Back to Studio

AI-friendly Markdown · structured for AI citations

Growth · Studio

How to Scope a Side Project MVP You Can Actually Ship on Weekends

16 min read
How to Scope a Side Project MVP You Can Actually Ship on Weekends

A practical playbook for nights-and-weekends builders to cut scope, pick a weekend-maintainable MVP, and ship without burning out or rebuilding forever.

To scope a side project MVP you can actually ship on weekends, start from the hours you really have, not the features you want. Count your honest weekly hours, cut that number for life and for optimism, then keep only the features without which your one-sentence promise would be false. For most people with a day job that means one user, one job, one happy path, and no more than three must-ship features that fit inside about six weeks of evenings. Everything else goes on a written "not in v1" list.

Last updated: Nov 11, 2026

Most side projects die in the feature graveyard, not because the idea was bad, but because the scope assumed a full-time team. You sketch a dashboard, an admin panel, OAuth for five providers, a referral engine, and a polished onboarding tour. Then Tuesday night arrives and you still have not shipped the one thing a stranger would pay for.

If you are building nights and weekends, scope is the product. This playbook is for founders who want an MVP they can finish and keep alive after work, not a fantasy version that needs a sabbatical. Research figures link to their sources; the worked example uses a made-up product and is illustrative. When you are ready to sell that thin version, pair this with how to find paying customers for a side project without quitting your job.

What does "weekend-maintainable" mean for an MVP?

An MVP is not the smallest demo you can hack together on a Saturday. Eric Ries framed the minimum viable product around learning the most about customers for the least effort. For a side project, add one more condition: it is the smallest product you can keep running for the next three months without burning out or rewriting it every Friday.

Use three filters before you write another line of code:

  1. One core job. A stranger can describe the outcome in one sentence. "It turns my meeting notes into a client-ready summary" beats "an AI workspace for knowledge workers."

  2. One path to value. There is a single happy path from signup to that outcome. Side doors, power-user modes, and "also works for agencies" can wait.

  3. Two-hour ops budget. In a normal week you can handle support, small fixes, and billing questions in about two focused hours. If the product needs more babysitting than that, the scope is still too wide.

If your plan fails any of those filters, you do not have an MVP yet. You have a roadmap wearing a launch date. Be honest about the third filter: unpaid evenings spent fighting deploys and confused DMs are still costs, even when your payment dashboard is quiet.

The goal is not to impress other builders. The goal is a version you can ship, charge for, and still sleep.

How many hours do you really have for an MVP?

Most scoping advice starts with features. Start with the clock instead. Basecamp's Shape Up method calls this the appetite: decide how much time a piece of work deserves first, then design something that fits, which it describes as fixed time, variable scope. In Shape Up a "big batch" is six weeks for one designer and one or two programmers. You are one person working evenings, so your six weeks hold far less.

Two adjustments make the number honest:

  • Life takes a cut. Work crunches, a sick kid, a weekend away. As a rule of thumb, assume about a quarter of your planned hours never happen.

  • Your estimates are optimistic. In the classic planning fallacy study by Buehler, Griffin and Ross (1994), honours students predicted they would finish their thesis in 33.9 days on average and actually took 55.5 days. Only 29.7% finished by their own best-guess date. That is roughly 1.6 times longer than planned, from people working on something they cared about.

Here is what that does to a six-week MVP window, counted in 90-minute build blocks (a block is one focused evening session):

Planned hours per week

Raw hours in 6 weeks

After a 25% life cut

After dividing by 1.64 for optimism

90-minute blocks you can count on

4

24

18

about 11

7

6

36

27

about 16.5

10

8

48

36

about 22

14

10

60

45

about 27.5

18

The 1.64 is 55.5 divided by 33.9 from the study above; the 25% cut is a rule of thumb, so swap in your own track record if you have one. The point is not precision. The point is that "8 hours a week for six weeks" is closer to 14 real sessions than 48 hours, and your scope has to fit in 14 sessions.

If you have shipped side projects before, use your own history instead: look at how many evenings your last finished feature took against what you guessed. Buehler and colleagues found the optimism shrank when people tied their prediction to past experience, so a quick look back is worth doing.

How do you cut features down to the paid promise?

Open a blank note. Write the promise you would put on a checkout page in one line. Then list every feature you have imagined and mark each one:

  • Must ship: without it, the promise is false.

  • Nice later: helps, but a determined user can work around it.

  • Theater: exists to make you feel professional (fancy empty states, dark mode, multi-tenant orgs you do not have).

Delete theater from v1. Move nice-later to a parking lot doc. Keep must-ship to three items or fewer. If you cannot get under three, your promise is still a platform, not a product.

