Monthly Community Update -- March 2023

Announcements
1

Hi all,

This is our March community update! Read last month’s update here.

Our focus this month continues to be increasing the rate at which we evolve the Bubble editor to be more useable and powerful, while making the platform more stable and reliable. We’re also increasing our investment into our ecosystem of partners and developers, to make it easier and easier to earn a living building using Bubble. Our long-term goal, as always, is to make Bubble the go-to choice for new software development, instead of traditional full-stack coding. We’re excited by the progress we’re making!

This month, we welcomed the following people to the company:

  • Missy, who is joining as our new head of Product Design
  • Greg, joining the Success team as a Technical Product Support Specialist
  • Tailer, joining as an Associate Visual Designer
  • Mike, joining as a Senior Engineer on our Platform / Cloud Infrastructure team

If you would like to join us, check out our careers page here. As always, we highly encourage community members with solid Bubble skills and a love for helping people to join us as Technical Product Support Specialists.

Changes we made this month

This was a big month for our homepage:

  • We shipped a completely-overhauled Showcase page with a fresh design and a new set of case studies: we hope this becomes valuable ammunition for agencies and developers trying to explain the value of Bubble to their clients!

  • We also redesigned the display pages for templates. For instance, here’s what our most-installed template’s display page now looks like. Our goal is to make the page more readable and appealing, and to boost template purchases and installations!

  • Finally, we improved the experience for new visitors to Bubble on a mobile device. The Bubble editor is optimized for big screens, but we want someone who checks us out from their phone to be able to browse our homepage and get a good understanding of what Bubble is all about.

We also shipped a number of product improvements. While none of these are game-changing by themselves, they all address painpoints that have come up from our community, and we’re excited to keep improving the overall quality of our product:

We also shipped a number of behind-the-scenes stability fixes to the way our servers deal with server-side workflows that process an unusually large amount of data or consume a lot memory, eliminating a ton of bugs that were leading to occasional workflow and request failures. We had been seeing a bunch of process crashes caused by memory or timeout- related issues, and have now mostly eliminated them, which should decrease the overall unexpected error rates on Bubble.

Finally, we launched an official plugin for Phyllo, which connects to multiple social media platforms via a single API to support influencer marketing and the creator economy.

This month in numbers

Note: we are replacing “new conversations” with “handled tickets”: new conversations was including a lot of spam, whereas handled tickets is limited to things where we actually wrote a reply, which is a more accurate representation of how much work our support team is doing.

  • Tier 1 (FAQs, account and billing issues) handled tickets: 5,396 (down from 6,037)

  • Tier 2 (app development questions and bug reports): handled tickets 2,331 (up from 2,140)

  • Average tier 1 reply time: 31 minutes (down from 34 minutes)

  • Average tier 2 reply time: 2 hours and 23 minutes (down from 3 hours and 11 minutes)

  • Tickets closed by the engineering team: 128 (up from 124)

  • Average days to closure by engineering for high priority tickets: 11.4 (up from 5.8)

  • Average days to closure by engineering for all tickets: 13.3 (down from 15.5)

  • Incidents and regressions: 19 (up from 18)

  • Of those, the number that are high-severity (greater than 20 bug reports): 2 (down from 3)

Things on our minds

As mentioned in last month’s update, we are doing a push on reducing the number of regressions and increasing platform stability. That push is not reflected in the regression numbers yet (February was basically the same as January in terms of number of regressions), but we did make a few important changes:

  • As mentioned above, we shipped a number of stability fixes. Those don’t directly address regressions, but they make workflows less likely to have errors
  • We hired a new engineer, Mike, onto the Platform team, and have an accepted offer for another senior engineer joining us soon
  • We rolled out a formal on-call rotation. Previously, incidents were getting responded to by whoever was available, with the founders being automatically alerted if no one else on the engineering team was around; we now always have a rotating designated first responder and escalation system.
  • We introduced a mandatory postmortem process for regressions and incidents that result in at least 5 bug reports, as well as formalized how we track and follow-up to ensure that automated tests get written in response to regressions

Another thing on our minds is leveraging AI to super-charge the Bubble app creation experience. We are still in the exploratory stage, but we are now recruiting AI researchers for a 3 - 6 month residency this summer to experiment with what’s possible. Our long-term vision is that users can seamlessly switch between editing apps via drag-n-drop and via AI-interpreted commands, and that AI can automate big chunks of the app development process, making it lightning fast to get an app launched.

