Session 16 found the gap and session 17 mapped the walls; this session is the doors. Five reduction paths, full set termination, restructure and repurchase, unsupported deployment, third party support, and the negotiated reduction, each with its mechanics and honest risk profile, the reinstatement mathematics that make most moves one way, the five component business case, and the retention counter offer that turns out to be the negotiated path working exactly as designed. The $1.2M bill ends the session at $805K.
The presenter in this session is an AI generated avatar. The curriculum and guidance are real, produced by Redress Compliance analysts from our consulting engagements and market network.
A taught session with three knowledge checks: the ghost set and the mixed set assigned to their different paths, the reinstatement mathematics priced two years after dropping support, and the 35 percent retention counter compared against the alternative it was built on. It closes with the $1.2M bill from session 16 landing at $805K after one program year.
The full narration of this session, section by section, for reading and reference.
Welcome back, session eighteen of forty. The last two sessions built the diagnosis: session sixteen found the gap in the support bill, roughly forty percent of a typical estate's spend delivering little or nothing, and session seventeen mapped the walls, two policies that make naive reduction impossible. Today, the doors. This is the playbook session: five reduction paths, each worked properly, with its mechanics, its natural targets, and, because honesty is the house style, its real risks, including the ones Oracle's retention team will emphasize on the phone. By the end you'll be able to assign every finding from your baseline audit to a path, build a business case that survives both repricing arithmetic and executive review, and execute the program through a renewal season, notices, timing, counters and all. One framing note: nothing today is aggressive or clever. Every path is contractual, published, and practiced across the industry every year. The goal is a program so procedurally boring that the only interesting thing about it is the number at the end. Let's open the playbook.
Five takeaways. One, you'll choose the path: five real options, and the skill is matching each reduction candidate to the right one, which mostly means asking session seventeen's question, what is the set, before anything else. Two, you'll price the risks honestly, each path has a profile, and pretending otherwise produces programs that die at the first security review. Three, you'll judge unsupported deployment properly: running perpetual licenses without support is a genuine right with genuine savings, standing behind a door that only swings one way, and the reinstatement mathematics get their own knowledge check. Four, you'll build the case, the five components that carry a reduction program through finance, security, and the executive committee. And five, you'll execute cleanly: notices done perfectly, everything on renewal dates, and the inevitable retention counter offer read for what it actually is, which, spoiler, is the negotiated path working exactly as designed. The stakes, in numbers, next.
Four numbers. Five, the real reduction paths, and note the pattern they share before we even name them: every one works at the set level, on a renewal date, with proper notice, exactly the shape session seventeen's pincer demands. The policies didn't eliminate reduction; they standardized its form. Twenty to forty percent, the reduction range disciplined programs actually achieve on bloated support estates, without touching a dollar of genuinely needed coverage. That range is not aspiration; it falls straight out of session sixteen's dissection, where roughly forty percent of the bill was shelf, ghost, and frozen. One hundred fifty percent, the reinstatement mathematics: support that lapses and is wanted back costs the unpaid years plus a penalty, which makes most reduction moves effectively permanent, a fact that belongs in every business case in bold. And one, roughly the number of retention counter offers you should expect per termination notice on real money. The counter is not a complication; for one of the five paths it is the entire mechanism, and pricing it in advance is what separates a program from a sequence of surprises. The paths, all five, now.
The five paths, each with its natural target from the audit's four way sort. Path one, full set termination, the clean kill: a complete, separable set that nobody needs, ghosts and pure shelfware sets, terminated whole at renewal. Final, repricing proof, and the cheapest money in the program. Path two, restructure and repurchase, the engineering route: for mixed sets where real value is trapped alongside padding, buy a right sized new set at negotiated prices, then terminate the old bloated set entirely. More work, bigger prize, full mechanics in a few slides. Path three, unsupported deployment, the bold path: stable systems on frozen versions keep their perpetual licenses and stop buying updates nobody applies. Real savings, real managed risks, and the one way door. Path four, third party support: whole estates on aging versions move to an independent provider at roughly half cost, session nineteen gives it the full hour it deserves. And path five, the negotiated reduction: support relief Oracle grants at leverage moments rather than lose a stream entirely, never offered, only won, and won in a very specific way you'll see at the final check. The paths compose: one program year can run all five on different parts of the same estate. Now the honest comparison.
