Why are my scores down after enabling Page Speed Optimization for a few days? (OLD)

Why are my scores down after enabling Page Speed Optimization for a few days? (OLD)

When using Page Speed Optimization from Tapita, most stores experience a significant instant page speed score increase when testing with Google PageSpeed tool. However, after a few days, some stores report that their speed scores are going down, and they think this is due to Tapita app. If you are one of them, please read this guide.

What exactly does Tapita page speed feature do?

When you test your page speed with Google PageSpeed tool, you will see that the final speed score is calculated using a set of different metrics:
  1. First Contentful Paint (FCP)
  2. Largest Contentful Paint (LCP)
  3. Total Blocking Time (TBT)
  4. Cumulative Layout Shift (CLS)
  5. Speed Index (SI)
Since we noticed that TBT is the metric that most Shopify stores perform the worst, Tapita app only focuses on optimizing Total Blocking Time (TBT) score.

To optimize TBT score, our app delay the JavaScript that have the most effect on your loading time.

An example of a store BEFORE optimizing speed with Tapita

This store has a pretty bad TBT score, among other things:



When we choose to show audits relevant to TBT score, we will see the issues that affect this score:


An example of a store AFTER optimizing speed with Tapita

Now, let's take a look at the same store after enabling Tapita page speed feature.



Here, we can see that its TBT was reduced from 1,120 ms (milliseconds) to 460 ms.

Looking at the TBT-relevant audits:



We can see how the issues relating to TBT was fixed by Tapita:
  1. Reduce the impact of third-party code: this was completely fixed so it no longer appears on the issue list
  2. Reduce JavaScript execution time: this issue is still there but with great improvement - from 4.2 s (seconds) to 1.7 s

What causes the fluctuation in speed scores?

An example of a store whose scores had been down after a few days of optimizing speed with Tapita

A few days after optimizing their speed with Tapita, this store's scores had been down a bit. Checking in details with Google PageSpeed, we see that their Total Blocking Time (TBT) has jumped from 460 ms to 1,390 ms.



Checking further into the audits, we observed some things:
  1. Reduce the impact of third-party code: this issue still remain fixed, since we do not see it in the list
  2. Reduce JavaScript execution time: this has increased from 1.7 s to 3.2 s. Checking with the store owner, it turns out they recently installed some new tracking code in their store.
  3. Avoid an excessive DOM size: this is due to this customer adding new content to their page, which makes the page much more heavy.


To fix this problem, we will then need to manually audit the page and optimize the elements on the page.

There is another issue that causes your scores to fluctuate: real-user-experience metrics.

These are the metrics that Google PageSpeed get from your real website users. When you do not have enough traffic, though, these metrics are missing:



When you have enough visitors traffic for Google PageSpeed to get the speed data from, it will show on your speed result page like this:



When this happens, these metrics will be taken into account to calculate your final scores.

This is why sometimes your scores might go up or down even when if you did not change anything to your page. It is simply because your users have different experience when loading your page. For example, for some reason, one day there is a surge of traffic from users from a very slow network connection to your site. this will cause the real-user-experience metrics of your site to go down, which affects the final speed score.