When to Build vs. Buy: A Simple Decision Framework
"Should we build this custom or buy something off the shelf?" is a question most small businesses answer with a…
Read →
Most custom software projects that go wrong do not fail in the code. They fail in the gap between what the owner pictured and what the developer understood. Knowing how to scope a software project closes that gap before any money is spent, and the document that does it can fit on one page.
This guide is for owners of small service businesses and Iowa nonprofits who have decided to build and need to define what they are asking for.
This guide begins after the decision to build. If you are still weighing a custom build against software you can buy, read When to Build vs. Buy: A Simple Decision Framework first. That article helps you decide whether to build at all. This one assumes the answer is yes and covers how to describe the project so it can be priced, built and finished. For the cost side, The Real Cost of Good Enough Software covers what tools that almost fit really cost.
A scope page is not a specification. It does not describe screens or database tables. It records the decisions that determine success, in language both sides can read. Nine items cover it.
Write one sentence that names the problem, not the solution. “Our crews’ schedules live in three places and customers get double-booked” is a problem. “We need an app” is not. If you cannot say it in one sentence, the project is not yet ready to be scoped.
List the roles that will touch the system: the owner, office staff, field crews, customers, a bookkeeper. For each, note what they need to do. A system built only for the owner often fails in the truck, where the crew lead has a phone and thirty seconds.
Name the workflows the software must support on day one, in plain verbs: “Customer requests a quote online; office approves; job lands on the crew schedule.” Limit the list to three to five. A sixth workflow is usually a second project, and folding it into the first is a common way budgets double.
Write down what this version will not do. The out-of-scope list surfaces assumptions. If you expect payroll export or customer text reminders and nobody wrote it down, each becomes a disagreement later. Wanted eventually but not now goes here, labeled “later.”
List what information goes into the system, where it comes from, and what leaves it. Name every other tool involved: accounting software, a spreadsheet, a payment processor, a calendar. Integrations are where hidden effort lives: connecting to a tool with a documented, supported interface is a different job from working around one without. Note any existing data that must be moved, and how clean it is.
Pick one or two measures you could check three months after launch. “Dispatch takes under ten minutes a day instead of an hour” can be verified. “The business runs better” cannot. A measurable target also gives you a way to say the project is finished.
State the budget range, the date that matters and why it matters, and any compliance obligations. A snow removal operator in Iowa needs the system before the first storm, not after. A nonprofit may have grant reporting rules or donor data handling requirements. Constraints shape the design, so they belong on page one.
List the two or three things most likely to hurt the project: data that is messier than expected, a third-party tool that may change its terms, a key staff member with limited time to give feedback, or a hard launch date.
Projects change, and that is not the problem. Unrecorded change is. Agree in advance on how a change is handled: the new request is written down, the builder states its effect on cost and timeline, and the owner approves or declines before work starts. Two sentences on this page prevent most scope disputes.
A scope page is often treated as a contract trap for the owner. In practice it protects both parties.
For the owner, it fixes what you are paying for. It lets you compare quotes, because each builder prices the same document rather than their own reading of a conversation, and it gives you a clear way to ask whether something is in scope.
For the builder, it defines what finished means. Without it, projects drift as every conversation adds “just one more thing” to an unpriced pile. That is bad for the builder and worse for the owner, whose project never closes.
Both sides benefit from writing down what was agreed while everyone still agrees.
Fill in the nine items before you request quotes, then give the same page to every builder you talk to. A builder who asks pointed questions about your data, integrations and out-of-scope list is thinking about the real work. A builder who quotes without asking is guessing.
A spreadsheet, a no-code tool such as Airtable, an automation service such as Zapier, or an off-the-shelf product such as Jobber may each cover part of what you wrote down. Custom software earns its place when your workflows are specific to your business and available tools need painful workarounds.
One page is the target for a first version. If it runs to three pages, the project is probably several projects, or the problem statement is still unclear. Detail belongs in later design work.
The owner should write the first draft, because the problem, users and success measure come from the business. The builder then tests it with questions and adds technical items such as integrations and risks.
Use the change request process on the page. The change is written down, the builder states the effect on cost and timeline, and you approve or decline it before work begins. Changes are normal. Unrecorded changes are the ones that cause disputes.
Yes. Every item on the page is about your business: the problem, people, workflows, data, goal and limits. Technical judgment is the builder’s job. Yours is to be specific about what the business needs.
This article was written by FORGE, the custom software and integrations platform from Intelligent Analytics and Data Management.
"Should we build this custom or buy something off the shelf?" is a question most small businesses answer with a…
Read →"Good enough" software rarely feels like a mistake in the moment it's purchased. It feels like a reasonable, cost-conscious decision.…
Read →Most businesses don’t think about their software stack until it stops working. By then, the cost is already there —…
Read →