[New Thread] More context on our changes to pricing - FAQs answered

[keith]

@dee what you are seeing in your app is exactly what I found in mine – that under this new pricing system, API calls are ridiculously priced (but in your case you’re doing 100x more calls than I am). If you want to understand my analysis (forgive me if you’ve seen this), it’s here.

As @gaimed has best observed, in general, the pricing for WUs (or perhaps the calculation for WUs - same difference) would seem to be off by about a factor of 10. In my own example, it seems Bubble wants to charge me about $300 for making the API calls my app needs to make (and, like you, I can’t “optimize” my way out of that), when I would be willing to pay $30 for that and Bubble still makes an incredible markup on those calls. So I guess in general, @gaimed is correct.

There’s so many messages in these threads that I haven’t seen them all. But as @TipLister observed, it’s quite ironic that Bubble did that big show-and-tell webinar with a few folks, like you, who have apps that use OpenAI’s API. I wonder how they are reacting to all of this. Does anyone know if any of them have chimed in here? They’ve gotta be pooping bricks right now.

I hate how this thread is throttled, so I’m atting several different folks. @bubble96, yeah, you as well cannot “optimize” yourself out of this. The new pricing is simply bonkers.

@lottemint.md, like you I was psyched for autoscaling so when I looked at my app yesterday I was doubly-bummed. But I can’t imagine anyone building anything on Bubble (regardless of how deep their pockets are) as it’s immediately obvious that the pricing is nuts and that indeed, nobody could scale anything under this system.

It’s just completely weird that Bubble isn’t addressing these specific pricing concerns directly.

Another thing I haven’t seen anyone address is that the new WU chart is very interesting not just for assessing future costs, but it very clearly show you what you’re paying for now. (At least, if we can trust that the WU model really syncs up with what things are costly and what are not.) For example, in the case of my app I’ve been talking about, the WU model implies that 90% of what I’m what I’m paying for now is those 2 workflows that I detail above. I guess I’m paying something like $180 today or thereabouts?

So, I guess Bubble thinks that my external API calls are costly (I guess I’m paying $118 per month for the luxury of pinging my GCF endpoint, and I pay Google 3 cents for the same), and $54 to validate if someone is subscribing to my app, which means that I’m paying $8 a month for everything else. That’s a really remarkable price for web hosting, wouldn’t you say? As every single API call I make more or less corresponds to the number of page loads that my app does.

Of course, all this shows is that the WU model is completely silly. I mean, c’mon.

[JGalt]

@brian.good500 completely agree - 10,000%. The abundance of well-made third-party plugins (some of which do a vastly better job at certain functions than Bubble’s own native equivalents) is what is allowing my current project to even be feasible. The level of self-sabotage being displayed here is unreal.

[dushdan123]

Bubble is becoming that one girlfriend that keeps cheating on you but you don’t wanna let go

[aaroncohn]

@tatiana.a oooof — this response isn’t really a great look nor a solve for most folks

• Most folks are still pretty confused on what makes up a WU
• When you look at the WU numbers, most people are getting sticker shock that the price they pay is going to 5x to 10x+…even after painful optimization.

To the team — @emmanuel and @josh — I do think there’s a real willingness for Bubble users to pay more.

And there are some good, reasonable solutions. I think you’d keep most users happy if everyone’s price went up 10% to 25% per month — but not 4x to 10x.

If I were on the Bubble leadership team, here’s the solution I’d pivot to:

• Raise the prices of each plan 10-20%
• Significantly increase the WU allotment on each plan, at least 3x to 5x

Computing power should be cheap. That’s what’s frustrating folks. And that’s where I think the team missed the mark.

[artemzheg]

So, I’d like to share more results of my tests for some basic simple searches and things creation:

Brand new App #1 [testing cost of fetching a single item from the database]
(spoiler: it’s not ~0.2 WU)

test data and results

Data Types: “User” (no custom fields). Only 1 user in DB (one, only one, I’ve checked it)
Page: 1 group element on a page [type of content = User, data source = Do a search for Users:first item
Workflows: 0
Tests performed: 4 tests
Results:

1. First test: total WU = 15 (page load 10.8, fetch data = 3, WF = 0, API = 1.2)
2. Subsequent tests: total WU = 6 (page load ~2.5, fetch data ~2.8, WF = 0, API ~ 0.7)

Conclusion: fetching a single item from the database doesn’t have a base cost around 0.2 WU, it’s x14 more expensive (2.8 WU)

Brand new App #2 [testing cost of fetching 5 lightweight items from the database]
(spoiler: it’s still not ~0.2 WU per item but cheaper than in App #1 results)

test data and results

Data Types: “User” (no custom fields). 5 users in DB
Page: 1 Repeating Group (no rows limit, “show all items immediately” checked) element on a page [type of content = User, data source = Do a search for Users(no constraints)
Workflows: 0
Tests performed: 4 tests
Results:

1. First test: total WU = 14 (page load 5, fetch data = 6, WF = 0, API = 3)
2. Subsequent tests: total WU = 9-11 (page load ~2.5, fetch data 5.5-6, WF = 0, API 0.6-2.6)

Conclusion: 1.1-1.2 WU per 1 item fetch which is cheaper than 2.8 WU from Brand new App #1 test results.

