Ready-made or custom software: what to check before you start

Describe the difficulty first

“We need a CRM” already selects a category of tool. Start one step earlier: what action repeats, who does it, and where does work break down? For example, requests from different channels may need copying into a shared spreadsheet, while the person responsible cannot reliably see their status. This is an illustrative situation, not a claim about a particular project.

Describe the current workflow and an observable improvement: processing time, repeated manual steps, or requests without an assigned owner. If you do not know the starting point, first watch how the work happens today.

Try an existing tool with your own scenarios

A feature list is not a test. Take a normal request, a difficult exception and a mistake that needs correcting. Work through each from start to finish with someone who will use the system.

Check permissions, migration of existing data, integrations and the ability to export data when switching tools. Configuration may solve the problem. A small extension may be enough. A complete custom build is another option, not a required starting point.

Identify what actually needs a new product

Custom development is worth discussing when an important business workflow cannot be handled adequately with existing tools. An unusual process is not sufficient evidence on its own: the process itself may need changing first.

A product intended for other companies needs a separate test. Does the problem recur across independent customers? Who makes the buying decision? What do they use now? Interest in a demonstration and willingness to pay are different signals.

Compare ongoing costs, not just launch costs

Choose the same comparison period for each option. Include implementation, subscriptions or licences where applicable, data migration, training, integrations, changes and support. Include your team’s time and the eventual cost of switching.

A custom system also needs an owner after launch. Who decides on changes, helps users and pays for maintenance? Without these answers, the development estimate is incomplete.

Finish with one testable next step

Before asking for a large project estimate, fill in a short note. It helps distinguish a missing tool from an unclear problem.

• What task repeats, and for whom?
• How is it handled today?
• What goes wrong, and how can you observe it?
• Which existing option have you tried?
• What must carry over: data, permissions, integrations?
• Who will assess the next step, and against what result?

When to postpone development

If you cannot yet identify the user, the process owner or a useful outcome, clarify those first. The investigation may lead to an existing service, an extension, a custom system or a decision to build nothing for now. Its value is in resolving a specific uncertainty.