Oracle will tell you the move from perpetual Retail licences to Merchandising Cloud Service is a modernisation decision. It is a licensing decision first, and the numbers only work if you fix the renewal mechanics before you sign the initial term.
Oracle will tell you the move from perpetual Retail licences to Merchandising Cloud Service is a modernisation decision. It is a licensing decision first, and the numbers only work if you fix the renewal mechanics before you sign the initial term.
Start with the fact Oracle's account team will not put in writing: there is no conversion path from a perpetual Oracle Retail licence to Merchandising Foundation Cloud Service. You do not trade in entitlement, you do not receive credit for historic net licence fees, and you do not carry forward the discount you fought for in 2011. You buy a new subscription service alongside licences you already own, and on the day you cut over and stop paying support, every dollar of that historic licence spend becomes sunk. In 25 years of taking these deals apart, I have never seen Oracle grant a genuine licence-for-subscription credit in Retail. What buyers occasionally get is a one-time "migration accelerator" discount on the first subscription term, which is a discount on new spend dressed up as recognition of old spend.
The structural asymmetry sits in how each model calculates its recurring fee. On perpetual, you pay a one-time net licence fee, then annual support at 22 percent of that net licence fee, where "net" means Oracle's list price at time of purchase less the original discount. The base is historic and fixed. Oracle's standard ordering language permits an 8 percent annual uplift, and in practice renewals land between 4 and 8 percent. On the SaaS side you are not buying one thing at all. You are buying a stack of separately priced SKUs, MFCS, Retail Pricing Cloud Service, Retail Integration Cloud Service, Order Management Suite and Xstore, each with its own metric: Annual Subscriber Revenue for merchandising and pricing, Hosted Named User per user per month elsewhere, and per lane or per store for Xstore registers and lanes. And the renewal fee is not a percentage of a fixed historic base at all. It is a percentage of Oracle's then-current list price, less whatever discount you can still defend at that moment. That is the whole game.
On-prem support is a percentage of a frozen historic base; cloud subscription is a percentage of whatever Oracle's list price says on renewal day.
Here is the comparison your business case needs, modelled on a mid-size retailer with an existing perpetual Retail estate carrying roughly $1.6M in annual support. Perpetual assumes support uplift at 6 percent, the midpoint of the observed 4 to 8 percent renewal band. Subscription assumes a $2.4M annual list rate discounted 60 percent in the initial term, with the discount compressing to 30 percent at first renewal in year four and uplift running at 10 percent thereafter, the midpoint of the 8 to 12 percent uncapped range. Third-party support assumes the 50 percent saving typical for stable Oracle workloads. Year one perpetual excludes the original licence purchase, which is already spent.
| Year | Perpetual + Oracle Support | Cloud Subscription (60% then 30%) | Perpetual + Third-Party Support |
|---|---|---|---|
| 1 | $1,600,000 | $960,000 | $800,000 |
| 2 | $1,696,000 | $1,056,000 | $848,000 |
| 3 | $1,797,760 | $1,161,600 | $898,880 |
| 4 | $1,905,626 | $1,848,000 | $952,813 |
| 5 | $2,019,963 | $2,032,800 | $1,010,101 |
| 5-year total | $9,019,349 | $7,058,400 | $4,509,674 |
Read the crossover, not the total. Against Oracle support, cloud stays cheaper through year four and crosses in year five. Now rerun it with the discount held at 60 percent through renewal: cloud never crosses inside a ten-year horizon. That single variable, discount compression from 60 to 30 percent at first renewal, pulls the crossover forward by two to three years and swings roughly $1.9M across five years on this profile. The revenue-band metric behind Merchandising Cloud Service compounds it, because a band uplift on top of discount compression stacks multiplicatively, not additively.
