Oracle EBS to Oracle Cloud ERP: a roadmap for mid-market companies.
Premier Support for Oracle EBS 12.2 ends in 2034. The procurement, design, and migration timeline for mid-market companies is 18–30 months. Trevera builds the roadmap.
Premier Support ends December 2034.
Working backward from sunset: 18–30 months of execution, plus 6–12 months of procurement and executive alignment. The practical decision window is 2027–2030.
Why 2034 means decide now.
Oracle EBS 12.2 Premier Support runs through December 2034. After that, Sustaining Support continues indefinitely — but excludes new patches, regulatory updates, and tax updates. For a US multi-state manufacturer or a multi-country distributor, that is the practical end of EBS as a viable production ERP.
Working backward from 2034: a typical mid-market multi-pillar migration is 18–30 months from kickoff to go-live. Add 6–12 months for procurement, business case, and executive alignment. That puts the practical decision window at 2027–2030 for most companies — and if you are not already planning, you are already behind schedule.
The companies that wait until 2032 will pay a premium for accelerated implementations, rushed change management, and reduced consultant availability as the industry's bench fills up. Plan now, execute carefully.
Four ways to leave EBS.
We scope all four during Assessment. The right answer depends on your customization volume, operational risk tolerance, and how much of the next 30 months you want spent in transition.
| Path | Best for | What it is | Timeline | Range |
|---|---|---|---|---|
| 1. Lift & shift to OCI | Buying 2–4 yrs | Same EBS, cloud-hosted | 3–5 mo | $150K–$350K |
| 2. Re-implementation (clean slate) | Most common | New design, modern process | 12–18 mo | $500K–$1.5M |
| 3. Hybrid (Cloud + EBS) | Niche modules retained | Cloud for core, EBS for specialty | 9–14 mo | $400K–$1.2M |
| 4. Phased pillar migration | De-risk by sequencing | Financials → Procurement → SCM → HCM | 15–24 mo | $700K–$2M |
Five phases. Real durations.
| Phase | Duration | Key outputs |
|---|---|---|
| 1. Assessment | 4–6 weeks | Customization + integration inventory; business case; sign-off |
| 2. Design | 3–5 months | TO-BE process; config decisions; customization scope |
| 3. Build | 5–9 months | Configuration; customization; integration; data migration |
| 4. Test + UAT | 2–3 months | Real-data testing; defect resolution; readiness sign-off |
| 5. Cutover + Hypercare | 3 months | Cutover weekend; day-one validation; 60-day hypercare |
Most EBS customizations don't migrate.
The single biggest source of EBS migration scope creep is "we have to bring all our customizations forward." Most EBS install bases have customizations in three categories:
- Native in Cloud ERP now. EBS could not do something natively in 2010, so you customized. Cloud ERP can do it natively today. Drop the customization.
- Genuinely unique business logic. Re-implement as a Cloud ERP extension using OIC or VBCS. Usually 10–20% of customizations.
- Historical decisions. Customizations that exist because someone made a decision 8 years ago and no one remembers why. Drop after stakeholder sign-off.
Trevera does the customization inventory and disposition during the Assessment phase — before you commit budget — so you have a realistic picture of what actually needs to move.
What "disposition" looks like in practice
Each customization gets logged with: original business need, current usage, native Cloud ERP equivalent (if any), recommended disposition (drop, replace with config, replace with extension, retain via integration), estimated rebuild effort, and stakeholder owner.
Disposition is a decision document, not an inventory. It moves through finance, ops, and IT for sign-off so the call is explicit and the budget reflects reality.
How much history actually needs to move?
Less than people think. Most "we have to bring all the history" requirements collapse under scrutiny. Open transactions, balances, and master data are non-negotiable. Recent transactional history is selective. Full archive lives read-only in EBS for 3–7 years.
| Data class | Move forward? | Notes |
|---|---|---|
| Open transactions | Always | Migrate forward |
| Trial-balance / period balances | Always | Validated against legacy |
| Customer + vendor masters | Always | Cleaned + deduplicated |
| Item master + BOMs | Always | Rationalized before migration |
| Recent transactional history | 12–24 months typical | Selective by criticality |
| Full transactional history | Rarely | Read-only EBS retained 3–7 yrs |
| Custom tables / schema | Selective | Per disposition outcome |
When hybrid is the right call.
Hybrid is right when one of these holds:
- You run an EBS module that has no Cloud equivalent yet — niche manufacturing or industry-vertical functionality Oracle has not finished migrating.
- The risk of replacing a critical operational system in lockstep with financials is unacceptable to the business.
- Budget reality forces sequencing, and Financials + Procurement is the first wave that pays for itself.
The pattern: Cloud ERP for Financials, Procurement, EPM, and Reporting; EBS retained for niche modules; OIC bridges the two; data lake aggregates for reporting. The hybrid window is typically 18–36 months before the EBS-side modules also migrate.
EBS migration questions.
Direct answers, no marketing softening. If yours isn't here, ask it directly on the contact page.
01 When does Oracle EBS Premier Support actually end?
02 Should we go to NetSuite or Oracle Cloud ERP?
03 What does the migration roadmap actually look like?
04 How do we handle EBS customizations?
05 What does it cost?
06 Can we go hybrid — keep some EBS modules and migrate others?
07 How is data migration handled?
08 What about lift-and-shift to OCI as a stopgap?
09 How does Trevera staff a multi-year EBS engagement differently from a Big SI?
10 What about customizations Oracle never gave us a Cloud equivalent for?
Have a NetSuite or Oracle Cloud project on the table?
We'll tell you what's realistic, what it'll cost, and where it could go wrong — before you sign anything.