SaaS platforms exist because most problems are common problems. A CRM built for ten thousand companies works for most of them because most companies sell, follow up, and close deals in roughly similar ways.
The question is not whether SaaS is good. The question is whether your problem is still a common problem.
The configuration ceiling
Every SaaS platform has a ceiling — a point beyond which the tool cannot be made to do what you need it to do. Below that ceiling, configuration is fast and cheap. Above it, you are either working around the limitation or paying for an enterprise tier whose primary feature is a higher ceiling.
The ceiling looks different for every platform. For a CRM it might be the inability to model a non-standard deal structure. For an inventory system it might be the lack of a location-specific costing model. For a reporting tool it might be the impossibility of combining data from two sources the vendor does not support.
When you hit the ceiling, you build a workaround. The workaround introduces friction. The friction accumulates.
The per-seat problem
SaaS pricing is typically per user per month. At ten staff, this is negligible. At a hundred, it is a line item. At five hundred, it is a cost that commands attention.
Custom software has high upfront cost and low marginal cost. SaaS has low upfront cost and high marginal cost. The crossover point varies — but it exists for every growing business, and it is usually earlier than the CFO expects.
The workaround tax
The real cost of a SaaS limitation is not the configuration time. It is the permanent operational overhead of the workaround.
Someone maintains the spreadsheet that bridges the gap between the CRM and the invoicing system. Someone runs the weekly export. Someone reconciles the two numbers that should always be the same but never quite are.
That labour is invisible in the software budget because it sits in the headcount budget. It compounds every year. And it makes the business harder to scale because the workaround cannot be easily trained, automated or inherited.
When to stay with SaaS
Custom software is not always the answer.
If your problem is genuinely common, SaaS is cheaper, faster and maintained by a team whose full attention is on that problem. If you are early and still figuring out your process, SaaS is the right tool for discovery. If the gap between the tool and your workflow is small enough to live with, the margin savings are not worth the build.
The right question is not "should we use SaaS?" It is "what is this limitation actually costing us, and what would it cost to remove it?" The answer to that calculation tells you whether the conversation about custom software is worth having.