When custom ERP is worth it (and when it isn't)

A good share of the ERP conversations we have end with us recommending against a custom build. That surprises people, since we sell custom development. But building the wrong system is worse for everyone than not building it, so we've settled on three questions we walk through before quoting.

First: is the pain a missing feature or a mismatched model? If your current tool is missing an export button, that's a feature request, and the vendor will probably ship it eventually. If the tool assumes your orders work one way and they actually work another, no amount of configuration fixes that. Feature gaps close; model mismatches don't.

Second: what are the workarounds costing per month? Count the hours spent re-entering data between systems, the errors that reach customers, the reports someone assembles by hand every Friday. If that number is small, a custom system won't pay for itself. In one recent case it worked out to roughly two full-time salaries, which made the decision easy.

Third: will the process still look like this in two years? Custom software is an investment in a process you plan to keep. If the operation is still changing shape every quarter, wait. Build when the process has settled and the tool is what's holding it back.

If the answers are mismatched model, real money, and a stable process, custom is worth pricing out. Otherwise keep what you have and ask again in six months. We're happy to be the ones who tell you that.