Certification converts three years of unlimited deployment into permanent entitlement at zero marginal cost, which makes the count date the most valuable afternoon in the agreement's life. This session runs the exit step by step: the T minus twelve runway, what counts and what does not, every move in Oracle's exit playbook with the calm answer to each, the declaration letter and its evidence file, and the five avoidable mistakes that turn winnable exits into renewals.
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 T minus six acceleration question (real deployments versus theater), the helpful scripts offer declined per the clause, and the free extension offered at T minus one. It closes with the certification letter itself: 210 processors certified from a 96 processor entry forecast, product by product, evidence behind every number.
The full narration of this session, section by section, for reading and reference.
Welcome back, session fourteen of forty, and this is the one the whole module has been building toward: certification, the ULA exit. Session eleven called it the most valuable afternoon in the agreement's life. Session twelve wrote its rules into the entry paper. Session thirteen spent three years building its records. Today we spend them. Here's the shape of the session: the twelve month exit runway, milestone by milestone. The counting rules, what counts on the count date, what doesn't, and the edge cases between. Oracle's exit playbook, every move they make, and the calm answer to each. The declaration letter itself, and what happens the day after it's delivered. And the five mistakes, each one avoidable, that convert winnable exits into renewals every single quarter. One promise up front: if session thirteen's register is real, nothing today is difficult. The exit rewards preparation more than any event in Oracle licensing, and you've been preparing for three sessions. Let's finish it.
Five takeaways. One, you'll run the timeline, the T minus twelve to T zero runway, five milestones, each with one job, so the exit is a project and never a scramble. Two, you'll apply the counting rules: the installed and running test, the categories that count, the ones that don't, and the two that depend entirely on language you negotiated at entry. Three, you'll read Oracle's playbook, the renewal call, the careful review warning, the helpful scripts, the finding lever, the slow walk, and you'll have the answer to each one written down before the phone rings. Four, you'll write the declaration, the certification letter that converts three years of unlimited deployment into permanent entitlement, and you'll know exactly what stands behind every number in it. And five, you'll refuse the conversions, the five mistakes that turn exits into renewals, including the friendliest and most dangerous one, which arrives gift wrapped at T minus one. The stakes, in four numbers, next.
Four numbers. T minus twelve, months before term end, when the exit runway starts, and note what that implies: the final dry run happens there, meaning it cannot be the first one. Session thirteen's annual rehearsals were the training; T minus twelve is the last practice before the game. One, the declaration. A letter, counts per product, with the evidence file standing silently behind every number. Silently is the operative word: the evidence rarely gets asked for when everyone can tell it exists, which is a pattern you've seen across this whole course. Zero dollars, the marginal cost of certifying more. Support stays at twenty two percent of the fee whether the letter says ninety six processors or two hundred ten, which converts every properly counted deployment into free permanent entitlement, and every missed one into a donation. That asymmetry is the economics of the entire afternoon. And three, the years of unlimited deployment converting in one document, or failing to. Session thirteen built the records; today spends them. If the register is real, this session is a checklist. If it isn't, this session is a negotiation, and you already know who negotiates well at exits. The runway, next.
The exit timeline, five phases. T minus twelve, the final dry run. The last rehearsal with enough runway to fix what it finds: register coverage gaps, cloud mapping, edge cases, all identified now, all closed by T minus nine. Anything found later than this gets fixed under pressure, which is why this milestone is non negotiable. T minus six, the deployment review, and this one surprises people: the roadmap gets reviewed for genuine, funded workloads landing in the next year, and where real, they deploy inside the term, properly, while the unlimited right still covers them. We'll examine the legitimacy question in the first knowledge check, because it matters. T minus three, the freeze and the count. Change discipline settles over the estate, no surprise migrations, no undocumented decommissions, and then the formal count runs: scripts, register, reconciliation, per product totals, evidence bundled per product. T minus one, the letter: drafted, legal reviewed, signed by the executive the clause names, numbers final. And T zero: delivered per the clause, inside the window, method and address exactly as written. Then the process language from session twelve's markup runs its course. Notice what's absent from this timeline: panic, and Oracle. The exit is your process, on your records, per your clause. Their playbook exists to change that, and it's coming. First, the rules.
