How fast does my website load? This is one of the key questions in performance, UX and search engine optimization. When looking for answers for this question, you should rely on the right tools. Because only a few tools provide really good data.
There are some great overview articles on the web about tools that can be used to measure the performance and load time of one's website easily and, most importantly, for free. These overviews are very helpful, but often make a mistake that can lead to more confusion than clarity. Namely, they compare two to three completely different types of tools with each other and sometimes do not go into enough detail about the differences.
This can then lead to attempts to make performance comparisons with completely unsuitable data. Whis is a pity, as those who measure with the right tools save themselves a lot of time and trouble.
One difference should be known when measuring the performance of a website. The difference between an analysisof potential and the tools that measure the loading time of a website. The first group includes tests such as Google PageSpeed Insights and Yahoo's YSlow. These tools indicate how well the tested website meets the requirements for fast page loading. Thus, both provide excellent input for onpage optimization. However, such tests do not measure the loading time of your website.
Most of the time, though, that's what you want to know - How long does it take my website to build, to be interacted with, to make money? And here's the rub - To our knowledge, there are really only two free tools that provide really good data here:
Therefore, in the following I will show what the measurement tools can do, how you have to operate them and interpret the data so that you can get the maximum out of the measurements.
Three Features that Make a Measuring Tool Particularly Valuable
The tools I mentioned are so good because they provide particularly valuable data. This means that the measurement results
- Consist of the results of several measurement methods that can be compared.
- Are issued in a high level of detail.
- Can be adjusted through various setting options.
In practice, this means that with a good measurement tool you can set how it measures the loading time of the website. This is mostly the choice of the test server and its location, and sometimes also the choice of the test browser.
The level of detail of the data means that not only time data is output, but also a detailed description of what was actually measured. Two tools may well understand something different under "load time". Of course, it is even better if standardized terms are used, such as Time to First Byte.
And finally, it is good if not only the loading time of your website is measured, but also information on the number of HTTP requests, the size of the loaded data packets or the results of potential analyses are output at the same time. This way, you can quickly see which adjustments you can make to optimize the loading time.
Other tools known to us offer such possibilities only to a limited extent or not at all and can therefore lead to problems of understanding and thus interpretation. For you, however, these advantages also mean a decisive disadvantage: The data are by far not as easy to understand as, for example, the results of Google PageSpeed Insights.
Measurement Data is Not Always Easy to Read
The most daunting part is certainly the so-called waterfall diagram. But it is precisely this that provides particularly important data. It is therefore worthwhile to take a closer look at this type of diagram.
In this post I will first explain WebPageTest in more detail. A tool that can quickly provide very informative information about your website and also measures the load time of your website.
My Favorite - WebPageTest
webpagetest.org is fast and very easy to use. In principle, you only need to do two things:
- first enter the domain to be tested
- then select the server location
The server location is - besides the choice of the browser to be tested - one of the most important settings to be able to track how and how fast your website builds up in different regions of the world. In such tests, the server location indicates from which location your website is accessed for the speed measurement.
Always make sure to use the same server location, especially for performance comparisons. Only then can you compare the results meaningfully. If the loading time of your website is measured once from the USA and once from Frankfurt, the loading times will differ significantly.
When you start the test, three large data packets are ejected after a few moments:
- Short Overview - Opportunities & Experiments
- Measured Numbers of Your Website - Observed Metrics
- Individual Test Runs with Waterfall Diagrams
The first part, "Opportunities & Experiments", gives you practical tips on which parts of your website are working well and where there is room for improvement.
Here you are presented with ratings and suggested solutions for various performance-relevant aspects of your website - similar to the tools from Google or Yahoo. It is important to understand that the loading time of the website was not measured here, but the potential for optimizing the loading time.
The loading time of the website is the central parameter with which performance comparisons in particular can be made. It is the direct answer to the question: How fast does my website load? The number of HTTP requests is also well suited for such comparisons. The rule here is - the fewer HTTP requests, the better.
