I think this is a perennial problem for people transitioning into/out of freelancing. Jonathan Stark has a whole YouTube channel largely dedicated to this problem - “Ditching Hourly”. e.g., see 2 of his latest episodes
He argues that the best model is “Value Pricing” (iirc), but also goes into depth on the pros/cons of subscription models
Well that’s the other benefit of price → scope rather than scope/hour estimate → price. You can expand/reduce the scope based on how much you know the client is willing to pay.
If the client is really fast at approving work, are you turning around multiple requests in a single day? Or is it more like request → implement → wait 48 hours (minus implementation time) → feedback → revision or next request?
In other words, how do you avoid having your calendar dominated by a single needy client that is hyper responsive and has a long list of requests?
Yes
It’s not a strict 48 hours either, if a client has a queue I may do a week’s worth in a day - as long as the client is fine with that and is getting good value out of what they paid for all is well. It’s nice to have the flexibility there.
Hi! I agree with you that it is possible to build HIPAA complient apps on Bubble. Your arguments make much sense. Do you send the user’s email & name to some external database to generate a secure token and fake email - and then register the user in Bubble using the fake email & token? Or you somehow hashing emails? What plugins or services do you recommend to use?
I wanted to leave some reflections/what went well because I’ve spoken to a few devs this week and lots were asking about how to develop their own career or agency.
I enjoyed reading this post because the agency has come a long way in the last 18 months. It’s a little nostalgic reading, and time flies. The agency has created so many opportunities for me personally and the people I get to work with, and for that, I’m immensely grateful and hope I can pay it forward for future devs.
FWIW, I don’t think I’m a great person to ask how to run an agency. I’m not a particularly good manager, and I don’t have nearly enough experience to have learned how best to handle anything that might come my way. I’m also not interested in building a mega-agency, so that’s not the goal. The goal was to build an agency that legitimately delivers good things for clients, good things for the community, and earns well. Essentially, a business where everyone, on all sides wins.
No Bubble MVP has any business being more expensive than $25K
Have a great day 
This is something I never fully understood as ‘what is a task’ or ‘what is a feature request’. For me I can see a feature request as simple as ‘make the email send 15 minutes after the meeting ends’ up to ‘I need to integrate a payment feature that has 3 different payment flows’. How do you quantify for clients what qualifies as a single feature versus what they request as needed to be considered multiple features so as to set the expectations for the different tiers you offer or how many features are in a Rocket Launch MVP?
This right here is a sign that you are better at managing or teaching than you might give yourself credit for. The fact that you were able to manage a situation to reflect and change course is the sign of a good manager, and that you learned from it shows you taught yourself. It is important for people to reflect on their strengths and weaknesses and find ways to lean more into their strengths and get others who can help fill the gaps.
The way we convey information to others, surprisingly, can have drastic effects on our bottom line. I think about this similar to how the difference between ‘customer favorite’ versus ‘recommended’ on a menu can drive more sales, and sometimes, just making things simpler will drive more sales.
Love this.
Overall George, this is the best post of yours I’ve ever read. Shows a lot of growth. Very impressed with you and what you’ve done. Keep it up and keep taking some intermittent breaks to reflect and if necessary change course a bit.
3 tasks (rule of thumb, every ‘and’ or every ‘view’ is probably a different feature). To be fair, a task is whatever we want it to be. But if we make tasks super tiny then clients won’t work with us and virtually all of our business is those subscriptions. So, we’re incentivised to just do a good, fair job for everyone.
They start on the base $1499/mo plan to get an idea and build trust and then upgrade soon after. MVP build isn’t feature based, and scope size depends on client, project, whether I would enjoy building it etc. $1499 filters out the projects we could never take on, but is also low enough that a business owner is comfortable taking that risk for a month.
Thanks, I appreciate that. But I think rather than being a ‘good manager’, it’s more that I accept what I’m not good at (management) and that’s okay.
Thanks so much! It’s great to hear that.
at this point, very interesting and inspiring. I like to read you usually, and also now… I think I’m not quite ready to be a good developer, althoug I’ve developped some app, and one of it is very complete. I’m waiting some month more and I will pass the certification… in any case, thank you for writing and inspiring
First of all, great post @georgecollier — I spent most of the day reading through the entire thread to better understand how you’ve structured your pricing and delivery model. Really appreciate how transparent and thoughtful the whole breakdown is.