The counting rules, five categories. Category one, installed and running, on list, in definition: counts, full stop. And note the breadth: production, test, and development environments all count if genuinely deployed, a point estates that only count production get wrong, in the expensive direction. Category two, planned but not deployed: doesn't count. Certification counts deployments, not intentions, not purchase orders, not roadmaps, which is exactly why the T minus six milestone exists, to convert genuine intentions into countable deployments while the right still covers them. Category three, off list products: never covered, never countable, they surface as findings, not entitlement, session eleven's lesson, now at the moment it prices. Category four, public cloud: depends, entirely, on what the certification clause says, session twelve's sentence decides this row, and for a migrated estate this single row can be half the count. And category five, standby and DR: depends, per session four's failover rules, what required licensing during the term counts like any other deployment, the ten day rule and mirroring distinctions apply exactly as you learned them. The pattern across all five: the count date creates nothing. It converts rights that were negotiated at entry and documented during the term. Which sets up the strategic question at T minus six. Knowledge check one.
Knowledge check one. At T minus six, the roadmap shows thirty processors of genuine expansion, funded, planned, landing next year, just after the term ends. The team proposes accelerating that deployment into the term. Legitimate? A, no, deploying ahead of need to raise the count is abuse of the ULA. B, yes: deploying real, planned workloads inside the term is exactly what the unlimited right is for, and they count if installed and running. C, only if Oracle approves the acceleration in writing. Or D, yes, and empty installs with no workload behind them would work equally well. Pause here. What does the count date actually test?
The answer is B, and it's worth being completely clear eyed about why. The company paid a fixed, seven figure fee for unlimited deployment during the term. Deploying genuine, funded, roadmapped workloads before the count date is that right, being used, at full value. Run the arithmetic: those thirty processors deployed in month thirty four become certified entitlement at zero marginal cost, support unchanged. The same thirty processors deployed in month thirty eight get bought at list, session nine benchmarks, call it a million dollars net. Same workloads, same quarter of business value, a seven figure difference, decided by a scheduling meeting. That's not a loophole; that's the contract's economics working exactly as written, and Oracle prices ULAs knowing buyers can do this. A overcorrects into leaving paid for value on the table, and timidity at exit is just donation with better manners. C invents an approval right that doesn't exist, deployment during the term needs nobody's permission, that is the entire meaning of unlimited. Now D, because the line matters: installations with no workload behind them, empty instances spun up the week before the count to inflate the declaration, that's the version that gets contested. The counting language says installed and running, and a certification padded with theater invites exactly the scrutiny that stalls exits and births disputes. Accelerate the real. Skip the theater. The distinction is genuine use, and it's visible in the evidence.
Oracle's exit playbook, five moves, and knowing them in advance drains most of their power. Move one, the T minus six renewal call. Warm, early, helpful: why go through all that counting, when another term makes it unnecessary? It's session eleven's machine, gear two, opening politely, and it's timed for exactly the moment a disorganized estate starts feeling exit anxiety. Move two, the careful review warning: certification recast as an audit like risk, declare wrong and face consequences. You know from session eleven that this works on fog and bounces off records, and you know which estate you are. Move three, the helpful scripts: an offer to run Oracle's collection tooling across your estate, to assist with accuracy. What the clause requires is what governs, and assistance that expands scope beyond it gets declined, with thanks, next knowledge check. Move four, the finding lever: off list or out of definition issues, raised now, at maximum leverage, priced against a renewal that would make them all disappear. Sessions twelve and thirteen exist precisely so this lever has nothing to grip, the gate held, the events were filed, the findings don't exist. And move five, the slow walk: questions, requests, silence, while the term clock runs and nerves fray. The timeline language from entry, and your own on schedule declaration, are the complete answer, deliver on time, per the clause, and the clock becomes theirs to worry about. Let's practice the scripts decline.
Knowledge check two. During the exit, Oracle offers to run its collection scripts across your estate, to help certify accurately. Your clause, negotiated per session twelve, accepts your own verified inventory as evidence. Are you obliged to accept the help? A, yes, refusing the scripts looks like hiding something. B, yes, Oracle always verifies certifications with its own tooling. C, no: the clause defines the evidence, your verified inventory satisfies it, and the offer is a request you may decline politely. Or D, no, and the offer itself breaches the agreement. Pause here. It's session eight's audit lesson, at the exit.
