How long does it take to build an MVP? Realistic timelines by type
Most MVPs I build take 45 to 60 days from the agreed scope. Here is how long my own builds actually took, from about a week for a one-feature tool to about 8 weeks for an internal web app, what slowed them down, and why a scoping sprint shortens the calendar.

Senior Shopify and web developer
The short answer
- Most MVPs and web apps I build take 45 to 60 days from the agreed scope.
- ShopMyESIM, a comparison site with traveller accounts and an admin console, had a working first version after two days of build and went live about six weeks later.
- BrandDesk, an internal web app for NutriStar that reads 100k+ orders from Shopify, took about 8 weeks to build.
- The calendar is decided by how fast decisions, content and real data arrive as much as by build speed. A scoping sprint settles those before the clock starts.
Want this done rather than explained? see my published prices
Most MVPs I build take 45 to 60 days from the day the scope is agreed. A one-feature tool can be done in about a week. A product with accounts, an admin console and live data from other companies takes the full range. The biggest variable is how quickly decisions, content and real data reach me.
How long my own builds took, by type
These are real durations from builds I've done, with dates, so you can check them against the work pages and build logs.
| Type of build | Real example | How long it took |
|---|---|---|
| One-feature tool | Alter, a macOS app of my own: copy text, press a key, say what to change | Built in about a week, 3 to 10 June 2026 |
| Comparison site with accounts and an admin console | ShopMyESIM, for a client | First working version in two days (14 to 15 Sep 2026). Live on 26 to 28 Oct 2026, about six weeks later |
| Internal dashboard on business data | BrandDesk, for NutriStar | Built in about 8 weeks, 2 Sep to 28 Oct 2026, 120 commits |
| Finishing someone else's stalled build | Nicobar's custom Shopify theme, with 30 to 40% of the work left by another agency | Launch-ready in about 2 months |
Across these, the build itself is rarely the long part. On ShopMyESIM, the first version ran after two days. The weeks after went into the work that only happens once real people and real data are involved.
Where the calendar went on one build
ShopMyESIM compares eSIM data plans for travellers. Its timeline, from my notes:
- 14 to 15 Sep: the first version. The public site, traveller accounts, the admin console and deploy tooling.
- 5 to 12 Oct: design. A design system, destination photography, polish and a redesign. The design went through three iterations and ended in a full rebrand to the ShopMyESIM name, logo and palette.
- 13 Oct: hardening. I traced page hangs and intermittent 500 errors to the database connection pooler, measured 15 of 18 requests hanging in transaction mode against 0 of 18 in session mode, and moved the serverless functions into the same region as the database.
- 26 to 28 Oct: real data and launch. The live plan data went in, the new brand went on, the site went live, and I handed it over.
- 3 to 10 Nov: growth features after launch. Email capture, trip reminders, a phone checker and a sticky trip bar.
What slows a build
When I write a scope, I give a range of days and write down what decides which end we hit. A build lands at the short end when:
- the brand direction is settled in week one;
- the real data arrives by an agreed day;
- each milestone is reviewed within one working day.
It lands at the long end when reviews run longer, the data arrives late, or partner links are still being negotiated. Every one of those is a decision or an input on the founder's side.
Three more things that stretch a calendar, from other builds:
- Measuring after launch. BrandDesk shipped with reorder lists, dead stock, real margins, competitor prices and a weekly founder brief. When I measured what the founders used, it was one feature, the weekly decision list. So I rebuilt the home page as a daily to-do list that closes items itself. Leave room after launch for a change like that.
- Problems that only show up in production. The page hangs and intermittent 500 errors on ShopMyESIM were a production problem, traced to the database connection pooler. A day of hardening fixed them. Leave room for one.
- Inheriting someone else's code. Nicobar's theme had stalled with another agency, 30 to 40% short of launch. I read their codebase first and picked up where they stopped, without a rebuild. Getting it launch-ready took about 2 months.
Why a scoping sprint saves time
The slow part of an MVP is deciding. A scoping sprint moves that to before the build clock starts, so the weeks you pay for are spent building.
What the sprint settles:
- What's in and what's out of version one, with a price on each cut so it can come back later without a new negotiation. Here is how a scope becomes one fixed price.
- The inputs I need from you, with dates. Brand direction, content, data files, partner accounts. Each one has a date in the scope, and the scope says what happens to the launch date if it slips.
- The data model. Where every piece of data comes from and who supplies it.
- How changes work. Anything that adds a system or changes the data model is priced before it's built, so the scope can't creep quietly.
I built a Claude skill, Enough, for the cutting part. It cuts a product idea down to the version that ships and writes every cut, with its reason, to a file in the repo.
The sprint costs ₹40,000 in India or $750 internationally, and it's credited to the build if you go ahead.
What to do next
Before you talk to any developer, have three answers ready:
- The one flow a user must be able to finish in version one.
- Who decides. One person who can approve a milestone within a working day.
- Where the data comes from, and when it can be ready.
Then start a brief. I reply within 12 hours with what I'd put in version one, and a realistic timeline for it. For prices, see what an MVP costs by type or every price I charge.
Frequently asked questions
How long does it take to build an MVP?
Most MVPs I build take 45 to 60 days from the agreed scope. A one-feature tool can take about a week. A product with accounts, an admin console and data from other companies takes the full range, and the written scope gives the exact timeline.Can an MVP be built in two weeks?
The build part sometimes can. ShopMyESIM had a working first version after two days, and launch came about six weeks later, after design rounds, a rebrand and the switch to live data. A two-week MVP needs one core flow and every decision and piece of data ready on day one.What slows an MVP down the most?
Waiting. Brand direction that isn't settled, real data that arrives late, reviews that take days instead of one working day, and partner terms still being negotiated. A scoping sprint names each of these, with a date, before the build starts.Do AI coding tools make an MVP faster?
I use AI coding tools every day, and they make me faster. They don't decide what goes into version one or supply your data, so those still set the timeline.
Sources
- Published prices, tanujrajput.com/pricing · checked 14 Nov 2026
- Alter build log, tanujrajput.com/projects · checked 14 Nov 2026
- My own project records for BrandDesk (not public) · checked 14 Nov 2026
- ShopMyESIM case study, tanujrajput.com/work · checked 14 Nov 2026
- BrandDesk case study, tanujrajput.com/work · checked 14 Nov 2026
- Nicobar case study, tanujrajput.com/work · checked 14 Nov 2026
Last updated November 19, 2026
