How do you scope a small business software project so the price holds?
Short answer: Scope a software project by picking one business process, writing down what goes in and what comes out, listing what is explicitly out, and adding a plain pass or fail check for each feature. Fix the budget and the date, and let the feature list flex to fit. Anything that does not fit goes into a later phase with its own price. Scope means the written list of what will and will not be built.
You want to hire someone to build software without the price doubling halfway through, and you need to know what to write down before anyone starts.
I'm Ranjit, and I run Blur Studio: 15+ years building software, 120 apps taken live, and clients who have gone on to raise $10M+. I build AI systems and processes that help businesses do a lot more with a lot less time. I also work differently from most of what else is out there, more on that further down this page. Say hello.
What scope means, in plain words
Scope is a written answer to three questions. What will be built? What will not be built? How will we both know it is finished? Fixed price only works with a clear scope; see fixed price versus hourly software development.
Most budget blowouts are not caused by bad developers. They come from a gap between what the owner pictured and what was written down. The owner says "a simple customer portal". The developer hears "a login and a list". The owner pictured payments, messages and a mobile app. Nobody lied. Nobody wrote it down.
Good scoping closes that gap before any money moves. It takes a few hours, and it is the cheapest work in the whole project.
Step 1: Start with one process, not a product
Do not scope "a CRM" or "an app". Scope one business process: what happens from the moment something starts until it is done.
Examples: a job request comes in and a technician is booked. A quote is sent and a customer says yes. A new hire is added and given access.
Write the process as four short lines:
- Trigger: what starts it
- Inputs: what information goes in
- Steps: who does what, in order
- Result: what exists at the end, and who sees it
If you cannot describe the process in a page, the project is too big for one phase. Cut it down.
Step 2: Fix the budget and time, let the features flex
Most owners ask "how much for all of this?" That question invites a big number and a long list. A better question is "what is the best version I can get for this budget and this date?" For realistic timelines, see how long custom software takes to build.
The team at Basecamp, a project management software company, calls this an appetite in their free book Shape Up. They write: "Estimates start with a design and end with a number. Appetites start with a number and end with a design." In other words, the time you are willing to spend is fixed, and the features bend to fit. The same idea works with a fixed budget.
I like this approach for small businesses. It protects your budget, and it forces an honest talk about what matters most.
Step 3: Write an in and out table
The "out" column is the most valuable part of any scope. It is where surprises hide.
Worked example (made-up business and figures): A heating and cooling company with 8 technicians wants to replace a shared spreadsheet and a paper job board.
| In phase 1 | Out of phase 1 (maybe later) |
|---|---|
| Office staff create a job with customer, address, problem and priority | Customer-facing login portal |
| Dispatcher assigns a technician and a time slot | Automatic route optimising |
| Technician sees their jobs on a phone-friendly page and marks them done | A native app in the App Store |
| Technician adds notes and up to 5 photos per job | Video uploads |
| Office sees a list of open, booked and done jobs | Online payments and invoicing |
| Three user roles: office, dispatcher, technician | Connecting to the accounting software |
| One-time import of the current spreadsheet | Reports beyond a simple job count |
Notice the specifics: "up to 5 photos", "three user roles". Vague words like "manage jobs" or "user friendly" hide the cost. Numbers and named things do not.
Also list the quiet items most scopes miss: who owns the code and data at the end, where it will be hosted, who pays for hosting, what happens to old data, and who handles bug reports after launch. On ownership, do not assume. The US Copyright Office explains that software made by a contractor is not automatically yours: a commissioned work counts as "work made for hire" only if it fits one of nine listed categories and there is a signed written agreement saying so. Otherwise ownership is usually settled by a written assignment. Ask a lawyer, and get it in the contract.
Step 4: Add a pass or fail check to every item
An acceptance check is a plain test you can run to say "yes, this is done". The Scrum Guide, the standard rulebook for one popular way of running software teams, calls the idea a "definition of done": a formal description of the state of the work when it meets the quality measures required.
For small projects, I keep it simple. Each feature gets one sentence starting with "Done when":
- Create a job: Done when an office user can enter a job with all 5 fields in under 2 minutes and it appears in the open jobs list.
- Technician page: Done when a technician can open it on a normal phone, see only their own jobs, and mark one done in 3 taps.
- Photos: Done when a technician can add 5 photos and the office can see them on the job.
- Import: Done when all rows from the current sheet appear, and 20 random jobs match the original by hand.