Cutting feels like losing value, so here is a useful reality check. Pendo's 2019 Feature Adoption Report, built on anonymised usage data from 615 Pendo subscriptions over three months, found that 80% of features in the average software product are rarely or never used, and that about 12% of features generated 80% of average daily usage. Those were funded teams with product managers. A solo builder guessing features in the evening is unlikely to beat that hit rate, which is why the must-ship list should stay short.

A useful gut check: could you fulfil the first ten customers by hand for a week if the UI broke? If yes, you are close. Manual fulfilment is ugly, but it proves the promise before you automate the wrong thing. That same honesty shows up when makers get brutal feedback from makers before a public launch: ship the thin truth, then listen.

Example: if the promise is "weekly competitor digest in your inbox," must-ship might be crawl sources, generate digest, deliver email. Nice later is a web dashboard. Theater is team workspaces and custom brand kits. Build the email first. Plenty of people will pay for a boring email that saves them an hour.

What does a scoped MVP look like on paper?

Here is a worked example with a made-up product. Imogen is a backend developer with about 8 hours a week. Her idea, Tallyrook, sends polite, escalating payment reminders for freelance illustrators whose clients pay late. Her promise: "Paste in your open invoices and Tallyrook chases the late ones for you by email." From the capacity table, 8 hours a week gives her about 14 blocks over six weeks.

She lists everything she had imagined and estimates each item in 90-minute blocks:

Feature

Must / Later / Theater

Estimate (blocks)

Why

Add invoices by CSV upload or a simple form

Must

2

No invoices in, no promise

Reminder schedule (3, 10 and 21 days late)

Must

3

This is the job itself

Three reminder templates, friendly to firm

Must

2

Without them the emails are generic

Signup, landing page and one paid plan via hosted checkout

Must

4

Needed to say yes to money

Accounting software sync

Later

6

CSV works for the first users

Dashboard with late-payment charts

Later

4

A weekly summary email does the job

Multi-currency

Later

3

Start with one market

Team seats for studios

Theater

5

Her first buyers work alone

Custom sending domain per user

Theater

4

Send from her domain with reply-to set to the user

Dark mode

Theater

2

Nobody pays for it in v1

The full wish list is 35 blocks, about two and a half times her budget. The four must-ship items total 11 blocks, which leaves 3 blocks of slack inside 14 for the bugs she has not met yet. "Reminder schedule" plus "templates" could arguably be one feature, which keeps her at three. That is a scope she can finish. The original list would have been a six-month project that stalled around week three.

Notice what the table does that a feature list does not: every "no" has a reason next to it, so when the urge to add accounting sync comes back on a Thursday night, the answer is already written down.

Which kind of MVP fits a weekend schedule?

Not every MVP has to be software. Pick the cheapest shape that still tests whether people want the outcome:

MVP type

What you build

Fits when

Watch out for

Concierge

Nothing automated; you deliver the result by hand

The job is valuable and low volume

Your evenings become the bottleneck past a handful of customers

Wizard of Oz

A real front end, manual work behind it

Users need to feel the product, but the back end is hard

Be honest in your terms about human review if it matters to users

Single-feature product

The one must-ship path, fully working

The core job is mostly software already

Quietly adding "just one more" feature before anyone pays

Pre-sale page

A page with the promise and a way to pay or reserve

You are unsure anyone will pay at all

Interest is not usage; follow up and deliver fast

If you have not talked to anyone with the problem yet, do that before any of these. The five-conversation method in validating a side project idea with five conversations before you build fits in two weeks of evenings and makes the must-ship list much easier to write.

How do you design for Tuesday night, not launch day?

Your constraints are not theoretical. You have a job, a family, sleep, and maybe 6 to 10 good hours a week. Scope against that calendar:

  • Build in 90-minute blocks. If a feature cannot be meaningfully advanced in one focused block, break it or drop it.

  • Prefer boring infrastructure. Managed auth, hosted database, hosted checkout, one email provider. Clever self-hosting is a second job. If you are still choosing tools, choosing a side project tech stack you can still maintain covers the boring picks.

  • Ship vertical slices. End-to-end for one user type beats horizontal scaffolding for five personas.

  • Freeze polish until someone pays. Icons, marketing-site animations and pixel-perfect spacing are optional until money says the problem is real.

Write a hard "won't do in v1" list and keep it next to your editor. When you feel the itch to add Google login "real quick," re-read the list. That itch is how weekends disappear.

Also plan recovery. If you miss two evenings because work exploded, what still works for users? Prefer designs that degrade gracefully: a delayed digest is better than a half-broken dashboard that needs you online to babysit jobs.

What does a 14-day scoping sprint look like?

Do not wait for a free month. Run a short, boring sprint with a calendar invite to yourself so it does not evaporate:

Days 1 to 2: Promise and buyer. Write the one-sentence promise and the one person it is for. If you cannot name a real human (or a sharp composite), stop building and talk to people first. Save their exact words. Your MVP should answer their sentence, not your mood board.

