Practices arrive at clients database software from a spreadsheet that has stopped coping, and they face three routes. Keep the spreadsheet and add discipline. Build something in a general database tool. Or buy a product built for client records. All three are reasonable and all three fail in a characteristic way, so the useful question is not which is best but which failure you can most easily live with over the next two years, given who in the practice will actually maintain it.
The spreadsheet, and the moment it stops
Spreadsheets are excellent until two people need them at once, until you need to know what changed, or until documents need to belong to a row. Those three limits arrive in that order and they arrive quietly. The characteristic failure is not a crash but a slow divergence: several copies with slightly different data, and a growing reluctance to trust any of them.
The database you build yourself
General purpose database builders give you exactly the fields you want and cost very little to start. The characteristic failure is ownership. One person understands it, that person eventually leaves or gets busy, and what remains is a system nobody can safely change. If you go this way, the important artefact is not the database, it is the written description of how it works.
The product built for client records
Buying gives you a shape somebody else decided: clients, statuses, documents, history, and usually onboarding. The characteristic failure is fit. Something about your practice does not match the shape, and you either bend the practice or keep a small spreadsheet beside it for the awkward part. Judge products on how close the built in shape is to what you actually do, not on the length of the feature list.
Questions people ask about clients database software
How many clients before a spreadsheet stops being enough?
The count matters less than the number of people editing and whether documents are involved. Two people plus documents breaks a spreadsheet at almost any size.
Is building our own cheaper?
In licence fees, usually. In maintenance and in key person risk, usually not. Cost it including the hours of whoever will keep it working after the person who built it moves on.
Can we start with a product and export later if it does not fit?
Only if you check the export before you commit. Test it in the trial: what comes out, in what format, and whether documents come with their client and date attached.