Without checks, "done" becomes an opinion. With them, you can sign off without a fight.
Step 5: Price the first phase only
Ask for a fixed price on phase 1 and a rough range, clearly marked as a guess, for later phases. Do not sign a price for work nobody has designed yet.
Here is the logic any decent quote follows. The price is roughly hours multiplied by an hourly rate, plus a buffer for the unknown, plus hosting and any paid services.
Worked example (assumptions are made up): Say the developer estimates 70 hours for phase 1 at $100 an hour. That is $7,000. A 15 percent buffer for the unknown adds $1,050, so $8,050. Hosting and any other services are billed separately each month. The point is to see how the number is built, not to anchor on these figures. Your own quote will differ. For real market ranges and the main cost drivers, read custom software development cost.
If a quote is a single number with no breakdown, ask for the breakdown. If you want a small, cheap first release to test the idea, see AI MVP development.
Step 6: Agree how changes are handled
Scope will change. That is normal. What hurts is a change nobody priced. Agree up front:
- A change is written down and priced before work starts.
- You can swap items of equal size without extra cost.
- New ideas go on a "later" list, not into the current build.
Agencies face the same problem from the other side. Their guides on handling scope changes without fights, a simple scope of work template and a free statement of work example are worth borrowing from. A statement of work is the signed document that lists what is delivered, when and for what price.
Common scoping mistakes
- Scoping everything at once. The bigger the document, the more guesses it holds.
- Describing screens, not outcomes. Say what should be possible, then let the builder propose screens.
- Forgetting the people. Who will answer the builder's questions, and how fast? A slow owner is a hidden cost.
- Skipping security. OWASP, a nonprofit that publishes software security guidance, lists Broken Access Control as the first item in its Top 10 for 2021. Put "users only see what their role allows" in the scope as a check, not as an afterthought.
- Leaving out data. Moving old records is often the messiest job. Put it in or out, on purpose.
When to run a discovery step first
If you truly do not know what you need, do not force a scope. Spend a short paid step on discovery: talking to the people who do the work, mapping the process, and ending with a written plan. The UK government's service manual describes discovery as the stage before committing to build, where you establish the problem to be solved, and says you should not start building during it. It also gives about 4 to 8 weeks as typical for large public services. A small business process needs a fraction of that, often days.
Red flags in a scope or quote
- "Unlimited revisions" or "all features included" with no list.
- No "out of scope" section at all.
- No mention of who owns the code, the data and the hosting account.
- A price for the whole project but no phases.
- No acceptance checks, only a delivery date.
How I'd approach this
When someone asks me for a build, I start with a single process and one hour on the phone. I write the in and out table with them live. If the out column is empty, I know we have not talked enough.
My rule is that the first phase should be small enough to finish in a few weeks and useful on its own, even if you never build phase 2. That protects you. If the build goes badly, you have lost a little, not a lot.
Bring a one-page description of your process, the spreadsheet or tool you use now, and the date you need it by. That is enough to get a real scope and a fixed price for phase 1. If you do the steps above before you talk to anyone, you will get better quotes from everyone, including me.
Questions people ask
What should a software scope document include?
The process it covers, a list of features that are in, a list that are out, a pass or fail check for each feature, who owns the code and data, hosting, the budget and date for phase 1, and how changes are priced. A signed statement of work holds this, as in the free statement of work example.
How long does it take to scope a small software project?
For one process, a few hours of your time plus a short call with the builder. Large or unclear projects need a discovery step first. Do not rush this part, because it is the cheapest time in the whole project.
How do I stop scope creep on a software project?
Write the out list, give every feature a done check, and price every change before work begins. Put new ideas on a later list. My guide on handling scope changes without fights goes into the conversation itself.
Should I ask for a fixed price or hourly rate?
For a well-scoped first phase, a fixed price protects you. For unclear work, a capped hourly rate with a weekly check works better. Whatever you choose, ask for the breakdown. See custom software development cost for how quotes are built.
Do I own the code if a developer builds it for me?
Not automatically. The US Copyright Office says a commissioned work is a work made for hire only if it falls in listed categories and there is a signed written agreement. Put ownership in the contract and ask a lawyer to check it.
Is it better to start with a small first version?
Usually yes. A small first release gets real feedback and limits your risk. See AI MVP development for how small to go.
More in this topic: Problem solving: custom software (22 guides) →