I’m a software developer with 8 years of experience, and I’ve been working with Bubble for the last 2. I mostly work in the no/low-code space now, and I’m aiming to get to a point where I can confidently charge $10–20K for full MVP builds, like you do.
A couple of questions came to mind that I couldn’t fully find answers to:
How do you manage to deliver an MVP in 2 weeks without a full scoping or design phase? I get that the scope is shaped to fit the budget and MVP mindset — but apps still usually involve some back and forth, feedback loops, and edge cases. Especially without using Figma or upfront designs, how do you keep things moving fast while still hitting expectations?
Regarding your ongoing subscription model (e.g. 2 tasks/week): Let’s say a client submits 4 tasks — 2 small bugs and 2 features. If one of those bugs takes 10 minutes and another task is a simple text change, do clients ever push back and question if they’re really “full” tasks? How do you set and maintain expectations around what counts as a task?
Appreciate any thoughts you’re willing to share, and again, thanks for building in public like this. It’s incredibly valuable for the rest of us trying to build in this space.
We’re going to be changing this soon and will start doing clickable prototypes. We haven’t focused much on new builds in the past but seems there’s a big opportunity there and we’re well positioned to actually do a good job at a fair price.
If it’s a text change it takes us longer to track than it does to just do it… so we don’t really track them. It all tends to work out on average. Clients can cancel at any time, because we want them to work with us because they want to, not because they have to. That means that if we do silly things like split tasks up into miniscule things, they’ll just leave.
Like, the business model literally boils down to do a good job, don’t try and cheat clients or devs, and work with people who will by a fair price for a job well done and a reliable partner.
[quote=“incomdies, post:2, topic:284974”]
Great post, George. Clean and clear
I sent you a DM. Plz check
That makes sense. Here’s how I understand your business model:
Work with clients who value your expertise and are open to your guidance. The key is to consistently deliver high-quality work so that clients trust your recommendations and never feel misled.
For example, if a client assigns two tasks for the week and you believe they’ll take longer than expected, they should be willing to break them down into smaller milestones based on your input—because they trust that you’re acting in their best interest.
Ok, it’s been another 4 months. I’ve graduated since then, so that’s cool. Here’s some more reflections that any other current or aspiring agency founders might be interested in.
Over the last few months, we’ve been trying to reposition our offering to focus on business value rather than development.
When you market yourselves as a developer, people price you as a commodity and compare to other full-time developers. They don’t realise the depth of what we know and how we make the problems these apps face go away. Our clients either run their entire business on Bubble, or their Bubble app is their business.
So, it’s a high value problem that’s expensive for them to not fix.
‘We develop Bubble apps’ is way harder to sell than ‘We stop your Bubble app holding your business back’. We’ve updated our language online and in sales to reflect that.
We’ve raised prices too. A long time ago, our minimum engagement was $749/mo. Now it’s $3k/mo. New build was originally $10k, now it’s $25k. The core offer hasn’t changed, but how we position it has, and the more demand you have, the higher you have to raise prices to keep a balance, because ultimately, we can only handle so many clients and have to prioritise the one that wants it most.
Interestingly, the best way to filter out difficult clients is to raise prices. The ones paying the least are the most demanding, and the ones paying the most are the most reasonable.
Our original client portal (getorchestra.com) was good, but became a liability for us. I still recommend it for freelancers/small agencies. But we have thousands of tasks in our portal, and it had a lot of performance issues at that scale.
We’ve moved to Linear. Each client has their own team (so each client can only see their own team). We pay $14/mo per developer and per client user. So, it’s not cheap, but it’s worth it for them to make the most out of us. Happy to share any findings / learnings from moving our client portal to Linear (given that it’s designed primarily for in-house product teams).
Man, Stripe was eating 5% of our revenue. 3.25% for international card payments, +0.5% for tax, +0.5% for billing, and 1% for paying out in USD because we’re registered in the UK (they’re not even converting the currency, they just want that extra fee as we’re not using their rip off currency conversion).
5% is a lot. But, it was worth it because of the convenience of paying by card and recurring payments. You don’t want clients to have to manually pay you every month because 1. you have to chase them up and 2. they re-evaluate working with you every month.
