Let's face it - no one likes to visit a slow website. We live in a fast-paced, give-it-to-me-now society that savors instant gratification.
When you frequently search new sites for information, you will probably encounter a few slow sites every now and then.
Parts of a site may be loading up slowly such as images or other media taking longer to appear even on a reasonably fast connection.
As a site visitor, your first reaction to a slow site would be to...
Hit the back button and look for another search result.
When site visitors hit the back button quickly after arriving on a webpage, this is regarded as a high bounce rate.
Google recently announced that page experience will play a big role in search rankings by May 2021. Their decision to use more of this data in their ranking should make any business sit up and notice if they care about their Google search rankings.
What Is Considered A High Bounce Rate?
An excellent bounce rate is between 26-40%. Typically, most sites average between 41-55%. Generally, anything above 70% is considered too high.
Since the topic of exit and bounce rates are not the main focus of this article and considerably lengthy to explain, we'll cover strategies for how to decrease bounce rates and how to set up your analytics to track these properly in a future blog post. Stay tuned for future posts.
For now, back to the original programming...
A high bounce translates to a negative user experience (UX) which leads us to...
Why Website Load Times Are An Important Factor for Search Engine Optimization (SEO)
Google doesn't want to deliver a poor user experience in its search results and dings a website's rankings for it. When visitors are "satisfied" with the web results, they tend to stay longer on a website. The time a visitor spends on a site is called a site session, an important metric to track with web analytics. A longer site session means the website engages the visitor more and usually results in better conversions.
Therefore, poor website loading times directly impact SEO and conversions.
As a website owner, you must know that having fast site load times is a ranking factor in SEO.
Google gives preferential treatment to sites that load fast.
In a sign of how serious Google is about web page speed, TechCrunch reported that their Chrome browser will begin labeling sites that are slow with a badge. They are yet to release exact details but it certainly won't be a badge of honor.
If you have any desire to rank on the first page of Google eventually, you need to ensure site load times are comparable to those in the current top search results.
The question now is, how to effectively diagnose potentially slow aspects of your website?
The Fallacy of Using Google PageSpeed Insights to Measure Website Speed
Many people obsess over a tool that Google built called PageSpeed Insights. Google PageSpeed Insights is commonly thought of as a tool to improve website performance.
Some spend countless hours in the pursuit of achieving a perfect score of 100.
All you need to do is input your URL to be analyzed and voila, out comes a score.
With the overall score, you also find recommendations to improve your overall score by fixing each of the "problem" areas.
The PageSpeed tool is powered by an open-source algorithm called Lighthouse. The Lighthouse software audits a website based on a variety of factors, including accessibility, performance, progressive web apps, and more.
For those who want to get into the nitty-gritty details, a more comprehensive view of how the site is actually assessed can be found on Google's Measure tool.
Here's the truth you need to know about the PageSpeed tool:
Google PageSpeed grade does NOT measure the loading time of your site.
What??! If it doesn't measure your site loading speeds, then what does it measure?
The PageSpeed score is akin to scoring a perfect 100% on a test. Does scoring 100% on a test necessarily mean you are smart?
Doing well on a test simply measures your ability to take a test. If you prepared well and used test-taking strategies, you stand a high chance of acing the test.
Therefore, scoring 100 on PageSpeed does not indicate that a site is fast by any means.
Your visitors don't care what score you have because all they want is for the site to load fast. Perception matters and it is not an actual number.
Google has also acknowledged in their 2018 "speed update" that their algorithm only impacts the slowest sites online negatively so most users will not be affected.
In fact, you'll find that many of the top websites and brands around the world don't score well on PageSpeed Insights at all.
Other Tools You Can Use To Measure Your Website Performance
A number of other tools also gauge your website performance such as Pingdom, GTMetrix (score is based on a hybrid of PageSpeed and YSlow), and WebPageTest. The scores you receive from all of them will vary.
What matters most is your website's average load time because this is what site visitors experience.
