How Long Does an ERP Implementation Really Take? Timelines by Scope, Size and Approach
Most ERP go-live dates are set before anyone has looked at the data or the localization list. Here's how long single-entity, multi-entity and multi-country projects typically take, and which decisions add months.
Realistic duration ranges by scope and company size
Your vendor will usually quote the shortest plausible date, and your board will remember it. Before you commit, build an ERP implementation timeline that fits your scope, not the vendor's best-case reference project. The biggest driver isn't the software. It's how many legal entities, countries and operational processes go live at once.
The ranges below are typical for projects run with an experienced implementation partner. They run from contract signature to go-live and exclude hypercare. Treat them as planning ranges, not promises.
| Scope | Typical profile | Finance only | Finance plus inventory, order-to-cash, procurement | Adding manufacturing or warehouse execution |
|---|---|---|---|---|
| Single entity, one country | 20–150 users; NetSuite, Business Central, Acumatica, Odoo, SAP Business One | 3–5 months | 5–8 months | 7–11 months |
| Multi-entity, one country | 3–10 entities, intercompany, consolidation | 4–7 months | 7–11 months | 9–14 months |
| Multi-country, mid-market | 2–6 countries, different tax and e-invoicing rules | 6–10 months for wave one | 9–15 months for wave one | 12–18 months for wave one |
| Multi-country enterprise | S/4HANA, Oracle Fusion Cloud, Dynamics 365 Finance and SCM | 18–36 months for the full program, usually a global template plus country waves of 3–6 months each |
So how long does an ERP implementation take? For most mid-market companies with real operations behind the general ledger, plan on 7–12 months. If someone quotes 90 days for a multi-entity distributor with a warehouse, ask which processes they're leaving out.
Where the weeks go in each ERP project phase
SAP Activate, Oracle's method, Microsoft's Success by Design and partner-built approaches all use similar ERP project phases. The names differ. The work, and the places where projects slip, don't.
Here's a hypothetical example. A 120-user distributor has three entities in one country and is moving to a cloud ERP with inventory, purchasing, sales orders and EDI with major customers. A sensible plan comes to roughly 36 weeks:
| Phase | Weeks | What actually happens | Common slippage |
|---|---|---|---|
| Mobilize and discovery | 3–4 | Team onboarding, current-state walkthroughs, scope confirmation | Key users unavailable; scope reopened |
| Solution design | 5–7 | Fit-gap, chart of accounts, item and costing design, integration specs | Sign-off delays from department heads |
| Build and configure | 8–12 | Configuration, extensions, integrations, reports, forms | Custom requests added mid-build |
| Data migration (runs in parallel) | Across build and test | Cleansing, mapping, usually two or three mock loads | Mock loads fail reconciliation |
| Testing (system, integration, UAT) | 6–9 | End-to-end scenarios, period-close test, defect fixing | Business testers pulled back to day jobs |
| Training and cutover | 3–4 | Train-the-trainer, final data load, opening balances, go/no-go | Unreconciled balances force a later date |
| Hypercare | 4–8 after go-live | Stabilization, first month-end close in the new system | Often left out of plans entirely |
Design and testing together take up close to half the calendar, and buyers tend to underinvest in both. Compressing design rarely saves time. It moves the arguments into testing, where they cost more to resolve.
The decisions that quietly add months
Slippage rarely comes from one big mistake. It comes from several reasonable-sounding decisions made early, before anyone has costed them in weeks.
Customization
Every custom workflow, pricing engine or approval chain needs its own specification, build and testing, then regression testing at each vendor release. SaaS ERP typically ships a few updates a year, so that cost returns after go-live. As a rough guide, each significant extension adds 2–6 weeks of elapsed time, more when extensions share developers or testers. Our position: use standard functionality for anything that doesn't differentiate you, and fund customization only where it protects revenue or margin.
Data quality
Duplicate customers, inconsistent units of measure, obsolete SKUs and unreconciled open items are the most common reason a cutover date moves. If your item master has never been cleaned, budget 1–3 extra months and give the cleansing to named business owners, not the partner. Decide early how much history to migrate. Open transactions plus opening balances is the fastest route. Several years of line-level detail can add weeks of mapping and reconciliation.
Localization and compliance
This is where multi-country timelines grow. E-invoicing and real-time reporting mandates differ widely: Mexico's CFDI, Brazil's NF-e, India's GST e-invoicing, Saudi Arabia's ZATCA e-invoicing, Italy's SDI, and the broader move toward EU digital reporting. Some ERPs cover a country natively. Others depend on a partner add-on or third-party compliance provider, each with its own certification and testing. Budget roughly 4–12 weeks per country where localization isn't native. Data-residency and privacy rules (GDPR in Europe, India's DPDP Act, and various rules in the Gulf and Southeast Asia) can affect hosting region and integration design, so involve legal and tax advisors early.
Internal staffing
Partners configure the system, but only your people can make decisions and test it. If key users give 10–20% of their time instead of the 40–60% the plan assumed, expect the timeline to stretch by a third or more. Backfilling the finance controller and operations lead is usually the cheapest schedule insurance you can buy.
