Today I’ll be testing if there is a noticeable page speed difference when using Burst. I’ve set up a test website on InstaWP. I’ll be using Pagespeed Insights to test for a noticeable difference. I’ll use the latest version for this test, which is version 3.2.0 at the time of writing.
Short answer: No, but some settings can make a difference. So make sure to check the results.
Empty fresh WordPress install (Baseline)
First, I’ll test a new WordPress install using the Twenty Twenty Five theme. The WordPress Core Performance team did a great job with the Twenty Twenty Five team, scoring a 100% on Performance, Accessibility and Best Practices. At the time of writing this, we use this theme on this website as well. We wanted to use Gutenberg because it is fast, and so we can understand what most new WordPress users will use.
I’m not sure if these results will stay online, but for now, you can see the pagespeed test here.
WordPress install with Burst Statistics (free)
So I installed Burst Statistics from the WordPress repository and did another PageSpeed Insights test. At first, I thought the page was cached cause the results were almost similar (they were even a bit better). I had to check the treemap to see if the Burst script was actually loaded, but it was. So the Burst tracking script does not slow down your website. I can stop here, but there is more to this than testing an empty install.
What about having lots of data in your database? What about cookieless tracking? Would that slow down the website? Well, let’s find out!
Does cookieless tracking slow down your WordPress website?
Before doing this test, you might expect cookieless tracking to have some impact on performance. Cookieless tracking uses a technique called fingerprinting, which requires a bit more JavaScript to collect the same data you’d normally get from cookies.
Let’s put it to the test.
With cookieless tracking enabled (and Turbo mode on, which is the default) the pagespeed score comes in at 99 out of 100. That’s barely noticeable, and in practice you’d never tell the difference. Turbo mode, which has been on by default for a while now, loads the tracking script deferred and only starts the tracking process once the page is fully loaded. This keeps performance impact to an absolute minimum. The one trade-off worth knowing about: visitors who leave before the page finishes loading may not be tracked. Whether that matters depends on your situation.
Disabling Turbo mode with cookieless tracking enabled
Our previous test with cookieless tracking was with the default settings, which means Turbo Mode was enabled. To get a picture of the effect of Turbo Mode, we need to test this separately. Let’s check what happens if you turn Turbo mode off:
The score drops to 88, a meaningful difference. So while cookieless tracking itself does require heavier JavaScript execution for fingerprinting, Turbo mode effectively neutralises that impact.
Cookieless tracking works great out of the box. Since Turbo mode is enabled by default, most users will never notice any performance difference at all. Just keep it on.
How does having lots of data in the database impact the pagespeed of your WordPress site?
As we are part of the TeamUpdraft family, I’ll obviously use UpdraftPlus Premium to add data to our test environment. I like to use it for tests like this because we can select the database tables we want to back up and we can transfer just the burst tables from our live website to this test site. So the site will still be fresh, but it will have quite a lot of Burst data (304 MB to be precise). I’ve also reset the settings to the defaults, so without cookieless tracking.
A perfect score! I think that concludes that Burst does not slow down your website, even after gathering lots of insightful data.
To be honest, the database size does have an impact on some things, but not on your website’s visitors. When you visit the Burst Dashboard, you load data from the database. The larger the database, the more data your server has to go through to get the right data. If you have a few years of data and you only want to see the data from the last 7 days, the database has to find the right data. Our setup is well optimised to handle this, but the size of the database plays a big role here. So if you feel like the Burst Dashboard loads not as fast as before, it might be time to archive or delete some data.
Are these tests accurate?
Pagespeed tests can differ quite a lot from test to test. But mostly it gives a good insight into what a change does, especially on a small website with not many variables. For these tests to be accurate to you need to have caching disabled and don’t use a CDN as these will skew the results. On active websites, you should obviously use caching and a CDN can be beneficial as well. Burst works without issues on cached websites or when using a CDN, It is recommended to clear the cache after changing settings or updating the plugin though. This will prevent some settings from actually changing, so keep that in mind.
For these tests, I used WP version 6.9.1, PHP version 8.0, no caching and no CDN. I tried to remove as many variables as possible, but your results can differ.
Conclusion
In general Burst Statistics does not slow down your website. But some settings can change this, like cookieless tracking. But Turbo Mode neutralises the impact of cookieless tracking almost completely.
If there are more tests you would like to see, please open a thread on the WordPress forums! I’d love to do some more tests and improve our product.