The third column is the one most business cases omit, and it is the reason Oracle discounts at all. Perpetual licences remain valid without support. You lose new patches and Oracle technical support, nothing else, and third-party support runs at roughly half the fee. Over five years that path costs $4.5M against $7.1M for cloud on the numbers above. It is not automatically the right answer, and Oracle's reinstatement policy demands back-support for every lapsed period if you ever return, so treat it as a one-way door. But it is a credible, quotable, board-defensible alternative, and it is the only anchor that reliably moves an Oracle Retail cloud quote. Get a third-party support proposal in writing before you open the cloud negotiation, not after. Reps price against the alternative they believe you have.
Strip away the transformation narrative and the 2026 Oracle Retail negotiation cycle has one driver: the Retail Lifetime Support Policy calendar. Merchandising System 19.0 went GA in January 2020, which puts Premier Support expiry at January 2027 and Extended Support at January 2030, with Sustaining Support indefinite after that. RMS 16.0 lost Premier in December 2023 and loses Extended in December 2026. RMS 15.0's Extended window closed in December 2025, 14.1's in December 2024, and the 12.0 through 13.2.9 estates, along with Price Management, Point-of-Service, Central Office, and Returns Management, dropped off Premier on 31 December 2019. Oracle's account teams know these dates better than most retail CIOs do, and they time the cloud proposal to land roughly twelve months ahead of the cliff, when your leverage is thinnest and your only visible alternatives are Extended Support uplifts or a rushed subscription.
| Retail release | Premier Support ends | Extended Support ends | Negotiation posture in 2026 |
|---|---|---|---|
| RMS 19.0 (GA Jan 2020) | January 2027 | January 2030 | Extended is a real option; use it to buy 24 to 36 months of negotiating room |
| RMS 16.0 (GA Dec 2016) | December 2023 | December 2026 | Cliff is inside the fiscal year; expect an aggressive cloud proposal |
| RMS 15.0 | Expired | December 2025 | Already on Sustaining; audit and PCI exposure is live now |
| RMS 14.1 | Expired | December 2024 | Sustaining only; no new CVE fixes |
| RMS 12.0 to 13.2.9 | 31 December 2019 | Not applicable | Sustaining only for six years and counting |
Two facts break the assumptions most boards operate on. First, Applications Unlimited, Oracle's public commitment that Premier Support on Continuous Innovation releases for on-premise applications will not be discontinued before 2037, does not extend to Oracle Retail. Retail sits under its own Lifetime Support Policy with hard, dated cut-offs. If your architecture team quoted 2037 in a steering paper, correct it in writing before the CFO builds a plan on it. Second, Sustaining Support is not a security posture. It lasts as long as you hold the licence and gives you technical assistance plus knowledgebase access, but it delivers only fixes, updates, and critical patch updates that were created during the Premier and Extended phases. A CVE discovered after your window closed will never be patched by Oracle. For a retailer processing card data, that converts a support decision into a PCI DSS compensating-control problem and an audit committee disclosure, not an IT backlog item. Map the release, the exact date, and the CVE exposure per component now, and read it alongside what an Oracle audit examines across the retail estate, because Sustaining Support customers get asked for deployment data more often, not less.
Applications Unlimited's 2037 promise does not cover Oracle Retail, and Sustaining Support will never patch a CVE discovered after your window closed.
Every on-premise Oracle Retail licence carries restricted-use entitlements underneath it: Oracle Database, WebLogic Server, and associated middleware, granted solely to run the Retail application and nothing else. Oracle's own Retail licensing documentation is explicit that prerequisite products, entitled products, and restricted-use licences do not apply to Oracle Retail Cloud products. There is no cloud equivalent and no conversion credit. The subscription price you negotiate is the whole entitlement. In practice, that creates two exposures we see repeatedly on the buy side. The first is the hybrid retailer who moves Merchandising to cloud, keeps Allocation or a legacy pricing engine on-premise, and assumes the restricted-use Database grant survives. It survives only for the Retail products still licensed and supported, and only within the original restrictions. The second, and the expensive one, is the retailer who drops on-premise Retail support entirely but leaves reporting schemas, data marts, integration staging, or a warehouse extract running on that same database instance. The moment the underlying Retail entitlement goes, those workloads are unlicensed full-use Database and WebLogic, priced per processor at list, with backdated support.