Now, we’ve moved to GoCardless. This is a bank-first payment platform that supports recurring bank payments. Its fees are 2.25%. We can have recurring payments, with lower fees. For very high value transactions, we bill manually via bank transfer so there’s no payment fees. But 2.25% is a fee worth paying for recurring payments the client doesn’t need to think about.
We have a lot of stuff on Airtable, but we reach limitations there because of our unique processes. We now have a custom internal portal built with traditional code. This manages billing, invoicing, and progress tracking (so we can remain on track with our clients). It means we can have something custom to our processes, and does not cost us anything substantial. Next.js + Supabase stack, for those of you that are no doubt curious. It’s also just useful to continue to assess the state of AI-assisted coding. We’ve done that once with Buildprint, so it’s interesting to see the progress. I think it would’ve taken a similar amount of time to build on Bubble, for what it’s worth.
Turns out social media is actually pretty good for attracting clients. Most of our clients are already in the Bubble ecosystem and know about us (we work on existing apps mainly). If they can see we’re an authority, it generates warm leads for us. Trying to improve my X and LinkedIn game
Hired three new team members! I love being able to get excited about working with new developers and stuff’s going great. We’ve paid out $2k in referral bonuses! Kind of a bargain for the quality we hired 
There really is a shortage of great Bubble talent. If you’re very good, and earning < $100k, you should be earning much more.
One of the markers I have now picked up on is that good developers can think through a solution and identify issues without even going through the editor. They can think ahead every possible issue they might encounter, and have a plan to deal with it before they hit it.
In the last three months, we doubled our MRR as a result of the above efforts. We now have a short waitlist which is awesome, and are continuing to get good leads in. All organic, from forum, socials, referrals.
For me, I began spending too much time working in the business rather than on it. In the weeds with admin crap. Now we have a new custom portal a lot of that will be taken away. In addition, I’ve further reflected and realised I actually just enjoy development. I’m involving myself closer on all client projects now (even ones I don’t own) because that’s where I deliver the most value and get the most value.
One of our clients has faced a delay with their development taking longer than expected. The root cause was bad estimation. But it was compounded by it not being communicated particularly well. Most business owners accept delays when they happen, but they need to understand the cause and, most importantly, when it will be resolved by. The second part never really happened. So, we’re going to get better at that.
Cold outreach is kinda tricky. A couple of devs I know have tried it and tried scanning apps for security issues and informing them, but the fact is that they ignore it and don’t particularly care. It’s pretty much a waste of time. It’s the proactive leads that think they have an issue and reach out to us that are nice to work with because they appreciate the importance of the issue. It’s better to make intentional messages to quality leads (rather than automation at scale). When they open your message, they’ll check your socials which is why authority on social media is important. We haven’t needed to do much cold outreach; it’s just something we can switch on if we were ever short of leads.
We hit a big ARR milestone recently, and set a new target of a nice MRR milestone by year end. Team is now 7 people including me, once we get to 10 we’re not going to hire anymore.
I’m gonna release a practical guide to Bubble soon, which contains maintainable and secure solutions to a bunch of common Bubble problems. I think the community will find it useful.
Impressive read! I like the analytical approach.
This is great. Awesome you take time to self reflect; important in so many aspects of life, not just business. Keep up the good work!
Thanks to share what you have done and how your business is growing.
However, one big question come to my mind: Why he is using so much external tool where he could build it with Bubble. And now, from what I understand, you are probably more focusing on differents solutions and not just Bubble?
Can you explain what is for you “great Bubble talent”? I have great skills (you know me mostly to answer API question, but I have a lot more skills) but I’m bad at design (however, I can replicate mostly any design…). So I always wondering what make a google Bubble dev for you? Because you have an agency, I think that you need to have skilled dev or designer who may have different skills but who cover all development needs (team work), no?
I’d rather spend 200 hours building for clients than replicating a perfectly good portal from Linear that has a whole team behind maintaining it
This isn’t what makes a good one, but often indicates a good one:
Because it means they 1. know Bubble inside and out 2. can plan ahead 3. can hold a mental model / abstraction in their brain
Thanks! Hope you continue to have success and have you here for a long time 
You provide a lot of value to Bubble ecosystem!
Read your original post a while ago and the forum suggested it to me again. Great to read your learnings this far @georgecollier. Thanks for sharing!