Five paths, honestly compared, four columns. Full set termination: saves one hundred percent of the set, low risk, fits separable and genuinely unneeded sets. If the ghost hunt found clean prey, this is its path, and there's rarely a reason to hesitate. Restructure and repurchase: saves the shelfware share of a mixed set, medium risk, mostly execution complexity and the bridge period, fits wherever value is trapped in bloat. Unsupported deployment: saves one hundred percent of the set's support, and I've labeled its risk managed but real, because that's the truth, security exposure and the frozen version are genuine, and they're handled by selection and controls, not by optimism. It fits frozen, stable, isolated systems, and nothing else. Third party support: saves about half, medium risk centered on the provider transition and the upgrade question, fits aging versions with no upgrade path planned. And the negotiated reduction: ten to thirty percent, low risk, fits whenever leverage exists and alternatives are priced, which, if you run the other four paths credibly, is always. One clarification on the risk column: it means operational and compliance exposure, not contractual danger. Every path is within your rights. The exposures are specific and manageable, and each one gets named today. First matching exercise, knowledge check one.
Knowledge check one. The baseline audit found a hundred eighty thousand of ghost support on a separable set, and two hundred sixty thousand of shelfware trapped in a set shared with production licenses. Same path for both? A, yes, terminate both at the next renewal. B, no: the ghost set terminates whole and cleanly, but the mixed set needs restructuring, because partial termination triggers repricing. C, yes, move both to third party support. Or D, no, the shelfware must simply stay forever. Pause here. Session seventeen's question: what is the set?
The answer is B, and the reasoning is the whole discipline in one example. Two findings, two set situations, two different paths. The ghost set is separable and completely dead: path one, full set termination, notice at renewal, a hundred eighty thousand recovered, no repricing because there are no survivors. Clean. The shelfware is a different animal: it shares a set with production licenses, so terminating just the shelfware is a partial termination, and session seventeen's table showed what repricing does to those, the saving collapses by eighty percent. Its path is restructuring: size what production actually needs, buy that as a fresh right sized set at a leverage moment with session nine discounts, then terminate the entire old set, production licenses replaced, shelfware extinguished, no survivors to reprice. More engineering, same destination. The wrong answers: A applies the clean path to the dirty problem and funds the repricing lesson with real money. C misfits the tool, third party support serves systems that need support cheaper, not licenses that need support gone; moving shelfware to half price support is still paying for nothing, just less. And D is the learned helplessness this module keeps retiring: trapped shelfware has an exit, it's a project rather than a letter, and the project pays two hundred sixty thousand a year forever. Map the set, then choose the path. Always that order. Now, the bold path.
Running unsupported, eyes fully open, four facts. Fact one, the right is real: perpetual licenses run forever, support is optional, and terminating support on a set while continuing to run the software is entirely legitimate, no matter how the retention call makes it sound. Fact two, what stops arriving: patches, security fixes, new versions, and the ability to raise service requests. Now be honest about the target systems: a frozen application on a five year old version that hasn't applied a patch since the last hardware refresh is receiving those things in theory only. For that system, the loss is smaller than the invoice implies, which is exactly why the audit's sustaining support category exists. Fact three, what must be managed, and this is where programs earn their keep: security exposure gets real mitigation, network isolation, compensating controls, and a security team signature, not a shrug; the version freezes permanently, so upgrade futures must genuinely be off the table; and the licenses remain licenses, auditable, so module one's records discipline continues untouched. And fact four, the near one way door: reinstatement bills the lapsed years plus a penalty, the hundred fifty percent mathematics, so the decision is made as if return were impossible, because economically it nearly is. Let's price that door properly. Knowledge check two.
Knowledge check two. Two years after dropping support on a set to save two hundred thousand a year, an upgrade is suddenly needed after all. What does getting support back cost? A, nothing, support resumes at the old rate. B, the new year's fee only, at current pricing. C, roughly the unpaid back years plus a reinstatement penalty, often making repurchase of new licenses the cheaper route. Or D, support can never be repurchased once dropped. Pause here. What does the one way door cost to reopen?
The answer is C, and the arithmetic deserves to be felt. Reinstatement bills the lapsed period plus a penalty: two years off at two hundred thousand becomes roughly four hundred thousand of back support, plus the penalty on top, plus the resumed annuity going forward. Total those, and buying brand new licenses with fresh support at session nine discounts frequently costs less than reinstating the old ones, an absurdity that is entirely intentional. The policy exists to make dropping support feel irreversible, and the correct response is not fear, it's finality: treat every unsupported decision as permanent at the moment it's made. Which is why the path's fit criteria are so narrow, frozen, stable, going nowhere systems, and why anything with an upgrade anywhere in its future belongs on a different path. The wrong answers: A and B describe pricing models Oracle has never offered, and estates that assume either one walk through the door casually and meet the mathematics two years later, at maximum surprise. D overshoots, the door does reopen, it just costs more than a new door, which for planning purposes is the same thing with extra steps. The planning rule, worth writing into every business case: before any unsupported move, write down the specific scenario that would force a return, price that scenario at reinstatement rates, and if it's plausible, choose another path. The savings are real. So is the commitment. Now, the engineering path in detail.
