type are important. Preload web browser resource prioritization hint is relevant to the nature of browser, server, website, webpage, and user-agent concepts’ relations to each other.
Preload is a resource loading hint (browser hint) to provide insights to a browser for loading a web page accurately for a better user experience. “link rel=preload” for key requests from critical request chain, and its attributes such as “as”, or “crossorigin” can help an SEO and Developer to improve the crawl ability, quality score, and usability of the web page for users, and search engine crawlers. Using, preload resource load hint with a wrong methodology, or with a less effective method can cause crawling efficiency problems, along with preventing search engines’ to see a web page’s styled version. To improve the rendering cost, and user-centric page speed metrics, along with the crawlability, using resource load hints such as preload is a must for an SEO Project. Resource Load Order Optimization should be a stage for the Technical SEO and Web Development intersection. And, Holistic SEOs who want to improve their skills for every vertical of SEO should be aware of usage tips for preload, and the nature of browsers for rendering a web page, for instance, critical request change, DOM-tree, CSSOM, or crossorigin attribute.
The “link rel=preload” guideline will be updated in light of new information.
- Importance of Keyword Search Volume for SEO - June 29, 2022
- Keyword Difficulty: Definition, Examples, Usage, and Importance for SEO - June 19, 2022
- Google Author Rank: How Google Knows which Content Belongs to Which Author? - May 31, 2022