What we’re currently working on

Speaking of AI, we are exploring having an AI- and human- powered chatbot in the Bubble editor to provide live support: we need to confirm this is a great experience before rolling it out, but we are excited by the possibility!

On the product development front:

  • We are re-organizing the list of data sources and operators in the editor to make it easier to browse and use, which we expect to launch in early March

  • We are making performance and reliability improvements to the editor drawing surface. We have this built, and are seeing crazy-high speedups on editor actions for apps that have thousands of elements on a page, as well as more modest improvements to average-size apps. We plan to release this as an experimental feature in the near future, and then release it for everyone once we’re confident it is stable.

  • Building on top of the new editor drawing surface, we have a proof-of-concept of a Table element. Our POC demonstrates solutions to the technical problems involved (having multiple surfaces you can drag elements into, and having different sets of properties for different rows and columns); we’re now designing how we want it to look and function.

  • We’re working on moving our interactive lessons into the new application creation flow, so that new users can get in-app coaching on how to build various features. We expect to start user testing by the end of March.

  • We are making some updates and changes to the API for building server-side actions in plugins, both to fix some bugs and to enable upgrading to node 16, which requires switching from a synchronous to an asynchronous API. We’ve gotten feedback from plugin developers on our proposed new API, and are ready to begin building it.

  • Our version control improvements are still on track for a March launch: we are currently beta testing them

  • Replacing capacity with auto-scaling / overages and updating our pricing accordingly: we are in the process of collecting feedback, with the goal of sharing details more broadly in April

  • We are continuing to work on overhauling our infrastructure, with the goals of enabling SOC2 certification, improving security, and setting us up for future infrastructure expansion

  • Moving off CoffeeScript and onto Typescript. We are now down to 1.8% CoffeeScript in our main codebase. Typescript is now up to 12.0%, with the remainder being Javascript.

Happy building,

Josh and Emmanuel

52 Likes
2

8 Likes
3

Making any progress on updating the client-side actions API? I’m much more productive than an AI.

20 Likes
4

So many great updates.

But most excited about these two!

Excited!

Thanks
Zubair

4 Likes
5

Totally agree. Version control because… well it’s version control. A very welcomed addition.

And AI support will be very interesting - hopefully by 2024 we’re building workflows using a chat UI.

2 Likes
6

:+1: :smiley:

Totally stoked about the potential of AI. Will you be collaborating with Neuralink so that we can create an app just by thinking about it? :smirk:

:brain: :calling:

2 Likes
7

“AI” is the key for bubble to open all the close doors.

1 Like
8

@josh, now this is interesting, can you lift the curtain a bit more?

Best,

Gerbert
MVP Design

5 Likes
9

I have near zero interest in chasing the AI hype. I like AI, but first I want versioning, autoscaling, faster editor, and all those other things you’ve seen requested by actual users.

Thanks for all the work you guys do :pray:

14 Likes
10

This :100:

4 Likes
11

Totally agree.
What’s the point in having an AI if plugin client side actions and custom events can’t still return values?

7 Likes
12

Just use custom states client side and light data types server side for that (not perfect but it works).

If AI can reduce build time even more by interpreting my speech and create /edit datatypes, privacy rules, workflows, screens based on what I say, awesome.

6 Likes
13

I am aware of the workarounds.
They are workarounds and they slow the development process.

We all saw the increased flexibility and speed when multiple inputs were added to custom events.
Returning values would be a massive improvement, and would put bubble on par with competitors in this regard.

It would be funny to have bleeding edge AI but a crippled plugin system and workflow logic.
But I understand that integrating Openai API probably requires less resources than adding returned values to custom events.

