A business management system project, planned to finish

The key facts

• Scope by records, not features: which truths will the system keep — stock, orders, money, people — and which stay with other systems.
• Phase by dependency: the record first, then the flows that feed it, then measurement, then automation. Screens and reports come last.
• Budget the tail: import, training, parallel running, and the first month's corrections — the part every plan underestimates.
• The buy path is the same plan: BSimple ($180/mo AUD up, trial) replaces build phases 2–4 with configuration — the records and sequencing still apply.

Scope: decide the records before the features

The plan starts with a written sentence per record: stock — will the system hold live quantities, for how many locations? orders — which intake channels, which fulfilment steps? money — kept here, or handed to the accounting platform (state it explicitly; the boundary must be a decision, not an accident)? people — kept here, or in the HR tool?

Then the refusals, written down with equal care: what this project will not attempt — no custom reporting beyond three reports, no integrations beyond the named two, no historical data before a cutoff date. The definition work and the ER thinking give the scope its structure; the refusals give it a finish line.

Phases: dependency-ordered, not ambition-ordered

Phase 1 — the record. Products, locations, parties, movements loaded and a baseline stocktake done. Nothing else works until quantities are true.

Phase 2 — the flows that feed it. Receiving and order intake on the new record; the old process retired for those two flows on a hard date. Partial adoption is how dual records are born.

Phase 3 — fulfilment and handoff. Picking, packing, invoice production and the accounting crossing — with the first month watched for mismatches daily.

Phase 4 — measurement and automation. Reports, reorder signals, portal rollouts to customers. Only after the record is trusted do the refinements help.

The tail, budgeted as a phase: importing history, training each role on their screens, parallel-running for one cycle, and the first month's corrections. The project-planning page scopes this tail for the build route; it exists on the buy route too, just smaller.

The buy path — the same plan, fewer phases

A vendor implementation follows the identical plan with phases 2–4 compressed into configuration: the records and sequencing are yours to decide either way, which is why this page serves both routes. We build BSimple, so weigh that: as a product it is phases 2–4 pre-built — live multi-location stock, batches, purchasing from demand, order-to-invoice, portals — leaving your project with scope, the baseline stocktake, training, and go-live. From $180/$250/$399 per month AUD, with the trial as phase zero: prove the record on your products before the plan commits to anything.