How to Really Test WordPress Page Load Times?

How to Really Test WordPress Page Load Times?

August 6, 2018 By Jake

If it sounds too good to be true it usually is!

Nothing makes my blood boil more is when I see an ad or a posting of someone offering to speed up your WordPress site, and then the falsely advertise the results. I consider them “Charlatans”.

Let’s look at an ad that was in a popular WordPress help group on Facebook, where users can ask “experts” for advice. Of course, the ad was from the admin, who happens to own a WordPress support firm and he is the only one who can advertise. (See the problem there.)

I would post this article or answer his claims in the group, but I would just get banned for revealing the “magic trick”. So I am going to post it here and hopefully save someone from wasting their money.

Here is the claim I saw yesterday. The site was taken from 10.2 seconds to 3.2 seconds. Sounds exciting and enticing, as it should. With the emphasis on load time and mobile first in search engine ranking, we all want our sites to be fast.

False Claims as to increasing WordPress page load times
Remember that saying, if it sounds too good to be true, it probably is! Well, one reader did comment: “I just try again a few seconds ago … I tried several times like at least 5 times and got a similar result like below.”

The magician’s response was “Cache has to rebuild when plugin added/removed, page update or expire schedule is met. If you test when rebuilding it will be slow.” He was semi-honest there. He did leave out the part that the load time of 1.8 seconds was for a visit after his “magical” plugin was installed. The response from the reader should tell you the optimizations are very resource intensive, and it will only get longer to install the site.

Also, the results you get the first time you run the GTmetrix test, will be higher than the second-time you run them, do to the cache being leveraged in the browser.  But the purpose of a fast page load time is to keep the first- time visitor from leaving your site.

This article will explain the proper way to, when to and how to test your page load times.

Where should I run my page speed test?

Well, the claims were made off of GTmetrix, so let’s start there. So 5 hours after the claim was made, I ran the test.

 

Well, the site is now loading 100% slower than the claims for it being optimized.  3.2 seconds to 6.5 seconds.  I ran the test again, and it got slower!

Why?  This account is on a shared hosting account.  Since the plugin that was used to optimize the site is very resource intensive, the host has probably “throttled” the account, so its consumption of resources does not affect other accounts on the server.

Also, website owners, when seeing a faster site, tend to want to add more items to the site.  This consumes more resources and slows it down.

If you run the same test, a few times, you will get different results every time.  GTmetrix is a starting point. It gives some great advice, but like other tests, it is not consistent.

When I am building a site or testing for speed I use Google Chrome and look at the Web Tools. 

Where should I run my page speed test?

The test should be run on as a first-time visitor to the site. This is what really counts.  This may sound like a “catch 22” since you probably visit your site a few times a day.

How can I test my page load time as a first-time visitor?

Launch your Chrome browser.  (If you do not use chrome and you own a website, go sit in the corner.) Make sure all other tabs are closed, and all that is open is just one tab at Google.com.

First, we need to clear all of the cached images and files from the site.

  1. At the top right, click More (Hamburger Menu – 3 dots).
  2. Click More tools Clear browsing data.
  3. At the top, choose a time range. To delete everything, select All time.
  4. Next to “Cached images and files,” check the box.
  5. Click Clear data.

Second, bring up the developer tools

  1. At the top right, click More (Hamburger Menu – 3 dots).
  2. Click More tools
  3. Choose Developer Tools
  4. Once it is loaded look at the top menu
    • Click on Network
    • Click on the camera – when you do you will see the message in the image below
    • Check the box disable cache

Now you can enter the URL you want to test and load the site.

Here are the results returned as a first-time visitor:

Load time: 13.42 seconds as a first-time visitor.

Now let’s run a test as a second-time visitor, by unchecking the box “Disable cache”.

Now it loads in 5.38 seconds.   The reduction is caused by leveraging the user’s browser cache.  It is the smart thing to do.  It speeds up the return visitor’s load times and reduces resource usage on the server.  But to claim it as page load time is false.  When a website is tested for page load time Google does it as a first-time visitor.

Is mobile page load speed important?

It most certainly is. Unless you have been living in a closet, it is a mobile world now.  More surfers are on mobile and tablet than desktop these days. With mobile first and load times now influencing your SEO indexing, that is really how you should be testing.

But if you notice, all the testing so far has been done simulating a desktop, because those were the claims that were made.  They usually do not show you the mobile load times, because they are always slower.

Here are the test results as a mobile visitor.  Repeat the steps above for a first-time visitor by clearing your cache, tick the box “Disable cache”, and then this time change the speed to 3G as shown below.

The first-time visitor on a phone yields a load time of 20.20 seconds.

Then for the return visitor, un-tick the “Disable cache button and run again.

Of course, the second visit loads faster at 6.64 seconds.  But this is not the first visit.

If you want the think like Google (The ones that test your site), you can always run your site through this site:

https://testmysite.thinkwithgoogle.com/

 

Now, Google says 15s for the first-time visitor at 3G speeds. There is an explanation as to why our tests in Chrome say 20.2 seconds and Google’s mobile test site is 5.2 seconds lower.

What is the perceived page load times to the surfer?

A visitor to your website perceives the load time to be when the content is loaded “above the fold” and when he/she can interact with the site.  This is what I try to stress to clients.  It is the user’s experience, not the test program that counts.

How do I test the load times that matter to the visitor of my website?

You can effectively use Google Chrome, once again, to do proper testing. In fact, we did when we ran the original tests.  We have frames that show a picture of what is loaded and at what time.

If you look at the time frame pictures, at 15 seconds you can see that the user perceives the site as loaded and can interact with the site, as the other parts load for another 5 seconds.

How much is a slow page load time costing me?

Across industries, the average landing page conversion rate was 2.35%, yet the top 25% are converting at 5.31% or higher. Ideally, you want to break into the top 10% — these are the landing pages with conversion rates of 11.45% or higher.

Know your numbers.

My statistics tell me that out of every 100 visitors that come to my site 2.35 persons will sign up.

My stats tell me the average customer will generate $200 in revenue per year.

My stats tell me I get 50 first-time visitors per day.

So in one month, my site gets 1500 new visitors.  I should convert 2.35%, so that means 35 conversions for the month. An increase in new revenue of $7,000 for the next 12 months from those new conversions.  That is $583 monthly or $16.67 per conversion. So every conversion is worth $16.67 in new revenue per month.

Google says I am loosing 32% of my visitors due to load time.  So that means my 1500 new visitors represent 68% of my traffic which stays.  So my gross traffic is 2200 per month. So if I get my site to load faster, so let’s say, 5 seconds, Google tells me I will only lose 10%.  So if I run those numbers, that means faster page load times, will increase my new visitors by 485 per month.  That is 14 new conversions per month, which equals new revenue of  $233/month or $2800/year.

When you put it all together, over a 2 year period of time, your total new revenue would be $64,308.00

Also, consider that the top 25% to top 10% of the industries have conversion rates of 5.31% to 11.45%.  I guarantee their sites do not take 20 seconds to load.

So you need to find out what the slow page load is really costing you.

So what do I need to do to get my page load times down to 5s or less on Mobile then?

That depends, but a $45 speed up special is not going to do the trick.

Things you need to look at and find out:

  • Can your hosting handle the resources necessary for proper load times?
  • Have you done any optimizing of images, CSS, or JS when building the sites, or do you rely on plugins to do the job?
  • Is your template code “clean and efficient” or are you using a page builder which bloats your code?

You may just need to do some optimizing at the core level, or it may be time to rebuild the site from scratch.

The Next Blog Entry will be about doing it right from the beginning.