I’ll keep going with the workarounds on bubble and enjoy the better developer experience when I work with other tools :man_shrugging:

8 Likes
14

To be honest, returning values are just a small thing because of the workarounds. Backend workflow debugging and missing log filtering options are the big things for me.

I’ve tried most of the No Code / Low Code competitors out there, and some of them provide a better developer experience, others offer more flexibility, and others offer faster development speed. But if you take these requirements together, the other tools never match Bubble.

Note: Another solution to have returning values for backend workflows is to use “Return data from API” together with the API connector to talk to the API’s in your own app.

8 Likes
15

I’m aware of this and all the other problems that you have with it :slight_smile:

That’s why we are still here and we want Bubble to improve with meaningful features :wink:

4 Likes
16

That’s great and highly expected @josh . Will it be an optional or a required change for Bubble users? Can Bubble please include me in the survey/feedback process as well. It will help us plan the pricing of our product accordingly.

1 Like
17

I am still stunned that no timeline has been given to make bubble GDPR compliant. As of today, if you are a European or UK user, the only way to make your app GDPR compliant using Bubble is to pay 4K/month for a European server, making it unaffordable for many if not most users.

9 Likes
18

@josh Thanks as always. I agree with what is noted above re AI. AI is another really cool marketing feature but there are some requested features/functionality, that I think most of us here consider “low hanging fruit”, and that would make our “development lives” much simpler. As @Keith noted above, there are those of us who are willing to do some pretty heavy lifting for the community if we could be given the proper tooling.

For example, I have for a long time now wanted to build a RoR like scaffolding tool such that someone can create a Data Type and have the CRUD pages automagically built (i.e.
rails generate scaffold LineItem name:string price:decimal)

Presumably the AI would have to interface with some editor APIs and you could just extend that to us?

4 Likes
19

Hey Josh and Emmanuel,

Thank you for sharing the March community update. It’s great to see the continuous effort in evolving Bubble editor, making the platform more stable and reliable, and increasing the investment in your ecosystem of partners and developers.

It’s good to see the product improvements and behind-the-scenes stability fixes that address community pain points. The ability to customize the flow for confirming a user’s email, adjust the opacity of an entire element in one setting, and use your own domain instead of Stripe’s domain on Stripe’s checkout page will undoubtedly help the community a lot.

was particularly intrigued by your plans to leverage AI to supercharge the Bubble app creation experience. The idea of seamlessly switching between editing apps via drag-n-drop and via AI-interpreted commands sounds fascinating. I believe that AI could also be used to automate other aspects of app development, such as generating logic/workflow for commonly used functions or even suggesting design elements based on user preferences and trends.

Thank you for sharing the community update, and I look forward to seeing what’s next for Bubble.
Have an amazing Day
Ali Imran
Linkedin | Youtube

2 Likes
20

Don’t tempt me

This is a change to how we implement it – there won’t be anything visible to users other than better performance and fixing some longstanding bugs. From a technical standpoint, we’re migrating it from very old jQuery code written in the early days of Bubble to Solid.JS

We’re keeping the AI research as a separate workstream from our normal product development for now, because we don’t want it to sap momentum from everything else we’re working on! I’m excited about AI long-term, but we’re exploring it carefully and will only ship stuff to our users if we think it actually adds real value: we’re not interested in AI-for-the-sake-of-the-hype.

It will eventually be required, but for current users there will be a very long transition period (> 1 year) to give everyone plenty of time to adapt. In terms of pricing your product, you can think about pricing a service built on AWS as a comparable – for almost all use cases, there’s a way of building it so it scales linearly with the total amount of activity on the platform.

We are aware this is a huge deal for our European users! Running a shared Bubble cluster in Europe is a tricky project from an infrastructure perspective because right now our US-based shared cluster is special-cased in many ways; we’re working to change our infrastructure such that we can spin up as many shared clusters as we want in any location we want, but it’s a journey and we don’t want to commit to a timeframe until we’re clearer on how the technical pre-work will go

Right now we’re just working on the server-side actions API, but we know the client-side one could use love too!

15 Likes