Loading..
Processing... Please wait...

Product was successfully added to your shopping cart.



Does site speed really matter to SEO?

Yes site speed helps SEO.

Here is an official proof by google published on Google Webmaster Central Blog:

..You may have heard that here at Google we're obsessed with speed, in our products and on the web. As part of that effort, today we're including a new signal in our search ranking algorithms: site speed. Site speed reflects how quickly a website responds to web requests....

Now question is - how google measures site speed? What should we work on to let the search giant know our site is fast?

I will answer this question in a bit (I don't want to wait - tell me now!) but first lets see what site speed measures exist.

 

1. How to measure speed

I am using webpagetest.org to measure site speed. It displays various metrics:

  • Load Time
  • Fully Loaded Time
  • Time to first byte
  • Start render time
Webpagetest.org speed metrics Webpagetest.org speed metrics

In 20 steps to increase site performance I talk about every one of them.

 

2. How google measures site speed - real life experiment

Research by Billy Hoffman published in 2013 shows us that google (probably) use time to first byte (TTFB) as ranking factor. Fully loaded time doesn't really matter to google! (But it still matters to conversion rate).

Does it still hold? I wanted to find out by sort of repeat Billy's experiment.

What I did:

  1. Setup a private webpagetest instance on AWS - to gather speed metrics. It costed around $20/mo - here is a forum thread on how to make it work.
  2. Found a Aol 2006 search data csv (Don't ask me where I found it - I won't say:)) and loaded it into mySQL - to have a real time search phrases people used.
  3. Modified a python script to search google using selenium and collect urls and its search positions. It really helped me to automate the crawl process. Here is an updated version.
  4. Wrote a script to query my webpagetest private instance and gather speed metrics: time to first byte, start render time and fully loaded time.

I analyzed around 1k search phrases and 50k search result urls.

Results I plotted using Jpgraph scatter plot with regression line. I was using regression line slope metric to see how Y-value (speed parameter) really depends on X-value (google search result position).

Here is a script I used to generate graphs.

2.1 Fully Loaded time vs Google Search Result position

Fully loaded time is defined by webpagetest.org:

The Fully Loaded time is measured as the time from the start of the initial navigation until there was 2 seconds of no network activity after Document Complete. This will usually include any activity that is triggered by javascript after the main page loads.

Below you can see how search rank depends on full load time.

 

Fully Loaded Time vs Google Search Result Position Fully Loaded Time vs Google Search Result Position

You can see the slope = 0.0011 and the regression line is almost horizontal. I would say no real correlation.

 

2.2 Start Render Time vs Google Search Result Position

Start Render:

The Start Render time is measured as the time from the start of the initial navigation until the first non-white content is painted to the browser display.

Slope is 0.0017 - no correlation.

Start Render Time vs Google Search Result Position Start Render Time vs Google Search Result Position

 

2.3 Time to first byte (TTFB) vs Google Search Result Position

TTFB:

The First Byte time (often abbreviated as TTFB) is measured as the time from the start of the initial navigation until the first byte of the base page is received by the browser (after following redirects).

Wow! Slope is 0.0049! Regression line is almost 4x times steeper than the fully loaded one.

TTFB vs Google Search Result Position TTFB vs Google Search Result Position

Now I can say there is a correlation.

 

Summary


Google uses time to first byte (TTFB) as a site speed ranking factor.

The experiment outcome is very predictive and expected - google bot can easily measure TTFB but fully loaded time (or start paint time) can take a while to be calculated and depends on many things (browser, network etc) thus volatile.

I described the whole experiment in details and open sourced all the scripts I used - you can double check my findings on your computational resources.

Here are a few resources on how to optimize time to first byte of magento site:

Want us to lower TTFB for you? Get a free quote now!

 

If you find this post interesting do not hesitate to sign up for our newsletter and join the 811 people who receive Magento news, tips and tricks regularly.

Thank You!

 

 

3 thoughts on “Does site speed really matter to SEO?”