Questions from the “Pagespeed” session

During the “Pagespeed” session, our participants submitted a number of questions that our experts answered below. Do you have any further questions? Then post it as a comment under this article.
1. How do I get the Chrome UX Report?

Here you can find the respective links:

2. How do you deal with snippets for cookie banners? Is there a best practice?

It is important to discuss the content with a lawyer so that all central elements are included. Otherwise the element should not appear foreign and should be adapted to the UX and the design. It shouldn’t look like a foreign body. It has also proven useful to display the cookie banner below the page

The cookie banner should also be loaded asynchronously and thus not affect the loading time or reaction time of the page. Here I would work with as little “gimmick” as possible and of course cache the data.
3. How should I deal with cookie statements that move the page?

Google shouldn’t necessarily see a cookie banner because Google cannot read or accept cookies. So the bot does not have to see this query. Additionally, the query shouldn’t shift the layout down either. An overlaying layer would be better.
4. How can I measure the influence of individual elements on the charging process? Does it make sense to block individual resources in order to measure the difference in loading time?

First of all, all external resources such as Tag Manager, Trusted Shop, etc. should be removed. The charging behavior can then be measured via Lighthouse. Here, too, we recommend that you do this not on the test system, but on the productive server, as this usually has a different caching setting than the test server (s).

The developer tools in Chrome show the individual elements and their effects quite well. In order to test the dependencies, one would ideally have to test with and without individual elements.

5. How would you ideally define the first view? Which viewport are you starting from: User? GA statistics? Googlebot (giant viewport)?

Google & users clearly.

Mobile 320 & 360 pixels wide

Desktop 1024 (Google) and at least standard template size.

This results in at least four formats that you should always keep an eye on.
6. Do Google Fonts lead to a faster display or loading time because the font used is already in the user’s cache?

That is correct, but so that the font does not become a bottleneck, the call should be extended to include the GET parameter display = swap.

7. When are CDN servers available?

But when it comes to CDN, it means that you don’t necessarily need it if you have a good hoster in Germany. A CDN is always interesting when users on different continents are to be addressed or when a server can no longer handle the load and one has to or wants to use load balancers.

In other words: Page speed problems cannot necessarily be solved by a CDN.
8. Which tools are recommended in addition to LEAP and Searchmetrics?

It feels like a new provider is added to the market every month. But what we still like to use is WebPagetest.org. We have been using this tool for years when it comes to page speed optimization, long before anyone spoke of web core vitals.

The Chrome Developer Console is great. However, correspondingly technical. In my opinion, really the “go-to” tool.
9. Do you also have experience with Webflow and the scripts loaded by default by the framework?

Yes, we have already optimized webflow pages. Do you care about the page’s page speed? Please be more specific so that we can help you.
10. What exactly does Searchmetrics measure in page speed and what is the value (FCP, LCP, …)? And which JS are also measured when “JS mitcrawlen” is activated in the crawl settings in Searchmetrics – including all external scripts such as Google Analytics and asynchronous scripts?

We would like to feed this question back directly to Searchmetrics and have their IT answered.

Searchmetrics Development Team: If it’s a regular (non-JS) crawl, the load time is the time until we get the response from the client’s server. Resources are not loaded here.

If it’s a JS crawl, we measure the time it took to finish rendering the page, including external JS resources. In addition, the cache is deactivated so that every page is re-rendered.
11. Where does the concept of “above the fold” actually come from?

That’s a good question, which we haven’t even thought about. What we do know, however, is that the faster a page loads, especially the visible area “Above the fold”, the more sales the website generates. Because what is always forgotten is that Pagespeed has a direct and indirect effect on the turnover of a domain.

The idea comes from the newspaper industry and includes everything that doesn’t require interaction. The reader does not have to turn the page or open it. Transferring this to the network means everything that is directly visible.

12. Which tool do you use to extract glyphs from fonts?

With a transfonter you can extract individual glyphs from the fonts.