That said, there is some value to optimizing your site based on some of the recommendations. After all, PageSpeed data still plays a small role in their ranking algorithm but not much from the data we see.
Considering the data output for both the desktop and mobile versions of your site, it helps to work on areas where it impacts speed more significantly such as caching, image optimization, and clean coding (fewer scripts run, combining CSS files).
There's no doubt if you fix problem areas based on the recommendations of the testing sites that your site will run a little faster. Do be careful, however, since there is a point of diminishing returns where the time put in results in little to no real benefits.
Some recommendations are also impossible to implement because it suggests you minify or add expire headers to files not hosted by you but by someone else. Unless you choose another alternative tool for your site, the "problem" simply cannot be resolved entirely.
What Tool Should You Use To Test Your Website Speed Instead?
We already covered a number of tools above. The one to zoom in on is Pingdom. It is the easiest and most accurate way to determine your website loading speed.
By using the Pingdom speed test tool, you can evaluate a site's performance based on different test servers around the world if your business site also serves customers abroad.
If you only deal with US and/or Canada customers, use only those servers in the US and Canada to run your tests.
Do note that repeated tests will need to be spread out over time because they throttle users to prevent abuse.
The tool returns four metrics:
- Performance grade
- Load time
- Page size
The one to look at for your website speed is website load time.
To get a more accurate load time number, be sure to run the test several times and calculate an average because each time you run the tool, it will yield a different number. Not to get too technical, but the reason for the discrepancies in each run is due to the various caching methods.
Let's have some fun and take a look to see how results compare for different sites.
The examples highlight why you should focus on web page load times rather than the PageSpeed score.
We took Youtube and ran both PageSpeed and Pingdom tests on it.
Here's what came back for PageSpeed (Desktop)
Here's what came back for PageSpeed (Mobile)
Here's the Pingdom test result:
As you can see, Youtube, Google's own property, scored pretty poorly for PageSpeed but had a respectable load time of 2.92 seconds considering all the videos it hosts.
In the case of scoring, is this a case of them not eating their own dog food?
Let's compare with a highly ranked PageSpeed scored site from the San Francisco server.
Their desktop site returned an impressive 99, near-perfect score.
What about their mobile site?
Not as great as their desktop result and it appears they still have some work to do on mobile.
How do they compare on load times?
They loaded up really fast at less than 250ms!
Let's compare with a US clothing brand and see how PageSpeed is not what it is cracked up to be.
Here are their PageSpeed findings for desktop:
Here's their PageSpeed result on mobile:
As you can see, their PageSpeed scores for both mobile and desktop weren't the greatest. What about their load time?
We tested based on the same San Francisco server and their load time was impressive - clocking in at 337ms!
The conclusion: PageSpeed scores are not an indicator of web page load times.
Hopefully, you see by now that chasing that perfect PageSpeed score is a fool's metric.
How Fast Should A Good Website Load?
The benchmark of good website performance for 2020 is 3 seconds or less. Most websites average about 6 seconds plus to load.
Realize, however, the goal of reducing load times to less than 3 seconds is hard to reach unless you have someone code the site from the ground up in many cases. Most sites deploy some form of content management system CMS (the most popular one is WordPress) to put a site together quickly and more cost-effectively.
With WordPress sites, webmasters typically install plugins to extend the usefulness of the software by enabling extra functionality such as contact forms, visual editors, and caching abilities. As you load more plugins into the system, the website slows down as a result.
If you discovered your website speed is less than ideal, there are several ways other than using the tools above to identify areas to fix.
5 More Ways to Speed Up Your Website
Since we are on the topic of optimizing a website, we have some bonus methods to speed up your site performance.
Use fewer or smaller images
Where possible without sacrificing image quality, reduce file sizes of your images. If there isn't a need to be graphic-intensive, cut down on the number of images used on any one page. Check the total file size (inclusive of images) of an individual web page and try to keep it under 500KB. Most aren't even close and average between 1-2MB. The max weight of any web page needs to be limited to under 2MB.