Days 3 to 4: Hours and cut list. Fill in the capacity table with your real hours, then run the must / later / theater pass with block estimates like the Tallyrook table. Capture the parking lot so you do not feel like you are losing ideas; you are sequencing them. Share the cut list with one trusted builder and ask what still looks fat.

Days 5 to 7: Happy path prototype. A clickable flow from landing page to signup to core job to a "you did it" state. No admin. No analytics suite. Screenshots or a short screen recording are fine if code is slow. The test is whether a stranger understands the outcome; a free 10-second landing page test is a quick way to check whether your promise reads clearly before you show it to anyone.

Days 8 to 10: Manual test with three people. Walk them through. Note where they stall. Fix only blockers that kill the promise and ignore taste feedback about colours. If all three stall in the same place, that is your next 90-minute block, not a new feature idea from social media.

Days 11 to 12: Payments and a support path. A hosted checkout, a real price, and a single email address or form for help. You are not launching a support org. You are making it possible to say yes to money. Put the refund policy in one plain sentence so you are not improvising under pressure.

Days 13 to 14: Soft ship. Put it in front of a small list. Before you do, a free launch readiness checker catches the basics people forget, like a missing description or broken preview image. Do not wait for launch-day energy. Quiet is fine; momentum comes from usage, not countdown timers. After you have a few buyers, the playbook in how to get your first 10 paying customers without a launch day beats another week of feature polishing.

How do you keep the MVP from melting after you ship?

Shipping is half the job. Keeping a side project alive is the other half:

  1. One changelog a week, max. Batch tiny fixes. Constant hotfixes are a sign the scope still leaks.

  2. Office hours, not always-on chat. Tell users when you reply (for example, weekday evenings). Boundaries protect quality.

  3. Test only what would hurt. The payment path and the core job deserve a few automated checks; the settings page does not. How much testing a side project actually needs walks through that budget.

  4. Kill unfinished experiments. Feature flags that stay half-on for months are technical and mental debt.

  5. Revisit scope monthly. Ask what paying users actually used. Cut what they ignore, even if you love it.

If maintenance routinely blows past your two-hour weekly budget, shrink the product again. A smaller product that stays up beats a broader one that goes quiet for three weeks every time work gets busy.

Keep a tiny ops checklist: renewals, failed payments, one backup reminder, and "what broke last week." Review it Sunday night in fifteen minutes. That ritual prevents the spiral where you open the laptop and discover three silent failures.

What are the common scoping traps?

  • Building for your future company. Multi-team permissions before you have one team. You are optimising for a press release, not a customer.

  • Copying SaaS market leaders. Their MVP is not public. You are seeing year five. Copy their job to be done, not their settings page.

  • Estimating in hours instead of sessions. "It's only 6 hours" sounds small until you remember it is four evenings, and one of them will get cancelled.

  • Waiting for "ready." Ready is a feeling that arrives after someone pays. Use a realistic first-week launch checklist when you soft-ship, not as an excuse to delay another month.

  • Scope creep from friendly feedback. Friends suggest features because they care. Thank them, park it, and ask whether they would pay for the current promise.

  • Secret second MVP. Rebuilding the stack mid-flight because a new framework is shiny. Finish the boring version first; rewrite only when paying users force a constraint.

How do you know the MVP is scoped right?

You know the MVP is scoped right when:

  • You can explain it in one breath without saying "and also."

  • The must-ship list fits inside your real block count with a little slack left over.

  • A new user can hit the core outcome without a call from you (or with one short setup call you can schedule).

  • You can take a sick week at your day job and the product does not panic.

  • Someone has paid, or has a clear next step to pay, without you adding another module first.

Side projects do not fail from lack of ambition. They fail when ambition outruns the calendar. Cut until the promise fits your real life, ship that, then grow the product with the same discipline you used to shrink it.

Frequently asked questions

How long should a side project MVP take to build?

Pick the time first and fit the scope to it. Six weeks is a practical ceiling for most people working evenings: long enough to build one real path, short enough that motivation and life do not erode it. If your must-ship list does not fit in the blocks you can count on in that window, cut the list rather than stretching the deadline.

Does an MVP need login and payments?

Login only if users need their data to persist or stay private. Payments yes, if the question you are testing is whether people will pay, which for most side projects it is. A hosted checkout link takes far less time than a billing system and still tells you the truth.

What is the difference between a prototype and an MVP?

A prototype shows how something might work; an MVP lets a real user get the real outcome, even if parts of it are manual behind the scenes. If nobody outside your friends can use it to get the job done, it is still a prototype.

How many features should an MVP have?

As few as keep the promise true. Three must-ship items is a good ceiling for a nights-and-weekends build; if you need more than that, the promise is probably describing two products.

side-projectsmvpscopingweekendsgrowth