Every founder who has tried to build software in India has heard the same four pitches, usually in the same week. A freelancer on Upwork at $18/hour. An agency quoting ₹14 lakh for "the full app". A staff-augmentation vendor offering "senior React Native, available Monday". And a friend insisting you should just hire in-house.
All four can work. All four fail in specific, predictable ways. This post is the honest version of that comparison — written by a team that has run all four models, including the ones that went badly.
The four models, in one table
Monthly ranges below are all-in Indian market rates for 2026 — what a buyer actually pays, not what the engineer takes home.
| Model | Typical cost | You supply | Fails when |
|---|---|---|---|
| Freelancer | ₹40k–₹1.5L / month per person | Spec, design, QA, integration, project management | Scope is vague, or more than two freelancers must coordinate |
| Staff augmentation | ₹1.6L–₹3.5L / engineer / month | Tech lead, backlog, code review, sprint cadence | You have no in-house lead — capacity without direction just burns budget |
| Dedicated team / pod | ₹5L–₹15L / month for a 3–6 person pod | Product direction and priorities | You treat it as a black box and skip weekly demos |
| In-house | ₹1.2L–₹3L / month per engineer + 3–5 months to hire | Everything, permanently | You need to ship in under 4 months, or the role is temporary |
Why the cheapest hourly rate is usually the most expensive outcome
The number founders compare is the rate. The number that decides the budget is cost per shipped feature — and those two move in opposite directions more often than anyone likes to admit.
A freelancer at ₹700/hour who needs a written spec, doesn't write tests, hands over code with no documentation and disappears for a week in December is not cheaper than an engineer at ₹1,400/hour who is in your standup every morning. The difference shows up as your time: rewriting tickets, testing manually, chasing status, and eventually paying someone else to understand what was built.
A useful sanity check before you sign anything: ask how many hours of your week the arrangement consumes. Under two hours means you've bought delivery. Over eight means you've bought a second job.
What "senior" means in the Indian market — and how it gets abused
The single most common way an engagement quietly goes wrong is title inflation. The CV says senior. The person has three years of experience, two of them on a codebase they didn't design.
Three questions that surface this in ten minutes:
- "Walk me through something you shipped that broke in production. What happened?" Genuinely senior engineers answer this fast and without defensiveness — they've all got one. Juniors describe a bug; seniors describe a system.
- "What did you argue against on your last project, and did you win?" Seniority is largely the willingness to push back. Someone who has never disagreed with a product decision has never owned one.
- "Show me the repository." Not a portfolio site — commit history. Cadence, review comments, and how they handle merge conflicts tell you more than any interview.
Fixed-price or time-and-materials?
This choice causes more disputes than pricing does, and the rule is simpler than vendors make it sound.
Fixed-price works when the scope is genuinely known — a clone of an existing flow, an integration against a documented API, a redesign of screens that already exist. You get budget certainty. You pay for it with change-request friction: every deviation becomes a negotiation, and the vendor is financially motivated to say no.
Time-and-materials works when you are still discovering the product. You keep the right to change your mind weekly. You pay for it with the discipline of actually reviewing burn — T&M without a weekly demo and a visible burn-down is how six-month projects become eighteen-month projects.
For a first MVP, we quote fixed-price against a scope we've written down together, with a named change-request rate. It's the combination that survives contact with reality: certainty on the core, a clear price on the surprises. Our published MVP tiers run ₹2–4L for a smoke test, ₹5–10L for a seed-stage build and ₹10–20L for a growth build, typically delivered in 6–12 weeks.
The five questions that separate real partners from resellers
- "Who exactly is on my team, and can I see their commit history?" A partner answers with names. A reseller answers with "a senior React Native resource".
- "Who owns the repository from day one?" The only acceptable answer is you. If code lives in the vendor's org until final payment, you are one dispute away from having nothing.
- "What happens when a milestone slips?" Every project slips. You want to hear a process — re-scope, re-baseline, written notice — not a promise that it won't happen.
- "Can I talk to a client you didn't pick for me?" The reference list is always curated. Asking to go one name off it is revealing.
- "What does support look like the week after launch?" Launch week is when the real bugs surface. If post-launch is a separate contract nobody mentioned, budget for it now.
How we're set up, and where we're honestly not the right fit
Xenotix Labs runs dedicated pods and staff augmentation out of India, with 37+ engineers and 110+ shipped products across 6 countries since 2020. We work IST-aligned with an overlap window for US and UK clients, and we hold a 4.7★ rating across 76 Google reviews.
Where we're not the right call: if you need a single freelancer for a two-week task, a pod is overkill and we'll say so. If you want the lowest rate on the market, someone will always go lower. And if you have no internal product owner and no intention of appointing one, no staffing model fixes that — including ours.
Pick the model, then pick the vendor
Most bad engagements start with the vendor conversation instead of the model conversation. Decide first whether you're buying hands (augmentation), delivery (pod or agency), or capacity you'll manage forever (in-house). The right vendor for one is usually the wrong vendor for another.
If you want a straight answer on which model fits your stage, tell us what you're building — we'll say if it's us, and if it isn't, we'll tell you what to look for instead.