Restructure and repurchase, the mechanics, four steps. Step one, size the real need: from the audit, what does production actually require, in licenses and metrics, today and across the planning horizon. Not what the old order contained, what the workload needs, counted with module one's rules. This number is usually startlingly smaller than the current set. Step two, buy the new set right: a fresh order, negotiated at a leverage moment with session nine benchmarks, structured with session seventeen's set architecture, carrying session eight's clauses. The new set is clean by construction, right sized, separable, capped, because you built it knowing everything this course teaches. Step three, terminate the old set whole: with the replacement live, the bloated set ends completely at its renewal date. No survivors, no repricing, no residue, the whole maneuver exists to create this clean ending. And step four, mind the bridge: there's an overlap period where both sets are briefly alive and paying, and that bridge is the maneuver's real cost, so price it into the case honestly, and then negotiate it, because Oracle, facing the loss of the old stream entirely, will sometimes credit the transition to win the new order. The restructure looks elaborate on a slide. In practice it's a purchase and a letter, sequenced. Now, the case that gets it approved.
The business case, five components, because reduction programs die in internal review far more often than in negotiation with Oracle. Component one, the baseline: session sixteen's audit, current, every stream sorted, set boundaries verified from the paper. Unverified baselines produce unverifiable savings, and finance can smell those. Component two, the repricing model: session seventeen's table built for every candidate move, and all savings quoted net of repricing, always. The fastest way to lose a room is to present gross savings that an Oracle account manager later corrects. Component three, the risk register: per path, the exposure, the mitigation, the named owner. Unsupported candidates get security sign off before finance sign off, in that order, because a security veto after budget approval kills programs dead. Component four, the calendar: every move mapped to its renewal date and notice period, because a perfect plan a week past its notice window waits a full year, and momentum rarely survives the wait. And component five, the sponsor: support reduction crosses IT, procurement, security, and finance, four departments with four agendas, and it needs one executive who owns the total number or it fragments into nothing. The pattern across all five: boring. Verified numbers, net savings, named risks, dated moves. Boring cases get approved, and approved cases save millions. Execution, next.
Execution discipline, five habits. Habit one, notice perfectly: written termination notice, per the policies' requirements, inside the notice window, to the specified address, with delivery proof retained. An hour of procedural perfection removes every cheap objection before it can be raised, the same lesson certification taught in module three. Habit two, never mid term: every move lands on a renewal date. Mid term gestures change nothing contractually and merely signal your intentions a season early, handing the retention team extra preparation time. Habit three, keep running the records: terminated support does not mean terminated licenses. The estate remains fully auditable, unsupported licenses still count in every compliance calculation, and module one's evidence discipline continues on every set, forever. Reduction programs that relax the records invite the audit that eats the savings. Habit four, expect the retention call: a termination notice on real money reliably produces a counter offer within weeks. Anticipated, it's leverage arriving on schedule; unanticipated, it's confusion. And habit five, bank the wins in writing: every discount, credit, and cap conceded to keep your business goes into the renewal paperwork, multi year where possible, session eight style, because verbal generosity from a retention desk has the shelf life of the quarter that produced it. Which brings us to that call. Final check.
Knowledge check three. Three weeks after your termination notice on a four hundred thousand dollar stream, Oracle counters: stay, at thirty five percent less, for three years, in writing. What now? A, decline on principle, the termination must proceed as planned. B, accept immediately, thirty five percent is thirty five percent. C, compare it against the alternative the notice was built on: if the discounted stream beats the alternative's economics and risks, take it in writing; if not, proceed. Or D, withdraw the notice and restart the analysis from zero. Pause here. What was the notice actually for?
