Ask three people at a plant how long the ERP rollout will take and the answers rarely match. The sponsor says six months. The project lead says nine. The warehouse supervisor says to ask again after the first full count.

All three can be right, because each is describing a different phase. The disagreement fades once the work is written down as an ERP Implementation Timeline, with an owner and an exit test for every phase. Without that, the schedule is one date that everyone reads differently.

Panorama Consulting Group’s 2026 ERP Report surveyed 170 organizations between January 2025 and January 2026 and found a median project length of nine months. Almost a quarter of the projects ran over schedule, and the most common cause was organizational: governance, resistance to change and process redesign. The software rarely got the blame. The median respondent had $200.5 million in annual revenue, so size that nine months to your own scope (figures as summarized by ERP Research).

The split below assumes nine months and a mid-size plant. Treat the week counts as a first draft.

Phase 1: discovery and data audit (weeks 1 to 6)

Most schedules are won here. The team lists every process the old system touches, including the ones nobody documented: the spreadsheet a planner keeps for rush orders, the label template on the shipping PC. Each gets a tag. Standard in the new system, configured, or customized.

Customization is where weeks disappear, so every item carrying that tag needs a named owner and a business reason.

The data audit runs in the same window. Pull the item master, the bills of materials and the open orders, then look for duplicates. Finding duplicate part numbers in week three costs an afternoon of cleanup. Finding them in week twenty costs a delay.

Phase 2: configuration and integrations (weeks 7 to 16)

Core modules configure quickly once the Phase 1 decisions are made. Integrations are slower, and they are what schedules underestimate. Scales, label printers, barcode scanners, the carrier, the EDI feed from the biggest customer: each is a small project with its own testing.

A GrayCyan client, a food distributor running Fishbowl, shows what this phase is worth. Its staff spent about 12 hours a day on manual receiving and inventory bridging. After scan-based receiving was connected to its purchase orders, with lot numbers captured at the dock, manual data entry fell 85% and any lot could be traced to its shipments in minutes.

Phase 3: inventory loading and testing (weeks 17 to 30)

This phase runs longest because it contains the inventory count, and a count cannot be rushed. Opening balances are only as good as the shelf they came from. Plan a full physical count close to cutover, load it, and reconcile every location that differs by more than a tolerance you set in advance.

If the target system is Odoo, the Inventory Setup in Odoo (warehouses, locations, units of measure, lot and serial rules) has to be settled before that count, because every later report inherits those choices. Changing them once transactions exist is disruptive.

Run user acceptance testing with the people who will use the system on a Monday morning, not with the project team. A receiving clerk who cannot finish a receipt in the test system will not finish one in production.

Phase 4: cutover and the first 90 days (weeks 31 to 36, then stabilization)

Cutover is a weekend, and the weekend gets rehearsed. Run at least one mock cutover with real data and a stopwatch. Freeze changes to the old system at an agreed hour, take the final count, load it, and run the checks written down in advance: open orders, inventory on hand, supplier balances. Decide before the weekend who can call a rollback, and by what hour.

The first 90 days after go-live belong on the schedule as well. Expect a steady run of small fixes, a label that prints wrong or a report that ignores one warehouse, and a slow climb in transaction accuracy as the floor builds new habits. Plants that treat go-live as the finish line staff the build and then leave users alone for the part where adoption is decided.

Exit tests worth writing down

• Phase 1 ends when the fit-gap list is signed and a sample of the item master has been cleaned.
• Phase 2 ends when every integration passes a scripted transaction.
• Phase 3 ends when count variances sit inside the agreed tolerance and acceptance testers finish their daily tasks unaided.
• Phase 4 ends when the rollback window has closed and a full week of shipments has gone out on time.

A nine-month median is a useful benchmark. Your own scope, and the dates on the physical count and the cutover weekend, decide where the plant lands.

About the Contributor

Nishkam Batta, Editor-in-Chief, HonestAI Magazine | AI Consultant, GrayCyan AI Solutions

Nish leads an applied AI company helping manufacturing and related companies automate operations with human-in-the-loop AI that integrates into ERPs, WMS, CRMs, and other enterprise tools, with an emphasis on explainable AI, clear audit trails, and measurable outcomes. His team builds agentic ERP systems that execute multi-step tasks inside approved guardrails so humans keep accountability, approvals, and override control.

Book a Consultation | LinkedIn