Decisions before development
You can build your own CRM. First work out why
“We do things differently, so we need our own system.” Perhaps. But that sentence does not yet explain what fails or how much the company loses. You can spend the money on development and then discover that the problem was how people agreed to work together.
What exactly is going wrong?
“We need a CRM” is already an answer. What was the question? Do salespeople lose enquiries? Is it unclear who replies to the customer? Does a proposal get assembled from five spreadsheets with a different price each time? Those are different jobs, with different system requirements. These examples are hypothetical.
Take one recent case and follow it from beginning to end. Who did what, where did they wait and what had to be entered again? If nobody owns the next step, software will not appoint that person for you. First agree on how the work should happen.
Give an existing tool some real work
I would first check whether an existing system can do the job. I would take an ordinary order, a difficult case and a correction, and work through each with the future user until it is done. Everything usually works in the vendor's demo. Your data and your staff will tell you more.
Check whether you can move the old data, set permissions and connect the system to the services you need. And how you get the information back if you leave. Sometimes configuration is enough; sometimes an extra module is needed. If a limitation remains, you can then describe it properly to a developer.
A bespoke system needs looking after
The launch price is only the beginning. Changes, support, infrastructure, training and fixes come afterwards. Someone on the business side will have to make product decisions and set aside time for them.
Compare the options over the same period. For an existing tool, count rollout, subscriptions, integrations and your team's work. For your own system, count development, those same integrations, support and later changes. Add data migration and a possible move to another system. Otherwise, you will compare the full subscription cost with one attractive programming quote.
When I would discuss building a system
When we know what work needs doing, have tested an existing tool on that work and can name the limitation preventing the business from earning, keeping its commitments or serving customers. Then we can cost a bespoke solution and discuss who will own it after launch.
If you plan to sell the system to other companies, you have another job: find buyers who can use a common solution. The first company's unusual process does not create a market by itself. It may need a custom project, without a future SaaS business in the presentation.
After that test, you will have a job with a clear result. You may find that development really is needed. Or that the money is better spent on configuration and helping people do the work. Both conclusions are more useful than owning a CRM nobody uses.