Integrations
WMS, TMS, carrier APIs, EDI, e-commerce platforms, banks and payroll each typically take 3–8 weeks, depending on prebuilt connectors versus custom middleware. Find out early which counterparties need their own testing windows, such as large retail customers' EDI certification or bank connectivity.
Phased vs big bang ERP rollout: which one fits your business
The phased vs big bang ERP rollout decision affects total duration more than any vendor choice. Neither is safer in every case. It depends on how tightly your entities and processes depend on each other.
A big bang takes every entity and module live on one date. You avoid temporary interfaces between old and new systems and run one cutover. The calendar is often shorter, but the risk sits on a single weekend, and a failed cutover hits the whole business. It suits single-country groups with shared processes and a handful of entities.
A phased rollout goes by entity, country or module. The first wave builds a global template and later waves reuse it. The program takes longer, but each go-live is smaller and the team learns between waves. It suits multi-country groups, companies with very different business units, and anyone facing a hard localization deadline in one market.
Here's a hypothetical worked example. A distributor operates in the UK, the Netherlands and India. A big bang might take 14–16 months because India's GST e-invoicing and the Dutch VAT setup must be ready together. A phased plan could take the UK live at month 9 on the template, the Netherlands at month 12, and India at month 15 after localization testing. The end date is about the same, but the first benefits arrive months earlier and the riskiest country goes last, when the team has the most experience.
Phased rollouts carry one hidden cost: temporary intercompany and consolidation workarounds while some entities remain on legacy systems. Put those interfaces in the plan, or they'll swallow your middle waves.
How to set a go-live date you can defend to the board
Good ERP go-live planning works backward from the business calendar, not forward from the contract date. A defensible date has evidence behind it, a buffer, and clear criteria for when it moves.
- Pick a natural cutover point. Go live at the start of a fiscal period, ideally not at year-end, when auditors and closing work compete for the same people. Avoid peak season: Q4 for retail, harvest cycles for agri-supply chains, Ramadan or Lunar New Year where they affect operations.
- Build the plan bottom-up with the partner. Ask for a week-by-week plan with named client resources and their time allocation, not just phase bars.
- Add 15–25% contingency on top of the partner's estimate for mid-market projects, and more where localization or data quality is unknown.
- Set go/no-go criteria now. For example: the final mock load reconciles to the trial balance within an agreed tolerance, no open critical defects, a period-close test completed, and key users signed off on UAT.
- Budget the cost of slipping. Partner teams commonly bill roughly $40,000–$150,000 per month on mid-market projects (typical 2026 market ranges; they vary widely by region and contract, and are often lower in India and parts of Southeast Asia). You may also pay for overlapping legacy licenses. Confirm both with your vendors.
Red flags that your date isn't credible:
- The statement of work says "fit-to-standard" but the gap list already has dozens of custom items.
- Nobody has run a mock data load before the go-live date is announced.
- Localization for a non-native country is "to be confirmed" with a third party.
- The client team has no backfill and still carries full operating duties.
- Hypercare and the first month-end close aren't in the plan.
Here's the rule we'd apply. Don't announce a firm go-live date until solution design is signed off and the first mock data load has been reconciled. Before that, give the board a range. After it, give them a date with the contingency already built in.
Common questions
How long does ERP implementation take for a small, single-entity company?
A single-entity company implementing finance only on a SaaS ERP such as Business Central, NetSuite or Acumatica can often go live in 3–5 months. Adding inventory, purchasing and order management usually pushes that to 5–8 months. Clean data and available key users keep it at the short end.
Is cloud ERP actually faster to implement than on-premise?
Cloud ERP removes infrastructure setup and tends to encourage standard processes, which typically saves weeks rather than months. The bigger time drivers, including data cleansing, design decisions, integrations and testing, are the same either way. Heavy customization can erase the speed advantage altogether.
Should we go live on the first day of our fiscal year?
It's tidy for reporting, but it puts cutover on top of year-end close and audit preparation, which stretches the same finance team. Many companies prefer the start of a quarter or month mid-year. Discuss the trade-offs with your auditors and tax advisors before deciding.
Do vendor rapid-deployment or preconfigured packages really shorten the timeline?
They can, often by several weeks to a few months, if you accept the package's predefined processes with limited changes. The savings vanish when the business asks for exceptions during design. Ask the vendor or partner to state in writing which processes the package covers and what happens to the timeline if you deviate.
Is hypercare part of the ERP implementation timeline?
It should be, though headline dates often leave it out. Hypercare is the stabilization period after go-live, typically 4–8 weeks, when the partner fixes issues and supports the first month-end close. Plan for reduced productivity and keep partner resources contracted through at least that first close.
How much extra time should we allow for each additional country?
If the ERP has native localization for the country, a template-based rollout often takes 3–6 months per wave. Where e-invoicing, statutory reporting or payroll needs a partner add-on or third-party provider, add roughly 4–12 weeks for certification and testing. Confirm local compliance requirements with qualified tax advisors in each country.
Written by the Carrier Deskapp editorial team. Prices, features and rules change often - confirm details with the vendor or a licensed professional before you decide.