Custom software timeline: weeks vs months

A small internal tool with stable requirements often shows a usable build in 6–10 weeks. Multi-org systems with history and APIs more often take 3–6 months. Unstable scope extends the calendar.

Ranges we can say in public

Small tools (accounts, one main flow, a few reports): often 6–10 weeks. Systems with approvals, many roles and external APIs: commonly 3–6 months. Those are magnitudes when scope is roughly stable — not a contractual slogan.

Code is rarely what slips the date

  • Key users can spare an hour a week to confirm.
  • The other ERP has no interface notes, so mapping is guesswork.
  • Masters still live in three coding schemes.
  • Go-live waits on a board meeting or a compliance window.

Do not reuse an app or WMS calendar

Apps add store review. WMS adds a floor walk and a trial loop. Software, mobile and warehouse are three critical paths that happen to share a vendor.

Rushing has a ceiling

You can add engineers on independent modules. You cannot add people to a partner who issues a test book next Thursday. Write that wait into the plan.

Frequently asked questions

Can we cut the calendar in half by paying more?

Parallel work helps. Discovery and counterparty interfaces do not compress the same way. Waiting for a finance test book is still waiting.

Why does test week explode?

Usually scope was never frozen, or acceptance is "leadership will have another look". Put acceptance in the memo or test week cannot end.

What should we send to get a first calendar?

Three must-have functions and one explicit non-goal. That is enough to sketch a first phase.

List what must go live and what can wait. We will put a first-phase calendar on that list at /en/contact/.

See the service Contact