WU Optimization from over 6 million WUs to less than 100k

Hey there,

With the new pricing coming in soon I hope everyone has been busy optimizing their apps. Bubble seems to be doing their best to write guides and be more transparent but I still have little to no faith in the implementation of WU or what could happen if someone causes a loop / system malfunctions and you are caught for a large bill. This happens fairly often with companies like Vercel.

I have an application running on Bubble which was built fairly optimized. We ran a lot of the complicated APIs via N8n so that Bubble wasnā€™t doing much heavy lifting but still it got to over 6 million WUs a month consistently. With the new pricing, this would cost us approximately $734 monthly. There are no major tricks here, we just moved to an external backend.

Through optimizing the app, we have made it faster, we have made it more performant and the feedback from users is that just everything loads faster and it is easy to navigate the large amount of data.

What concerns me about Bubble and I donā€™t want this to be a Bubble bashing post is that not once in any documentation is using an external backend mentioned as the best way to reduce your WUs. It quite clearly is.

This is where I think that the Bubble strategy of trying to do everything, Jack of all trades, master of none will be the downfall. I can now reduce my bill significantly with Bubble due to the migration.

I also have the added benefits of privacy. My data is in my external backend and only the people who are authorized via user permissions etc can see it. With the few threads popping up of Bubble ā€˜accidentallyā€™ locking accounts / turning off apps, this also lets me sleep peacefully at night.

How much time and effort was put into learning the skills and pathway to reduce an app WU like this? Probably more than I like but now I am very happy because I have an external backend which I can use to make Discord Bots, to make a mobile app, to build a solid API on top of etc etc.

Bubble is a great product but I think the Jack of all trades approach will drive people away and/or explore other options. My hand was forced to optimize so I optimized significantly and now when I switch on the new pricing my Bubble price and backend price will be less than I was paying monthly for Bubble before, so thatā€™s a win in my books.

10 Likes

Great post.

More evidence of Bubble being penny wise and pound foolish. You will now be paying LESS than previous and the actual costs of your app havenā€™t been reduced in many meaningful way so Bubble will make less from your subscription. Itā€™s the consequence of Bubble charging inflated amounts for WUs that arenā€™t tethered to reality (compare with AWS costing less than 1/20th) so when you work around it, as you have @stuart8, Bubble gets left holding the bag.
Itā€™s simple unit economics that they refuse to acknowledge and that is hurting their product and bottom line :exploding_head: (not to mention all the wasted months on trying to actually calculate the WUs - which is still inaccurate).

Stuart, just pointing out the < and > in the subject of the post are reversedā€¦

5 Likes

I always get them wrong, itā€™s like thing I google the most :joy:. Alligator always eats the biggest number or something but apparently not then with how I did it. Regardless, changed it.

1 Like

Using external databases has been a common among experienced Bubble devs even before WU was announced.

Thereā€™s no 1 size fits all solution. Bubble or not just use the tools that meet your needs.

1 Like

@stuart8 great post!

I believe that there one reason Bubble can be missing the bandwagon here. Technology is build on top of choices made 12 years ago. Currently we have a very different tech stack to work with. There is no reason that an average bubble app cannot work with hundreds of other apps on a single ā‚¬100 server a month. So there is not enough margin for Bubble now I think on the small plans.

So then you have to make a decision, either go fully corporate and charge ten times the money which is happening now with dedicated. Or you should rebuild from scratch such that it becomes a platform for all.

Rebuilding will take 12 months and will break all existing apps.

Going corporate might work as the economics are working there. A small sized project in a corporate environment does about 100k in cost each month. A dedicated Bubble server will cost only a few percent of that but it also means that a few developers can go home. Saving them probably 30k a month.

Why I think it will not work in the end is because corporates do not own anything after building with Bubble. Also, there are many established lo and no code frameworks in the enterprise space doing a great job with ownership of code

Isnā€™t it a given that not using Bubble will reduce your Bubble costs :joy:

7 Likes

I think itā€™s well known that using an external backend would significantly reduce your WU consumption and already there are many Bubble applications that are using it as a front-end only.

1 Like

Pepperidge Farms Remembers :frowning:

https://forum.bubble.io/t/announcing-a-new-pricing-and-the-ability-to-buy-server-capacity/

1 Like

@stuart8 How do you handle realtime data and auto refresh of data?

I.e. you create a new thing, how do you display it smoothly without page refresh?

I think the topic should be renamed as ā€œSaving on WUā€ instead of WU optimization. From my point of view, this is not optimization.

6 Likes

Anybody here who has done an exact price comparison?

Like that:

Do a search with 100 (1000, 10.000ā€¦) bubble vs. external backend

Fetching results via api is less than fetching from bubble database?

Same for:

Make changes to a thing ā€¦

Make changes to a list of thingsā€¦

Fair is fair. We could talk about migration pains, complexity, one-platform-convenience, etc. But at the end of the day if this new setup serves you better, then thatā€™s that.

On our end, I can tell you that the team is thinking about WU pretty much every day. There are roadmaps in place to make things better, and I share everything I come across with the team (like this thread) for discussion.

Thanks for sharing about the switch and if you want to share more details, Iā€™ll be happy to keep sending them their way :rocket:

13 Likes

Itā€™s not really news that you can use alternative backends. Thatā€™s a strength of Bubble. You are not locked into the database.

When itā€™s the right thing to do, move your database and costs elsewhere for the right reasons.

2 Likes

Having an external backend and this tool combined seems to be the ultimate WU optimization.

using websockets. Very simple.

Legacy plans will be forced to switch to the current WU plans in October.

If you are down to using websockets, what is stopping you to go to a fully custom code. Your database is external as well. And Bubble frontend is not the best in class. I really wonder what keeps you still at Bubble? Not questioning, really just curious. Because for me (who is not still interested in costs yet), I am interested in the abstraction and the technology and speed of development.

6 Likes

Can you tell us about some of these roadmap items and their expected release dates? Itā€™s been 16 months and Bubble hasnā€™t really given us anything in regards to WU. I literally check the monthly announcement thread every month and they donā€™t even mention WU nor answer any questions about it. It seems like they are trying to ignore itā€™s existence. The reporting is still atrocious and the fact that we still canā€™t select which fields are returned in a search means we will be burning WU that we donā€™t even need because of Bubbleā€™s limitations.

2 Likes

Iā€™m sure @fede.bubble could speak more on this than I can, but what I am aware of, I believe it is mostly focused on better ways to visualize and analyze the WU metrics. Some of what I have seen as potential new layouts and features are awesome for that purpose, but sadly I am not aware if Bubble is planning on doing anything to help us reduce our WU consumption through things like defining which fields to return on a search, lowering what certain actions/functions cost in terms of WUs or ensuring that the system is not as full of bugs as it has been for the past 16 months in terms of overcharging, or under reporting (ie: WU pricing chart doesnā€™t show everything they charge for, and at times it contradicts what the editor shows was charged)

3 Likes

Sorry if I missed it, but what external database is being used? And what kinda cost is it incurring compared to the native bubble option?

Also for the experts out there can you tell me if itā€™s possible to use bubble dB for some things but have a separate external table/database for others? For example Iā€™m happy with the built-in database for user management, logs, and most other things, but when it comes to letting the users add ā€˜likesā€™ (Ie frequent writing to the database) it really wracks up WU. So an externally hosted ā€˜likesā€™ table could be an idea? Is that possible?