The second section consists of the measured numbers, i.e. the metrics. Below this is the visual page buildup in the form of a series of images in steps of 100 ms. This is also referred to here as a "filmstrip". Images that show a significant change in the page layout are highlighted in color.
The third part shows you an overview of the test runs. Here, the HTTP requests and the content of your website are broken down by type and source. This way, you can quickly see if a certain type of content, including images, consumes a particularly large amount of space or produces a particularly large number of requests and thus slows down your website.
The waterfall diagram shows you in detail the order in which the requests run and how long each request takes to load. The waterfall chart is therefore a detailed view of the loading time of your website.
You will also find a key at the very end of the test results that explains what the symbols and colors mean.
In our case, you will see at the top the first contact with the server and then the loading times of each script.
In this way, you can look at each individual request and compare it with another measured value. In this way, you can determine whether the loading time of your website has improved after an optimization measure and, if so, by what factor.
Advantages - Fast, Relevant, Informative
Overall, WebPageTest is a good tool to measure the loading time of a website. You can quickly detect
- how fast a website loads
- how big it is
- where it still has potential for optimization
- how this potential can be harnessed
- how the construction of your website works
If this information about WebPageTest is not enough for you, I recommend you to go further in depth.
Note Felt Loading Time
The loading time of a website is important, no question, both for the experience and for conversion. But for people, it's not the actual load time that matters, but the perceived speed of the website - or in other words, the time it takes to visually build the website. This is because the website is experienced without knowledge of the measured values. The perceived loading time is therefore an important value for optimizing the conversion and bounce rate. This can also be measured reliably.
The connection between page load time and conversion rate in stores has been well studied in the past. The result - every second counts. Because on average, people are impatient. A loading time that is one second longer can result in a seven percent lower conversion rate.
With every second that your website takes longer to load, you lose revenue. Above-the-Fold Optimization may be the antidote - because this measure reduces the perceived loading time, which in turn can lead to better conversion rates.
Under certain circumstances, your website may take ten, twelve, or thirteen seconds to load, but after only three seconds it looks as if it is completely loaded. Your website would thus be perceived much faster than it is.
Time Series Analysis and Video Function for Perceived Loading Time
1) The Time Series Analysis
Comparing the loading time of your website over several days or weeks can be very informative. For example, if you optimize your website step by step over several weeks.
2) The Video Analysis - The Best Indication of The Perceived Page Load Time.
The video function or image series, which take a screenshot at certain intervals during the loading process, are best suited to measure the perceived loading time. You can see how long it takes until a visual representation is visible. This simple trick allows you to get a clear differentiation between the measured and the perceived loading time. WebPageTest offers this feature in the "Observed Metrics" section as "Visual Page Loading Process".
GTmetrix
I would like to briefly cover GTmetrix here. Essentially, the two tools WebPageTest and GTmetrix can do the same thing. GTmetrix is as easy to use as WebPageTest: You enter the URL of the website you want to test and the tool spits out the result after a few moments. The disadvantage
- GTmetrix offers only one server location, which is in Vancouver
- GTmetrix tests only with Firefox
This means that the measurement results from GTmetrix always apply to a desktop computer with Firefox in Vancouver. As long as you only use GTmetrix for your performance measurements, this does not affect the results. But if you want to know how fast the website loads for German visitors, you have to use another tool.
Conclusion
The measurement of the load time experience is also elementary for the analysis of the load time. Because it allows comparisons on a level abstracted from mere data. We can only recommend tests that enable video analysis or comparable functionality. Because in some circumstances, the measured load time is quite high, but the perceived load time is absolutely justifiable. The value allows you not only to better understand how your offer is perceived, but also whether an optimization of the website with regard to performance is worthwhile at all. Under certain circumstances, it may be more worthwhile to invest in conversion optimization than in the loading time of the website.
When it comes to performance measurement, WebPageTest is the ultimate in free measurement tools for us.
With which tools have you had the best experience with performance analysis so far? Have you had any experience with other measurement tools that might provide even better data?
