All posts
01Custom Software·7 min read

When off-the-shelf software stops fitting.

The signs your business has outgrown a SaaS tool, the workarounds that compound, and how to tell whether custom software is the answer — or whether you just have not configured what you already pay for.

Published 12 May 2026Flowuity · The Practice

You bought the software. You configured it. You forced your team to use it. It still does not fit. The workarounds eat hours every week. Errors compound. The part of the business that gives you your edge is the part no software supports.

This is the most common shape of a leak. It is not catastrophic. The business runs. Things ship. Customers are served. But every week, somewhere in the work, a person is doing something a machine should be doing — typing the same data into a second system, opening a spreadsheet to translate one tool’s output into another tool’s input, sending an email at 5pm because the booking platform does not trigger the confirmation properly.

The first question is whether you have actually configured what you have. Most off-the-shelf software ships with a sane default and a deep settings page. Most buyers exhaust the default in week two and never open the settings page again. Before you decide that the software is the problem, hire someone for ten hours to push the configuration to its actual limit. If after that the tool still does not fit, you have a real signal.

The second question is what part of the business is genuinely yours. Every business has a few processes that are commodity (invoicing, payroll, basic CRM) and a few that are not (the way your team qualifies leads, the way your operations team handles a particular failure mode, the part of your customer experience that customers actually remember). Off-the-shelf software is built for the commodity. It is not built for the part that is yours.

Custom software is not the same as bespoke software. The distinction matters. Bespoke software builds every screen from scratch. Custom software builds the parts that are yours and leans on standard components for everything else. Authentication, payments, email — these are solved problems. The custom layer wraps your unique workflow on top of solved pieces.

The maths is more favourable than most owners assume. A custom system built on modern tooling — TypeScript, Next.js, Postgres, a hosted auth provider, a hosted email service — is no longer a multi-year project. The expensive part is not the code. The expensive part is the discovery: understanding the workflow well enough that the build does not need to be rewritten when the first user hits it.

The wrong reason to build custom is that the off-the-shelf tool is slightly annoying. The right reason is that the tool does not understand what you do, and your team is paying the gap. If five people each spend two hours a week working around a tool, you are losing about half a senior salary every year to the gap. That number is what custom needs to beat.

If the answer is custom, the next decision is how to scope it. The honest version is: build the smallest version that closes the largest part of the leak, and live with it for a quarter before you build more. The wrong version is to specify everything upfront and discover the truth too late.

→ Find out what your business is leaking. Book a Discovery.

End of post

If this resonates, the next step is a conversation.

Book a Discovery
More from the practice
02·AI Implementations

AI inside the work, not next to it.

Why ChatGPT in a browser tab does not move the needle on your business. What “AI inside the work” actually looks like — three patterns we ship most often, and where each one earns its keep.

05 May 20267 min
03·Strategy & Advisory

Inside a Flowuity Discovery.

A look at the two-week paid engagement that begins every Flowuity build. What we read, who we interview, what the memo contains, and why a clear no is the most useful outcome.

02 May 20266 min
04·AI Agents

The work that does not need a person.

Lead qualification, support triage, reconciliation, scheduling. When an AI agent is the right answer, how the economics work, and where agents fail in ways that humans do not.

28 April 20268 min