I want to tell you about a conversation I have more often than any other.
A founder comes to me with a product idea. They have $8,000 to $10,000 to spend. They have done some research. They have decided they want to build an in-house team — a developer, a designer, maybe a project manager.
So I ask them to walk me through what that looks like.
By the time we finish that conversation, their $8,000 to $10,000 has become $50,000 to $70,000. And it has not even started yet.
Here is how it happens every time. A non-technical founder building their first product cannot manage a development team directly. They do not know enough about the work to review it, question it or direct it. So they need a technical manager or a business analyst to sit between them and the team. That person alone costs Rs. 8 to 15 lakhs a year. Then there is the developer — a good one costs Rs. 12 to 20 lakhs. The designer, Rs. 6 to 10 lakhs. And all of these people are permanent costs. They are on the payroll even on the weeks when there is nothing to build.
Two clients I worked with closely — the team behind LeSociety in Canada and the BISSAFE team in the US — both came to me at this crossroads. Both were considering in-house teams. Both decided to work with us instead. Both have told me since that it saved them a significant amount of money and, more importantly, months of time they would have spent hiring, onboarding and managing people instead of building a product.
This is not me telling you to always use an agency. I have told founders the opposite too. Here is the actual framework I use.
What each option actually costs — honestly
Let me give you real numbers, not the optimistic ones.
A senior full-stack developer in India costs Rs. 15 to 35 lakhs per year depending on experience and location. In the US or Canada, you are looking at $120,000 to $160,000. That number does not include the 3 to 6 months it takes to recruit them, the 1 to 3 months before they are actually productive, or the management overhead that comes with every employee.
A development agency charges differently. For a small startup project — a focused MVP with one or two user types — you are typically looking at $10,000 to $35,000. That is a one-time cost for a defined deliverable. When the deliverable is done, the cost stops.
Freelancers sit in between. Cheaper upfront, but harder to manage, harder to scale, and continuity is always a risk. They work well for specific, bounded tasks. They rarely work well for building a full product.
The right comparison is not monthly salary versus project cost. It is monthly salary multiplied by how many months you need the work done versus what you would pay an agency for the same output.
When an agency is the right call
Speed is the most important factor here.
An agency has a team that is ready to start. No recruitment. No notice periods. No onboarding. If you have a clear brief and funded runway, a good agency can have a working product in front of real users in 8 to 12 weeks. Building an in-house team to do the same takes 3 to 6 months before anyone has written a single line of product code.
For a pre-revenue startup, those months are not just a cost — they are a risk. The market moves. Competitors build. Your own understanding of the problem evolves. Speed to learning is everything at this stage.
Agency work also makes sense when you do not yet know exactly what you are building. A good agency will help you define the product, challenge your assumptions and push back on features that add cost without adding value. That's something an in-house team rarely does — because employees build what they are told to build.
When to build in-house
Once you have product-market fit — meaning real users, real retention and ideally real revenue — the economics start to change.
At that point you know what you are building. You are not exploring anymore; you are executing. You have a product that works and a team that understands it deeply. Hiring a developer who lives and breathes your product full-time starts to make sense.
The economics flip at roughly 12 to 18 months. A developer at Rs. 20 lakhs per year is significantly cheaper than paying an agency retainer for the same output. In-house wins on long-term cost — but only if that person is fully occupied and you are not still figuring out what you are building.
My advice: do not hire your first full-time developer until you can give them a full backlog of clearly defined work that will keep them busy for at least six months. If you cannot do that, you do not yet need in-house.
The hybrid model that actually works
The honest answer for most startups is not a choice between agency and in-house. It is a sequence.
Phase one, before product-market fit: use an agency. Build fast, learn fast, spend only what you need to learn. Do not hire permanent headcount when the product might pivot in 90 days.
Phase two, after you have PMF: keep the agency for new development while you hire one strong in-house developer who owns the codebase, manages quality and understands the product deeply.
Phase three, once you are scaling: bring most development in-house. Use the agency selectively for specialised work — AI features, security audits, big new product lines — that require skills your team does not have.
This is what the most successful founders I have worked with have done. They did not treat it as a binary choice. They treated it as a progression.
The question I always ask first
Before any conversation about agency versus in-house, I ask founders one question: do you have product-market fit?
If the answer is no — and for most early-stage founders it honestly is — then the priority is learning, not building. You need to talk to users. You need to test assumptions. You need to iterate quickly on a small product, not manage a permanent team while you figure things out.
That is not a conversation most agencies will have with you. Most agencies want to start building as soon as possible. The more they build, the more they earn.
I have lost projects by having this conversation. Founders who wanted to build a full product, heard my advice to start smaller, and went somewhere that would just take their money and start coding. Some of those products failed. A few came back to us later.
The ones who listened built better products, spent less money, and got to market faster. That is the part that keeps me having this conversation.