Do the inventory before you sign anything, not after the cutover. Enumerate every instance, every schema, every WebLogic domain, and every scheduled job touching the restricted-use estate, and record which Retail product number justifies each one. Then price the full-use replacement for anything that does not map cleanly. That number belongs in the cloud business case as a migration cost, because Oracle will not volunteer it. Our detailed treatment of the boundary sits in the restricted-use Oracle Database hiding under Oracle Retail applications, and it should be circulated to whoever owns database compliance, not only to the retail programme team.
The entry price on an Oracle Retail cloud deal is the number your CFO looks at, and it is the least important number in the contract. The standard Cloud Services Agreement reserves Oracle's right to apply prevailing list price at renewal, and the discount you negotiated is generally not contractually preserved past the initial term. That is the whole mechanic in one sentence: a 60% initial-term discount is a term-limited concession, not a rate, and it can compress to 30% or to nothing when the renewal quote lands. Across deals we have seen, first renewals after an initial term carried 7% to 12% uplifts where no cap existed, and Oracle's practical posture on uncapped cloud renewals runs in the 8% to 12% band. Nobody signs up for that. They sign a five-year TCO model built on Year 1 pricing and discover in Year 4 that the model was fiction.
Understand why this happens and you stop treating it as bad luck. Oracle cloud reps are credited on Annual Recurring Revenue, not first-year billings. A rep who discounts Year 1 to 40% of list and loads the recovery into Years 2 and 3 is not being aggressive, they are being paid correctly by their own comp plan. The discount you extract in the closing week of Oracle's quarter is the cheapest thing they can give you, because it does not touch the ARR they get measured on. This is why the negotiation you must win is the renewal rate, not the entry rate, and why a discount concession offered late in the cycle should be traded for cap language rather than banked as a win.
The discount you extract in the closing week of Oracle's quarter is the cheapest thing they can give you, because it does not touch the ARR they get measured on.
Quantify the fix so it survives your own finance review. A 3% cap applied to a $4M annual stream returns roughly $1.2M of avoided uplift over five years, which on most Retail estates dwarfs anything you will win on the entry discount. The benchmark language is the lower of CPI or 3% to 5%, applied to per-unit rates, across the full initial term and the first renewal period. Some buyers push harder, to "not to exceed 2% or CPI, whichever is lower," and get it when the deal is large enough. Three drafting details decide whether the cap is real: it must attach to per-unit rates rather than total contract value (otherwise band growth defeats it), it must survive into the first renewal term rather than expiring with the initial term, and it must not carry a carve-out for "changes to the service" that lets Oracle reprice on a module rename. If your rep says renewal caps are non-standard, they are. They are also routinely granted. Do not sign the order form without one, and read the revenue-band pricing mechanics before you agree what the cap applies to.
The second overspend is quieter and larger. Contracted volumes routinely exceed actual usage by 10% to 25%, and in Retail that shows up as an Annual Subscriber Revenue band bought above your real comparable-store revenue. Oracle sells the band you might reach, not the band you occupy. Size to current measured revenue, negotiate a documented step-up rate for the next band, and make the step-up price fixed at signature rather than list-at-the-time. Buying headroom you do not use is the same as paying an uplift, except you pay it from Year 1.
Then close the sleeper clause: what actually counts inside Annual Subscriber Revenue. Pin every element in writing before signature, because Oracle will not define it for you and the ambiguity always resolves upward.
Freeze that definition for the full term and the first renewal so Oracle cannot reclassify mid-term. The same discipline applies to Xstore, which is sold per POS Lane or per Store on either subscription or perpetual depending on deployment. A hybrid retailer running a phased rollout can pay twice for the same lane throughout dual-run unless the order form carries an explicit dual-run credit or transition allowance. Fix the lane and register counting rules in the same conversation, not after the migration plan is set.