Typical file types for images are JPG, PNG, and GIF. JPG and PNG image files are most commonly used on web pages while GIF files are best used for animated images.
To reduce file size, JPG files are preferred over PNG files. PNG images tend to produce higher quality images but JPG can be adjusted for the right combination of quality and file size.
Run JPG files through image compression/optimization tools to reduce the overall file size. Be careful about compressing images too much as it degrades the quality at a certain point and images will be very grainy and unprofessional.
Let's see an example of how compression and image size play a role in a file's weight.
We'll work with a picture of a kingfisher starting at a resolution of 300 pixels and a size of 1280x853 pixels.
The image above is the original photo but after running it through the image compression software ImageOptim, we get this below:
File size reduced slightly from the original of 135KB to 133KB but the result is still great. You can hardly see the difference.
Now, if we reduced the original image size proportionally to 640x427 pixels (without any compression) while keeping proportional scale and image resolution the same, we get this below:
Still a stunning photograph where you can hardly see any difference. The file size of the reduced size photo is now 36KB compared to 135KB in the original sized photo.
The result of simply reducing the photo by half the size: almost a quarter of the original picture's weight just by reducing the size by 50%!
The lesson here: if you don't need many large pictures to deliver an impact on your webpage, reduce the image sizes and run the image compression programs on it to reduce file sizes to be more manageable.
If you are using WordPress as your site builder and CMS, there are many image optimization plugins you can use.
Keep in mind what we said earlier though - more plugins mean greater site slowdown. Here they are below:
Use 3rd Party Video Hosting Instead of Hosting Your Own
If you are currently hosting videos on your site, consider hosting them on external sites such as YouTube and DailyMotion. By hosting externally, you reduce hosting bandwidth and possibly higher hosting costs.
The drawback of free video hosting sites is having to contend with the ads and even competitor's videos appearing.
Use Accelerated Mobile Pages (AMP) for your website
For those with slow times on their mobile speed tests, installing AMP may be a quick way to solve a few of those problems. In fact, Google encourages sites to be AMP
Note that AMP doesn't have any effect on the desktop version of your site.
AMP is basically a stripped-down version of the desktop HTML to deliver faster load times for mobile devices. To serve AMP pages on an HTML website, place the rel=amphtml tag into your site's HTML.
For WordPress sites, downloading the AMP plugin takes care of it.
Use a content delivery network (CDN)
Because much of this content is off-site, you save on hosting bandwidth and get a performance boost at the same time.
Some services you can consider:
Despite its benefits, drawbacks to CDNs include the additional cost, increased website complexity, and some network filtering that blocks CDNs from serving up the content.
Switch to faster web servers
Perhaps the most expensive option but a viable one is to upgrade your web hosting. Simply upgrading from shared to a virtual private server (VPS) will yield a performance improvement. For large company sites and where high site traffic is expected regularly, dedicated web servers are highly recommended.
In addition to optimizing elements identified using the PageSpeed or GTMetrix tool, you can also use the 5 bonus ways such as image optimization, use of CDNs, and other methods to further optimize your overall website performance.
For load times, the use of the Pingdom website speed tool is more indicative of web page load time.
Google places a priority on its search users and it makes sense - that's where the money is. If you focus on delivering a good experience to your site visitors, you'll do fine in Google's eyes.
Now that you are armed with these tools, take action and optimize your site by tackling the easiest ones first.
How did you score on load times after optimization? Did you see much performance improvement after fixing problems identified by PageSpeed or any of the other tools? Let us know in the comments!
About the author: Anthony Yap is an SEO strategist for over 15 years. To him, top rankings are not the be-all-and-end-all because it doesn't necessarily mean money in the client's pocket. If something doesn't put money into his client's pockets, he considers those vanity metrics. He focuses on getting more profits for his clients with strategies that move the needle.