Bubble app slow? Nine fixes that actually work (and one that doesn't)

Bubble apps are fast when they are small. Around a few thousand records and a few hundred users, pages start taking five seconds and founders start asking whether they picked the wrong platform. Usually they didn't. Usually the app is doing too much work on every page load.

The nine fixes

1. Stop loading whole lists to count them. :count on a search that returns thousands of records pulls every record first. Use a dedicated counter field updated by a backend workflow.
2. Search server-side, not in repeating groups. Filtering a repeating group's data with :filtered runs on the browser after the full list arrives. Put the constraints in the search itself.
3. Cut the fields you fetch. Every text field, every image, every nested thing on a list item is downloaded. Use lighter data types for lists and load the full thing on detail pages.
4. Move heavy workflows to the backend. Anything that touches more than a handful of records belongs in a backend workflow, scheduled, not in the page.
5. Kill duplicate workflows. A button that triggers two workflows doing the same update is the most common cause of "it fired twice".
6. Index by structure, not by search. Store the relationship (a list of things on the user) instead of searching for things that reference the user.
7. Lazy-load below the fold. Groups that aren't visible shouldn't fetch data until they are.
8. Reduce plugin count. Each plugin adds scripts to every page. Remove the ones from experiments you never shipped.
9. Check privacy rules. Missing rules don't just leak data, they also change what Bubble has to evaluate on every search.

The one that doesn't work

Upgrading the plan. More capacity makes slow queries run slow in parallel. It does not make them fast. Fix the queries first; upgrade if you still need to.

When it is time to leave

If your product needs real-time updates for many users, heavy computation, or an investor wants to see code you own, the ceiling is real. That is a migration, not a fix, and it can be done without losing a user. But most "slow Bubble app" problems are a week of work, not a rebuild.

JetBuild has rescued fifty-plus Bubble apps. A rescue starts at a fixed $2,000 and most are done in one to two weeks.