Bubble means speed and Bubble is furiously fast // Google´s Light House Speedtest with Bubble.io Apps - How to master it

[meinharttv]

Hi there,

I never produced content, so I hope I do nothing “wrong here”. I just wanted to show you my work on secondra.com trying the last days to speed up the whole thing. Thanks to all the great girls & guys making bubble such a great venture, by the way! See the speedtests in attached files, in comparison to amazon.com / aribnb.com and secondra.com (bubble app), with a great result for not having millions for network routing.

Summary:
Bubble is fast, as it helps as we know to reduce development time & cost.
Apps on Bubble are very fast in execution, if something is done around the apps to tune this in.

But lets start with the meta-idea, the ideas of Google and the web.

1. Google´s sick idea how web apps should look like today
Google has its own idea of how a web app should look like. I say “look” because they think, as they are super sucessful, that people “like” what they “design” or how they work, live and tell us an app should look like. Some think it is a post-communistic approach / a so called “Material Design”. Which is a joke. And if someone does it in another way, so well: I can be a superb funded enterprise, or you get…a BAD LIGHTHOUSE TEST, which ends up in bad rankings, and a very aweful feeling as a coder / bubbler and thinking:"…well maybe it was a bad idea to take bubble.io to create apps, because it is not a coding language as Ruby or React or other, and this must be worse than these best of breed frameworks or programming languages…:"

Which is not true. In no way at all.
Now let´s talk about what I did.

2. Steps towards super speed and 95 Score in Google Lighthouse Test
   I started with a Score of 5 (percent) two weeks ago. Even the app was not so slow, it was like I was trapped. So I started

• setting up a PWA:
  this was not so easy, as I thought, but in the end worked out to let the service worker work, and doing the important caching and more on my app. I did just configure it (some trys and errors, haha, until it worked…). The manifest still does not work, but as far as I see, has no influences on speed, more on other topics as behaviour on the mobile devices.

• intensely change the configuration of showing groups & objects, rg on load entire / above fold, and then test hundreds (or thousands) time the effort for the page speed. Here I took the tool GT Metrix, and reduced results on just “Stop on load” as this is more the realistic usere centirc speed “impression” - as pages load and load even after user can work with it for some time… . Repeating groups I did defer in load, once with hiding, but then when users enters working, I said the rg should begin with scrolling >1 - ideally hiding it beneath fold of page. This speeds up a lot.

• still not enough, I wanted a more crispy feeling on load, so I went on network things, to cloudflare, set up an account, echanged the name server adresses on my google domain account (where I have my web address bought). Then some hours again of config and always retesting the efforts on speed, leveraged web caching for my app on cloudflare´s network. And upgraded the plan there, but I am not sure it is necessary, they have also a free quite impressive plan, but I was so excited I just wanted more!

• then went back to reconfigure & test again on the app, exchanged the first service worker, testing, and now the desktop app is really fast I think.

Resumee:

• Bubble is fast. Speed takes time & testing and some technical mirage.
• Bubble has no disadvantages in no way for me to any other option, but has a lot of advantages.
• Finally, even when speed is superb: the economic success is not guaranteed at all. People do not love a company for the fast app they have. They love ot for the spirit, the speed of adapting to market needs (which is bubble´s super strength) And sorry for my typing errors…;))

See Google Lighthouse with a normal / average page in my app (landing is of course slower, as I have 20 popups and dozens ogf other things plus a repeating group with products loading), but doesnt matter, - its fast!

See the comparison i did some days ago with secondra.com, airbnb.com and amazon.com

Mobile secondra.com Lighthouse test:

[Taiheta]

Your page takes a long time to load

[meinharttv]

yes, of course the landing as I said, is full loaded with stuff and pops – fully loaded, means ALL the page of course. But you must not optimize speed against this measurements, is it doesnt matter in most cases. I.e. if you have a RG, it loads on and on - until the list is full loaded with photos. Plus: a lot of things load on even after page is fully visible and usable. That is the difference. And please, load several times (at least 2) and you will see that happens in the caching.

[Bubbleboy]

@meinharttv you seem to have some errors , might be good to resolve this as this as it could impact speed etc

[meinharttv]

yes, thank you, I know. Have to tackel that! The geocode API I tried, but couldnt resolve it by now…
I think some errors are just “normal” anyway, but of course anything has to be checked.
Thanks,

[rico.trevisan]

Thanks for all the thought you put behind this article!

[meinharttv]

yes, you know, maybe it is for some a kind of help, not just staring at the one figure, and optimizing into nirvana. But it is a lot of work, development in pure to optimize, and if youre a one man show or so, crazy lots of work…

[Taiheta]

@meinharttv

[meinharttv]

just no necessary anymore thanks!

[Taiheta]

From what I understand is that you are trying to cheat speedtests, that is irrelevant when your page takes such a long time to load. When websites take longer then 3 seconds to load I don’t bother. You should fix how you are searching for and displaying data, something is not right with your development.

[JohnMark]

Hi @meinharttv Alex

I’ve tested your page a few times today, and the speed is good, less than 6 seconds each time (New York zone).

[meinharttv]

yes, thank you, by now I have just the “MArketplace”, so I let users go there, in future they will be landing on a more simple index page, of course, always something to improve. Thanks!

[meinharttv]

Hi, I now did resolve my “pop up forrest on landing” which was causing the landing page problems. As I started with “coding” on bubble and began with onbording, I did create cascades of popups - this makes too much DOM load for browsers, and problems, I did not know that in January when staring the project-…— since this week, corrected that mistake of mine, now I have just some pop ups, for onbording only one with groups stacked, as should be - and the landing page speed is also now very good. Some folks checkoud out my landing, but I was talking about my terms page, which was fully ok.

[nocodeventure]

Here in the Netherlands it’s not acceptable for clients to get websites that take more than 1 second to load. So I’m basically stuck and can’t promote Bubble untill this has been solved.

[Taiheta]

Hi, I now did resolve my “pop up forrest on landing” which was causing the landing page problems.

Good to see you admit there was a problem finally. This speed is much better, try remove those errors to get even faster.

[meinharttv]

Great, I like people like you, as not promoting keeps the secret inside our crowd and helps us to propell the business smoothly. Go along, please, a long way!

[meinharttv]

I cannot, as google´s API (new, and updated on Google Account with all things ok on Google) do not change the status in the browsers expert console, saying “API Key not OK”. And: as using a lot of maps on asingel RG / page, causes erros says google. I ca´t change that as this is integral material on the whole project, displaying the map with a toggle on any object. I have no solution for that by now.