Do not open the cloud conversation until you have a walk-away number, because Oracle prices against your perceived lack of alternatives. Step one is to establish the fallback stack on the perpetual side: get two third-party support quotes (the market benchmark is roughly 50% of the Oracle stream) and model the Sustaining Support path in parallel, understanding that Sustaining gives you technical support and existing patches but no fixes for new CVEs. That combination is your leverage. Step two is the five-year comparison, and it must be built with discount compression baked in, not Oracle's flat-rate spreadsheet. Oracle's Cloud Services Agreement reverts renewals to then-current list less any negotiated discount, and the discount is generally not preserved past the initial term, so a 60% Year-1 discount can compress to 30% or zero. Model the renewal at a 7% to 12% uplift (the observed range where no cap existed) and see whether the business case survives. In my experience across these deals, roughly half do not survive that single change of assumption. Only after both artefacts exist should you negotiate.
Bring peer data to the table. Oracle publishes no list price for Retail Cloud Services (Capterra's 2026 listing still shows "contact vendor"), so benchmarking against comparable retailers of similar revenue band is your only price discovery mechanism. Without it you are negotiating against a number Oracle invented.
No. The two licensing models do not convert into each other. Migration replaces your on-prem licences with a new subscription service, so the perpetual investment is sunk unless you keep an on-prem instance running under its own entitlement. Any credit Oracle offers is a commercial concession on the subscription price, not a legal conversion of entitlement, and it should be documented as a discount schedule rather than described as a trade-in.
RMS 19.x moves to Extended Support through January 2030, typically at an uplift, and then to Sustaining Support indefinitely. Sustaining Support keeps your technical support and knowledgebase access but delivers only fixes, updates, and critical patch updates that were created during the Premier and Extended phases. It produces no new security patches for CVEs discovered after that point, which becomes a PCI scoping and audit-committee issue for a POS-connected estate long before it becomes an IT problem.
On a $4M annual stream, holding uplift to 3% instead of Oracle's uncapped default returns roughly $1.2M of avoided cost over a five-year term. The benchmark to push for is the lower of CPI or 3% to 5%, applied to per-unit rates and covering both the initial term and the first renewal period. The renewal period matters more than the initial term, because that is where the compression from a heavily discounted Year 1 rate actually lands.
No. Applications Unlimited commits that Premier Support on Continuous Innovation releases for certain on-premises applications will not be discontinued before 2037, with annual review. Oracle Retail sits under a separate Retail Lifetime Support Policy with hard, dated cut-offs. If a sales team cites the 2037 date in a Retail conversation, ask them to put the specific product and release under that commitment in writing, and check it against the Retail LSP PDF.
It disappears. Prerequisite products, entitled products, and restricted-use licences do not apply to Oracle Retail Cloud products, so there is no cloud equivalent to inherit. If you have adjacent workloads, reporting schemas, or integrations sitting on that restricted-use database, they lose their licensing basis when the Retail on-prem entitlement lapses and become full-use exposure. Inventory every schema on that instance before you sign the cloud order.
It is a legitimate option and, more importantly, it is the anchor that makes Oracle discount the cloud deal. Perpetual licences remain valid without support; you lose future patches and Oracle technical assistance only, and third-party support typically saves around 50% of the annual fee. The catch is reinstatement: if you later return to Oracle support, Oracle requires payment of back-support for every lapsed period, so treat this as a one-way door and price it as such.
Oracle prices Fusion ERP Cloud per employee, not per user, which inflates true cost. The buyer side guide to module economics and the modernization discount.
Gated with a work email on the download page. No sales follow up you did not ask for.
Get the White Paper →
500+ enterprise clients. 11 vendor practices. Industry recognized. One conversation can change what you pay for the next three years.
One buyer side briefing a week. Renewal signals, audit moves, and the levers that work. No vendor spin.