The answer is C, and understanding why reframes the entire program. This counter offer is not an interruption of your plan; for the fifth path, the negotiated reduction, it is the plan. Support relief gets granted exactly here, when a credible termination is in motion and a real stream is genuinely about to end, and at no other moment, which is why polite requests for support discounts accomplish nothing in twenty years while termination notices produce them in three weeks. So the decision is pure arithmetic, and you already did it: the notice was built on an alternative, third party at half cost, unsupported at zero, a restructured repurchase, with its economics and risks priced. If two hundred sixty thousand a year, locked for three years, in writing, beats that alternative, taking it is victory, a thirty five percent reduction won by preparation, and nobody should feel out maneuvered into it. If the alternative still wins, the termination proceeds calmly, because it was always real, which is precisely why the counter appeared at all. The wrong answers: A punishes yourself to prove a point nobody will remember, momentum is not strategy. B takes a first counter unexamined, and first counters are rarely final, at minimum the uplift cap and terms deserve session eight treatment before signing. D surrenders three weeks of leverage at its moment of maximum effect. The standing rule: never send a notice you wouldn't execute, and never execute one without reading what comes back. That's the entire fifth path, and it just paid a hundred forty thousand a year. Let's total the program.
The one point two million dollar bill from session sixteen, one program year later, category by category. Active, current, needed, seven hundred thousand: untouched in coverage, renegotiated in price, renewal run properly with the uplift capped at three percent, landing at six hundred eighty. The genuinely valuable support got cheaper without getting thinner. Shelfware in mixed sets, two hundred sixty thousand: the restructure path, a right sized replacement set purchased at negotiated discounts, the old bloated set terminated whole, new support stream ninety five thousand. The trapped money, freed by engineering. Ghost support, one hundred eighty thousand: full set termination at renewal, notice letter, delivery receipt, zero. The easiest line in the program, exactly as promised. And sustaining support on frozen versions, sixty thousand: moved to third party at roughly half, thirty thousand, coverage continuing under a different model, session nineteen's story. Total: eight hundred five thousand, down from one point two million, a thirty three percent reduction, with not one dollar of needed coverage touched, every move contractual, noticed, and dated. The program cost a few weeks of analysis and one properly run renewal season, and it repeats its savings every year afterward. That's the playbook, executed. That's the session.
Session eighteen in three sentences. One, five paths close the gap, full set termination, restructure and repurchase, unsupported deployment, third party support, and the negotiated reduction, and the set boundary from session seventeen chooses which path fits each finding. Two, unsupported deployment is a real right behind a near one way door, priced by reinstatement mathematics at roughly one hundred fifty percent, so it's decided as permanent and reserved for frozen, stable, isolated systems with security's signature. Three, execution is notices done perfectly on renewal dates, records that never stop, and retention counters read as the mechanism working, compared against the priced alternative and banked in writing when they win. Next session, the alternative that disciplines every one of these conversations: third party support. The economics at roughly half price, the legal history that established the industry, the security question answered honestly, and who it actually fits, because it is neither the boogeyman Oracle describes nor the free lunch the providers advertise. The most leveraged phone number in the module. See you in session nineteen.
Homework, about an hour, and it drafts your actual program. One, path your findings: every line from the baseline audit assigned to one of the five paths, or explicitly marked keep as is. One added column, and the whole program exists in draft. Two, model one restructure: your worst mixed set, price the right sized repurchase against the current stream, net of the bridge period. If the number works on your worst set, the path is real for your estate. Three, test one unsupported candidate: the most frozen system you own, write its return scenario, price that scenario at reinstatement rates, and see if the path survives the honesty. Four, draft one notice: a termination letter for your cleanest candidate, written to the policies' actual requirements, address, window, form. Unsent, it costs nothing; drafted, it makes the program feel as real as it is. And five, list the leverage: every renewal date and pending purchase in the next twelve months, because those are the moments when retention counters can be provoked, compared, and banked. That's the hour, and honestly, that's the program. See you in session nineteen.
Five reads, all free on redress compliance dot com. First, optimizing your Oracle license footprint before renewal, the reduction program in full methodology, today's playbook with worksheets. Second, the Oracle support renewal contract checklist, the notice mechanics and clauses behind every move you'd make. Third, Oracle renewal negotiation strategy, provoking and banking the retention counter as deliberate strategy, the fifth path at article length. Fourth, the Oracle contract renewal strategy guide, the renewal season run as a coordinated program across an estate. And fifth, the Oracle vendor management guide, the organizational ownership question, because five paths need one owner. That's session eighteen. Five doors through session seventeen's walls, a business case built to be boring, notices that land on dates, and a retention counter that turned out to be the fifth path introducing itself. The bill is down a third and nothing needed was touched. Next session: the alternative that makes every renewal conversation honest, third party support, examined properly. See you there.