A Study of 150,000 Websites’ Speed and Analyze How to Improve PageSpeed?
It is very easy to state that site speed influences your search engine optimization. But do you know why and how does website page speed influence SEO? An experiment was conducted to find out the ignored speed factors or website page loading time that have a direct impact on Google rankings. Almost 150,000 URLs were analyzed. This kind of study helps a lot to reveal a specific speed factor that is used by Google to rank a website. Speed is not the only influence, rather a particular type of page speed is what marks the difference.
The aim of the study
Our readers may be curious to know what is the reason of taking all these troubles to conduct this massive study. Well, there are many reasons behind this:
- The first reason behind this study was curiosity. We were eager to know how the page speed influences the rankings and what can be done to better the ranking.
- We had no major study on the speed of the page and Google rankings. The last massive study was conducted 2-3 years ago. Hence, we wanted to come up with an up to date analysis of data on page speed.
- During the last two years, we have seen Google changing their algorithm several times and emphasizing page speed again and again. So, we felt that it is important to know how the changing algorithm influences the speed and the ranking.
- We wanted to find out how user experience and page speed influence each other.
How we proceed
A faster website is suitable for SEO and it also helps to increase the conversion. But everyone related to this field knows that Pagespeed metrics and Google’s PageSpeed Insights are not able to provide a timely result. Even you cannot bank on a Pingdom test to furnish reliable data on page speed.
There are many other factors regarding the speed of the page. These factors include the impact of cache-validating, caching, code minification, DNS lookup, encoded headers, image size, parallelized downloads, plugins, render rates, redirects, server response, static content, and time to first byte (TTFB). So, the matter is not so simple, rather it is a complicated one. This study was required to acquire more insights.
We needed to gather, analyze, refine, parse, configure and finally digest all the gathered data resourced from the 150,000 URLs. For each URL, we observed six metrics which are integral parts of page speed. These six metrics are:
1. Time to First Byte
When your browser is going to a URL, the server receives a message to ask for the HTML document at that URL. It helps to calculate the time for the first little bit of website data that is sent to the browser you are using. When the time is fast, it indicates that the rest of the site will also be loaded fast.
2. Start Render
Rendering takes place when your computer converts the code into the visual display you see on the screen. When the first visual part of your website appears, it is called start render. This is helpful for the users as it conveys them that the website is going through something.
3. Visually Complete
When the rendering is completed, it can be called visually complete and in this phase, the user can visit the full website.
4. Document Complete
Even when the site is completely visible, some operations still go on in the background. Once these operations are finished, the server signals that the loading of HTML is completed. This is perhaps the most technically apt measure of site loading.
5. Fully loaded
When the document is done, the code starts running and loads more items. But it cannot stop the user from making interaction with the site. Hence, it cannot be considered as a part of the load time. When all the loading activities stop for 2 seconds, then this metric should be measured.
6. A number of file requests
When you are seeing a site to be loaded, it is requesting a number of files like CSS, JS, or image files. Loading multiple files at the same time can slow the loading time. So, naturally, they should always be minimized.
The process in detail
We used the free tool webpagetest to examine the performance of the websites. And the result looks like:
|First byte||Start render||Visual Complete||Doc complete||Fully loaded|
|0.85 second||2.22 second||10.43 second||2.11 second||4.12 second|
We used application program interface to crawl each of the URLs to conduct extensive speed testing and analysis. As we had to crunch a lot of URLS, it took more than 150 AWS EC2 servers. And the entire matter was completed in two and half days. Each of our testing agents was running Chrome on 1024* 668 windows desktop.
In order to determine the crawling of the URLs, we gathered almost 6000 random keywords with a search volume on monthly basis (12 per month) to ensure the fact that we were correctly using a representative sample of all types of keywords, both long tails, and head terms. We recorded the top 20 Google results (non-mobile) for each of the keywords.
The result of the experiment
As an outcome, we found that faster site speed is related to higher Google ranking. We found the top ranked websites having high speeds in the start rendering metric. The result indicates that, generally, the overall load time is faster for the top six positions. The website, at seventh position, was found to be almost 20 percent slower than the top ranked URL. This implies that the improving start render speed directly influences the higher Google result.
We also studied the results of other researches and found that the top-ranked websites have comparatively fast speeds for TTFB or time to the first byte. TTFB helps to better the user experience. The importance of TTFB will be even clearer to you if you read Google’s company motto. 10 principles were mentioned here in the initial stages.
Among these 10 principles, the most important ones in this context are:
- Focus should always be given to the user and the rest will follow
- Fast is always better than the slow
Both of the points emphasize the fact that TTFB is an important user factor. When Google provides a higher ranking to a site with high TTFB speed, the ranking indicates that Google is pointing the site for having an excellent user experience.
Now, let’s come to the discussion of other correlative significance of comparatively less important site speed metrics such as document complete and start render. They should be mentioned after TTFB on the basis of their importance in the page speed.
How you can improve TTFB
There is a series of articles that discuss how you can improve the speed of your site. Most of these articles are generalized. But, here we will focus on some specific issues related to the speed of a web page. Let’s get started with checking the present TTFB of your website. How to do this? Just go to www.webpagetest.org and type your website URL in the specific field. Next, click on the “start test” option. Now, you need to adjust the test location in order to find out the server that is geographically nearer to the place where most of the users live. If you know that your users use another browser than chrome, then you can change the browser setting as well.
You will find the test finished within one minute. Now, have a look at the first byte number given in the result. This first byte number is the TTFB of your website. If you need a detailed analysis, you can click on the “first view” chart. This is the process of checking your TTFB. Now, how can you understand whether your TTFB needs improvements or not?
A table is given here to make you understand when your TTFB is good and when you need to work on it.
|Speed (in seconds)||Comments|
|0.3-0.5||Good. Minor changes can do well|
|0.6-0.9||Average. Improvements are required|
|1.6 or more than it||Very poor|
I hope you have pointed out the current status of your TTFB with the help of the guidance provided here. Now, employ these six tips to improve your site’s TTFB:
1. Start using a CDN
2. Application code should be optimized
3. Database queries should be optimized
4. Try to reduce HTTP requests
5. Must ensure fast server response time
6. The use of an RFPL (response first, process later) will do good for caching
After discussing the entire study, some factors should be summed up once again to highlight them as per their importance. First, never forget that speed is important but it is not the only factor working behind the ranking of a website. Secondly, you should keep in your mind that Google always prefers the sites that have an ideal UX (user experience). The third important thing to be noted is, of course, TTFB. You should optimize your TTFB in the best possible way.
If you find the study helpful and want help to improve your website speed for SEO purposes, please feel free to consult B3NET, a leading US-based digital marketing agency.