Keeping an Eye on Your Page Speed: The WordPress Performance Framework

Jan Hornung Last updated 21.10.2020
6 Min.
The RAIDBOXES WordPress  -Performance-Framwork
Last updated 21.10.2020

Performance checks are time-consuming. Especially if you optimize not only one, but several sites . Therefore, it must be clear which optimization measures are worthwhile at all and what you are fighting against windmills with. And of course you have to keep an overview. Today I have a framework in the luggage, with which you can precisely record all your optimization measures and precisely determine which measures really benefit the loading time of your website.

Whether Webpagetest, Google Test My Site or Google PageSpeed Insights - there are now a whole range of tools available to you as a webmaster with which you can measure the loading time of your web pages in detail. In particular, the Google tools Test My Site (for mobile sites ) and PageSpeed Insights(for mobile and desktop versions) give you tangible tips and instructions with which you can reduce the loading time of your web pages: Set up caching, shrink images, minimize CSS files.

The same applies to monitoring tools, which record the PageSpeed ratings and loading times of your sites at regular intervals.

So you set about implementing all the usual measures in painstaking detail and in the end you find that your site now actually loads significantly faster. Very pleasing. But which measure was the most important in the end?

We know from experience: Not all optimization measures make sense

In more than 450 detailed performance analyses, we have made the experience that the really relevant problems are not always addressed. Result: For example, a lot of time and money is spent on setting up a CDN, even though this ultimately brings no performance advantage (e.g. because the visitors come exclusively from Germany). Some interventions can therefore be safely saved and the time invested more sensibly.

For this very purpose, we have built a small framework internally that we use to check the results of our own performance optimizations. With it, we can track exactly which optimization measures have led to which results. And we also use it to measure the influence of Plugins and Themes on page performance. 

So today I'm going to show you

With better WordPress performance ranking and conversion rate increase

The fact that the loading speed of a site is a ranking factor was already officially confirmed by Google in 2010 and is therefore actually old hat. The faster the site is, the better it will rank.

Sounds trite, but you have to internalize it.

Google itself recommends, for example, as an ideal value for the Time to First Byte (TTFB), i.e. the time until the server responds, is 200 milliseconds.

RAIDBOXES WordPress   Performance Framework error message from Google PageSpeed Insights regarding TTFB
This message from Google PageSpeed Insights will only be displayed if the TTFB of your site is less than 200 milliseconds.

The correlation between the loading time of a website and the ranking on Google was also independently proven in 2013 by Moz, one of the American big players in SEO. In these two studies, the colleagues find a strong correlation between ranking and TTFB:

Interestingly, the Americans could not find a connection between the total loading time of site and the Pagerank! So you see: Not every optimization measure pays off directly on the search engine optimization.

The concrete WordPress performance is especially relevant for online businesses.

The topic of page speed is also of enormous importance for online businesses outside of search engines. Because the WordPress performance influences conversion rates and user behavior.

0.1 second more = 1 percent less turnover

As early as 2006, Amazon determined via A/B tests that 100 milliseconds of delay in loading can mean an average loss of 1 percent of sales. A full second delay adds up to $1.6 billion in lost revenue per year for Amazon.

In a slimmed-down form, this principle thus applies to all shops: If the shopper has to wait too long for site , the probability of a purchase is reduced. The Google tool Test my Site already takes this into account in its input mask with the saying:

"Most sites lose half their visitors while loading."

Raidboxes wordpress performance framework the homepage of google test my site
In addition to the saying, Google Test my Site also outputs a value that shows how likely users are to bounce from a site because it loads too slowly.

Also for content applies: The slower the site , the less is read.

In 2016, the Financial Times tested how readers react to a delay of one to five seconds in loading time.

It was found that the slower site became, the fewer articles visitors read per visit. Thus, both subscription and advertising revenues suffered from a too slow site .

The conclusion of the media house: Half a year later, the website was completely overhauled, with a focus on page load time. The redesign of a website and its load time optimization alone have thus helped to boost the core publishing business.

This connection is also evident in the services business

And also for service providers, the loading times of their own website are important. This is shown, for example, by the case of the Canadian software developer Intuit.From 2012 to 2013, the provider was able to improve the loading time of its site by over nine seconds.

And this optimization also had tangible business advantages: When the loading time of the website was still over seven seconds, the conversion rate could be increased by three percentage points with every second reduction. Even when the loading speed was already below four seconds, every second of improvement still added one percentage point to the conversion rate.

Let's summarize: For shops, publishers and service providers, website load time has a demonstrable impact on conversion rates. Not to mention its relevance as a ranking factor.

The reason for this is assumed to be in the perception of the users. These articles show impressive examples:

That's why it's important not to get into shadow boxing matches when it comes to performance optimization, but to really optimize where you can positively impact the site visitor's perception.

Maintain an overview, identify adjusting screws: This is what our tool

So, after all the facts and figures, now something solid. By the way, you get access to the framework via one of the input masks in this article.

But what are the concrete benefits of the framework?

During the optimization and analysis of hundreds of WordPress sites we noticed that the analysis tools GTmetrix and Webpagetest create a test history, but that it quickly becomes very confusing if you test several sites regularly.

For example, we found out that especially with slim sites Plugins , which compress CSS and JavaScript, doesn't always make sense. At least not if the site runs under SSL and uses HTTP/2. So you can save the effort at this point completely or at least put the measures on the back burner.

Our framework is based on the Google spreadsheet tool. We ourselves use a semi-automated variant, where the tests are performed automatically with Webpagtest.

Who can use the performance framework

Especially for SEOs, agencies, designers and consultants it makes sense to have many sites in view at the same time. This is also where we see the greatest benefit: With the framework, you can test different measures directly against each other and systematically find out what brings advantages and what does not.

How to use the RAIDBOXES performance framework

The name is a bit bulky, but the use is quite simple:

Step 1: Search for measuring tool and collect basic data

As a starting point, you collect all the loading times of your site and record them in the table. You can use Webpagetest, GTmetrix or Pingdom Tools for this purpose. We recommend Webpagetest. The Google tool simply offers the best settings for really good measurement.

Be sure to select the correct spreadsheet for the test you are using. No matter which tool you use, the following values are usually recorded: Time to First Byte, Start Render (time when the visible page load starts) and the Load Time (how fast the site loads in the perception of your users). Values such as the perceived load time are calculated automatically by the tool.

It is also important to note sites -URL and the date of measurement.

Step 2: Optimize and measure again

Now you perform your optimization measure (e.g. compress your images), make another measurement with the same tool and enter the values in the corresponding fields. The framework now automatically calculates whether - and if so, by how much - the individual values have changed. Of course, it is important that no other changes have been made to site .

That's it. That's all that needs to be taken care of. Although we have one more tip at this point:

It is important that you always carry out several measurement runs. Only in this way, and by calculating average values, will you arrive at reasonably reliable values. The colleagues from HootProof explain why in their big "WordPress -Optimization-Guide". In our article on the use of Webpagetest or our performance e-book, we also show how you can reliably produce good measurement values with the Google tool.

Click through and try it out

Simple but powerful: With the framework, you not only keep track of your current optimization cases, but also of the performance of your ongoing WordPress projects. This makes sense especially if you manage many sites at the same time and want to better understand your optimization process.

If you have further questions about the Performance Framework, just leave a comment or send me an email directly .

RAIDBOXER from the very start and our Head of Support. He loves talking about PageSpeed and website performance at BarCamps and WordCamps. The best way to bribe him is with an espresso – or a Bavarian pretzel.

Comments on this article

Post a comment

Your email address will not be published. Required fields are marked with *.