Google had already announced way back in 2010 that website speed would be considered as a ranking factor. However, the company has not specified just how it calculates website speed and hence a test was conducted to find out actual factors that may have been considered by Google to determine search rankings.
Since Google does place a lot of importance to good content, the company also wants users to get the best possible user experience and a slow website will not be able to please visitors. However, there are only a few metrics that can help correlate between specific speed metrics to search ranking. The test involved extracting top 50 search result URLs based on ranking and evaluating around 100,000 pages by using single and long tail search terms.
WebPageTest, an open source tool was used along with the Chrome browser to test all pages by capturing 40 page metrics for every examined URL. The results were based on several factors as mentioned below.
One factor that was tested via this test was actually split into two. The Page Load Time of the website was split into two factors with the first being time taken for data to be loaded even as you begin interacting with the displayed data by clicking on it. The second factor was complete loading time in which all elements of that page were loaded, which included advertisements, images, etc. The test found no correlation between both the above metrics.
The test was further expanded to calculate the time taken for the very first byte to get received from a server by your browser. This metric calculated the time taken by the web server to process a request and deliver a response and the time taken by that first byte to travel from the server to your browser. This TTFB or Time To First Byte revealed a correlation between higher TTFB and lower search rankings. This metric was quite strong implying that Google did consider it as an important one for ranking purposes.
Another factor considered for the test was the size of the page. The test revealed a surprising correlation between page size and page rank that resulted in decreased page rankings due to decreased page size. This indicated that smaller firms that had less resources to create large pages with more content could end up with lower rankings as compared to larger companies that presented larger pages that were further optimized for improved rankings.
Surprisingly, the test also did not reveal a correlation between images on a page and ranking even though more images translate into increased loading times. This factor backs the page load time results that indicate no correlation between rankings and page load times.
The test results may not reveal a correlation between page load times and search rankings, but the Time To First Byte test does reveal that improved rankings can be achieved with better TTFB timings. This basically means that Google gives more importance to website performance related to the back-end rather than the front-end.
One possible reason is that Google may have found it easier and faster to measure TTFB metrics rather than considering metrics that were also dependent on the browser and web page structure. TTFB timings can be affected by load on the web server, network latency between the server and visitor, and the website’s back-end ability to quickly generate the content. Websites can hence utilize CDNs or Content Distribution Networks to lower the network latency.
The conclusion of this test reveals that websites with strong back-end infrastructure and ability to deliver content faster through the back-end can achieve higher rankings. You should focus on utilizing CDNs, optimizing the database queries and application code, and making sure that your web servers are quick and responsive.
You can also use tools such as WebPageTest to measure your TTFB and those of your competitors to make suitable adjustments. While front-end performance may not be directly related to rankings, it is still a very important factor to ensure a satisfying user experience. Visitors will bounce away quickly in case your page load times are slow and happy visitors are more likely to share content, make purchases, and come back for repeated visits. All these factors will actually improve search rankings.
As the above tests reveal, you do need to focus on both your front-end and back-end site performance to ensure that Google delivers better search result rankings for your site.