App #3 [testing cost of empty page load]

test data and results

Data: not relevant for this test
Page: empty page, 0 elements
Workflows: 0
Tests performed: 5 tests
Results:

1. First test: total WU = 32 (page load ~6, fetch data ~25, WF = 0, API ~ 1.)
2. Subsequent tests: total WU = 3-4 (page load 2.1-2.7, fetch data 0-0.4, WF = 0, API 0.4-4)

Conclusion: 25WU to fetch data on an empty page during 1st test?

Brand new App #4 [testing cost of “Create an account for someone else”]

test data and results

Data Types: “User” (no custom fields)
Page: 1 button triggering a WF
Workflow: Single WF with 1 step → “Create an account for someone else” action
Tests performed: 3 tests
Results:

1. First test: total WU = 29 (page load ~6.4, fetch data ~3.6, WF = 18, API ~ 0.9)
2. Subsequent tests: total WU = 23-24 (page load 2.3-4.3, fetch data 0-1.2, WF ~18, API 0.5-0.8)

Conclusion: “Create an account for someone else” is more expensive than “Sign the user up” action (the last one cost is ~ 16 WU, details here).

Brand new App #5 [testing cost of creating a single new Thing in a client-side WF]

test data and results

Data Types: “Thing” (no custom fields)
Page: 1 button triggering a WF
Workflow: Single WF with 1 step → “Create a new Thing”
Tests performed: 5 tests
Results:

1. First test: total WU = 26 (page load ~11, fetch data ~1, WF = 13, API ~ 1)
2. Subsequent tests: total WU ~17 (page load 2.3-2.5, fetch data 1-2, WF ~13, API ~0.7)

Conclusion: creating a single Thing (that has no custom fields) costs ~17 WU in total or ~13 WU in WF workload

Brand new App #6 [testing cost of creating 2 new Things in one client-side WF]
(spoiler: nope, not “create a new thing” x2)

test data and results

Data Types: “Thing” (no custom fields)
Page: 1 button triggering a WF
Workflow: Single WF with 2 subsequent steps → “Create a new Thing” + “Create a new Thing”
Tests performed: 5 tests
Results:

1. First test: total WU = 29 (page load ~6, fetch data ~0.9, WF = 21.5, API ~ 1.2)
2. Subsequent tests: total WU 25-31 (page load 2.4-6, fetch data 1-2, WF ~21, API ~0.6)

Conclusion: I thought creating 2 new things will cost x2 of creating a single thing, but it’s a little bit cheaper.

Brand new App #7 [testing cost of creating 20 new Things in one client-side WF]
(spoiler: bulk sale)

test data and results

Data Types: “Thing” (no custom fields)
Page: 1 button triggering a WF
Workflow: Single WF with 20 subsequent steps → “Create a new Thing” x20
Tests performed: 3 tests
Results:

1. First test: total WU = 184 (page load ~5.5, fetch data ~5, WF = 172, API ~ 5)
2. Subsequent tests: total WU 180-185 (page load 2.4-7, fetch data ~4.5, WF ~172, API ~5.5)

Conclusion: ~9 WU in total (or ~8.6 in WF workload) per 1 thing creation

[TheDaniel]

Fun sidenote - I’m one of the “users who wouldn’t have to pay more” at my current WU usage. If the price changes stay, my business is still dead on Bubble. Why? Because I haven’t launched yet! If I scale to even a couple of users, I will start losing money.

So even though I would show up in the “green zone” of Bubble’s pricing analysis, I have to leave the platform if these prices stand. And I am certain there are many more in the same scenario.

I’ve seen the narrative of it only being the 8% heavily affected posting here about not being able to stay, which is untrue.

[OliHaydes]

For once - please listen to your community. Following through with these changes will single-handedly kill your platform and I can’t see many people who will stick around with this.

[gazinhio]

Unless she comes back begging for forgiveness in the next few days then you don’t have a lot of choice but to dump her, too high maintenance.

[axeldor67]

You forgot to add this thing for a next feature.

[jobs]

APRIL 2023 BUBBLE TUTORIAL:

1. Never use backend workflows. (This is the only way you can securely do many things on the web… but just change your business model.)

2. Cut Features. This isn’t really bubble related but if you’re thinking of doing something with data just don’t. Make it simple.

3. Switch your business model so you can make $10+ per user. Each user on bubble will cost you at least a few dollars as proven below.

4. Hire a WorkLoad Optimization expert like @ojnhoi who can help you keep your retirement savings untouched.

A serious talk about costs:

So I finally gave in and reviewed the logs. I found the issues with the optimization. (Or rather, the things Bubble has determined to de-incentivize thru the means of holding my wallet ransom.)

But just finding the issue didn’t do workloadofSHI- for me!
My user signup (takes 2 minutes) went from 2,159 WorkLoad to 2,309 WorkLoad per user.
But I don’t like to use workloadofshi’s as a reference. I’ll repeat myself in terms of COST.

