When you implement the tracking code into your website, you may want to know how that will affect the loading time.
When you implement the synchronous code in the <head> tag, the changes are applied before the content loads. This means that the time load of the website has a minimal increase.
What is the loading speed?
a) When implementing the tracking code asynchronously, the loading time won’t be affected. The average of the perceived loading time delay is 50 milliseconds. If there are active experiments on that page, it could last approximately 150 milliseconds. If you have an active A/B test on one page, for example, there might be a small flickering effect displayed.
b) When implementing the tracking code synchronously, the user experience is better, without any flickering effect. However, in this scenario, the perceived loading time delay is between 60 – 300 milliseconds.
What happens if the Omniconvert server stopped responding?
In this extreme scenario, the active experiments won’t be triggered and their results won’t be affected at all. Worst case scenario, there might be a small delay of 1 second in the loading time of your page. However, that won’t affect all the other elements displayed on your website, or other available resources.
What is the protocol for extreme situations?
Omniconvert’s protocol for emergency situations clearly states that our tracking code will skip loading and won’t apply the changes if the time load is above 1 second. This security time load checker is used to prevent any influences in the time load and the users’ experience.
If you want to avoid that small perceived loading time delay, you could also disable the Omniconvert code temporarily, until the issue is fixed. We strongly recommend you use this feature instead of removing the Omniconvert code from your website, to avoid losing all your experiments’ data.
What if my website’s pages load slowly or do not render properly?
When you type in an URL and the page loads, there is a complex delivery system built around data packets ( images, CSS files, plugins, etc. ).
If the internet connection is slow, our tracking code has a security time load checker (called Loading Tolerance) that will skip it and it will not apply the changes in order not to influence at all the time load and the user experience.
You have the option to edit this setting. A higher Loading Tolerance means that we will wait for a longer time for everything to load. This might also mean that, if the script is synchronously added, we might block loading other elements for a longer period of time. It is up to you to decide how much should this value be increased. Currently you cannot set it lower than 1000ms equivalent of 1 second.
What is the flickering effect?
This happens when the original content is loading first and after that, the changes are applied. This can distract the users from the CTA experiment and the results not to be accurate.
When you install the tracking code please consider the following:
- Position: place it to be the first in the <head> tag, if possible. We highly recommend this position for the changes to be applied before the content loads, in order for the user to have the best experience.
- Type of code: with the asynchronous code, the loading time will not be affected, but for the A/B test experiments the flickering effect will be visible. With synchronous code, the loading time will be affected with 60 – 200 milliseconds top, but it provides the best user experience.
- Tag Manager: place our tracking code to load first. This helps the changes to be applied before the original page loads.
Click here to read more about integration with other plugins.
- JQuery: we recommend using the ‘jQuery bundled in Omniconvert code’ option because the experiments will be affected if the version used by you differs from the version used by our platform.