Your own software should be considered for specific needs that are not supported by a reasonably ready solution. The decision should take into account both implementation and subsequent maintenance and access to data.
Describe problem without technology selection
Start with tasks that require manual work, repeat data or circumvent the limitations of the current tool. Specify the frequency and people involved in the process. Not all inconvenience warrants creating a new system; sometimes it is enough to arrange configuration or working method.
Compare the ready-to-use solutions with the requirements, not with the idea of an ideal panel. Select the necessary features, easy to simplify and those that can wait. The same list will be the basis for a custom-made chat.
Check Match Cost
A ready product may require a change of process or additional integration. Your own system gives you the ability to more accurately fit, but requires maintenance, development and organization of support. Do not present any option as automatically cheaper or better.
For example, a company that prepares custom valuations may need its own calculation module and at the same time use the ready-made tool for the rest of the work. The combination of solutions can be worth analyzing if the limits of responsibility and data flow are clear.
Establish responsibility for the life cycle
Before the decision, discuss:
- the method of transmission of the documentation and code in accordance with the contract;
- access to data and their export opportunities;
- maintaining the environment and updating;
- the rules for reporting errors;
- development of functions and valuation of changes;
- the possibility of another team taking over the job.
These arrangements are not an addition to the project. They determine how the company will use the tool after the first implementation has been completed.
Test the most important assumption
Before expansion, you can prepare a limited prototype or attempt to integrate. If the problem is calculation, check them on representative cases. If data exchange, verify available interfaces and behavior with errors. The effective screen will not confirm the feasibility of the process itself.
Make the decision based on scope and dependencies, taking into account team time. Helpful are problem selection for applications and The role and work of users.