My user signup (takes 2 minutes) went from $0.39 to $0.42 per user.
5 minutes on my app for 1 user cost me $2.06

This is based on worst case scenario of 0.0018 per WorkLoad
Lets use the best case scenario of 0.0008 per WorkLoad

5 minutes on my app for 1 user cost me $0.90

My app hit 3% Capacity during this test.

So to compare CAPACITY VS. WorkLoad

Capacity could allow 33 users online 24/7 for $29/month.

WorkLoad : could allow 1 user online for 88 minutes each for $29/month.
WorkLoad : could allow 1 user online for 5 hours each for $99/month.
WorkLoad : could allow 1 user online for 18 hours each for $299/month.
WorkLoad : could allow 2 users online for 22 hours each for $599/month.
WorkLoad : could allow 6 users online for 20 hours each for $1,499/month.

It’s the same $29 app but the new method can’t even come close to what it can do today.

[lucas023]

Mine updated too. 2k WUs to fill out a multistep form

Goodbye Bubble and the idiotic leadership team. Great way to cut off your entire user base. I am looking forward to the day that Bubble becomes open source. Until then…

[dorilama]

From my initial tests:

• loading an empty page: 2.2WU
• creating a new thing with 2 text fields: 8.4WU
• do a search for thing: first item: field (15 things in database): 2.1WU

10 runs per test, new private browsing window every run.

[rafaelfernandes]

@tatiana.a @emmanuel

I was here thinking.

Will you instead be out there thinking of a better way to capitalize on our trust in you, WHY you don’t try to solve several system problemas in your back log?
In case, for example, why…in years, no one fixed the dev screen tree view panel problem?

Then you will tell me:

• There was no time? Do you need another 3 years?
• There was you will give me: Nobody abiur called in support. I open!Actually Emmanuel himself knows, everyone knows.

Clean up your mess before you invent a way to bankrupt our companies.
**

Por favor, MAIS RESPEITO COM TODAS AS EMPRESAS!

**

**

It is very clear that their ways of thinking are to offer something that looks beautiful to take away what really matters. You guys are basically selling us more tethers to your system! A future where many companies will let go of the day they chose to partner with Bubble.

**

What you are doing is shamefully unethical.

[artemzheg]

@dorilama

creating a new thing with 2 text fields: 8.4WU

That’s an interesting difference. I’ve already seen something around ~15-17 for 1 thing creation somewhere on the forum.
How did you optimize it?

[aj11]

There is a current experimental feature that fixes the element tree issue.

[dorilama]

@artemzheg

How did you optimized it?

It’s a confidential secret that I will disclose in a webinar in 2 weeks. I can’t tell much now

Jokes aside, that is just the raw numbers of the single operations.
A real word app would have much more WU:

• load empty page: 2.2 WU
• create thing: 8.4
• if cookies are on automatically retrieve current user: 2.1 WU (I suppose it should be similar to do a search for)

It’s already 12.7WU for an empty page and a single click on a button.
Of course it’s not a proper test with only 10 runs, and we shouldn’t be here wasting time reverse-engineering an opaque pricing metric

[giosegar]

Exactly. Just like how in video games, 90% of people don’t get the first achievement that comes from finishing one level. Either they’re completely ignorant of the fact that the 8% represents their actual customers and power users, or they’re willfully using the statistic to twist reality.

[italogmoura]

The Unfair and Deceptive Acts and Practices (UDAP) law in the United States is a federal law that protects consumers from unfair and deceptive business practices. The law is enforced by the Federal Trade Commission (FTC), which has the authority to investigate and prosecute companies that violate the UDAP.

The UDAP law prohibits a wide range of unfair and deceptive business practices, including false or misleading advertising, false or misleading statements about the quality or effectiveness of products or services, aggressive sales tactics, hiding important information from consumers, unauthorized or hidden charges, among others.

The UDAP law is a federal law, but many states also have similar laws that protect consumers from unfair and deceptive business practices. Additionally, companies can also be held liable for UDAP violations in class action lawsuits brought by groups of consumers.

Enforcing the UDAP law is very important to protect consumers and ensure fair competition in the market. The FTC is a very active agency in enforcing the UDAP law, and companies that violate the law can face significant fines and other sanctions. Therefore, it is important for companies to be aware of their legal obligations towards consumers and to practice ethical and fair business practices.

[xtechaus]

Danm. I thought they would apologise and say they made a big mistake, didn’t expect them to double down on the changes💀

[rafaelfernandes]

So I believe entrepreneurial friends from the US need to come together to identify someone specialized (like a lawyer), so that he/she can create a real bridge of trust between us (Bubble Clients) and the Bubble company, considering an integral alignment with such law.

If US entrepreneurs don’t act, we’ll just be here complaining. At least at this point.
I believe that someone should take this issue seriously.

I’m currently here in Brazil talking to several entrepreneurs and, it’s unanimous: Everyone understands that the imposed model is completely absurd:

1. Either we accept a critical financial situation, 2) Or we accept the bankruptcy of our companies.

I don’t think all the entrepreneurs here are rich enough to say, OK, fine!

The word unfair is quite correct in this case.

Very good!