The answer is C, and this is session eight's script lesson replayed at the exact moment it pays for itself. Obligations live in clauses. Your certification clause, negotiated at entry when it cost a sentence, names your verified inventory as acceptable evidence, so your verified inventory is what the certification stands on, and the scripts offer is a request, to be answered on its merits. Consider those merits: broad collection tooling, run across the entire estate, at the precise moment of maximum leverage, produces output that leaves your control and sees everything it sees, including things that have nothing to do with certifying listed products. There is no version of that trade that improves your position. The professional decline is one sentence, and it's worth writing down: our certification follows the process defined in the agreement, and the supporting evidence accompanies the declaration. Done. A negotiates against yourself with atmosphere, the same instinct that hands over unvetted audit data in week one, declining to exceed the contract is not concealment, it is the contract. B asserts a verification right that exists only if your clause grants it, most don't, read yours. And D overshoots in the opposite direction: offering isn't breaching, and treating every ask as an attack burns goodwill the confirmation process may still need. Calm, contractual, cooperative, in that order, at the exit as everywhere else in this course. Now, the letter itself.
Writing the declaration, five elements. First, the counts, per product, in the contract's metric: each listed program with its total, the register reconciled against the scripted inventory, session two's counting rules applied to the final numbers. This is the letter's entire substance, everything else is scaffolding around these figures. Second, the evidence file behind it, and note the phrasing: behind it, not attached to it. The declaration states the counts; the evidence, register extracts, scan output, cloud tags, feature usage reports, sits assembled, per product, ready if the process asks. Evidence that visibly exists rarely gets summoned; evidence that doesn't exist gets assumed against you. Third, the scope statement: the entities and environments covered, matching the customer definition, cloud included per the negotiated language, so the letter itself demonstrates that the count respects the boundaries. Fourth, the signature the clause names. Certifications typically require an officer of the company, so brief that executive properly, thirty minutes, before the letter reaches their desk: what the number is, how it was built, why it's defensible. Officers sign confidently when they understand what they're signing, and confident signatures read differently. And fifth, delivery per the clause: the notice method, the address, the window, followed to the letter. Procedural perfection costs an hour and removes every cheap objection before it can be raised. Then comes the day after.
The day after certification, five housekeeping moves that lock the result in. First, the perpetual position: the certified counts are now your entitlement, fixed quantities, perpetual licenses. The entitlement library gets its largest single update, session seven's species column marked perpetual, ULA vintage noted, because ten years from now someone will need to know where these licenses came from, and the answer is this file. Second, support, unchanged: the annuity continues at twenty two percent of the old fee, and the first post exit invoice gets checked against that expectation, line by line, because renewals have a way of drifting upward when nobody reconciles them. Third, and culturally hardest: the counting resumes. Unlimited is over. New deployments need licenses again, bought at session nine benchmarks, and module one's disciplines come back out of storage permanently. The estates that stumble here are the ones that let three unlimited years erase the habit. Fourth, the confirmation file: the letter, the delivery proof, the evidence bundle, and any acknowledgment from Oracle, archived together where the next decade can find them. Audits ten years out will ask what this position rests on, and this file is the answer. And fifth, the growth plan: future needs price à la carte at benchmark discounts now, and if growth genuinely reignites at scale, session fifteen's frameworks reopen the question deliberately. Which leaves only the failure modes, and they deserve their own slide.
Five mistakes that convert exits into renewals, and every one of them is voluntary. Mistake one, arriving with fog. No records, so the careful review warning lands with full force, and the renewal becomes the safe looking door. Note the timing of this mistake: it was made in years one and two, it's merely priced at exit. Mistake two, counting late. The count starts at T minus two, no dry runs behind it, and rushed counts systematically undercount, missed environments, forgotten entities, unrebuilt evidence. Undercounting at zero marginal cost is pure donation. Mistake three, the cloud surprise: discovering at exit what the certification clause says, or doesn't, about the half of the estate that migrated. Session eleven's classic, fully matured, and by exit the options are exactly two: accept the smaller count, or renew. The sentence at entry cost nothing; its absence at exit prices in seven figures. Mistake four, negotiating scared: treating findings, warnings, and deadlines as emergencies rather than positions. Every scared response teaches the other side what pressure works, and prices the next demand accordingly. The whole course has been an argument against this mistake. And mistake five, taking the extension, the friendliest trap in the set, which arrives at T minus one wearing a bow. It gets its own knowledge check, right now.
Knowledge check three. T minus one month. The records are solid, the count is done, the letter is drafted. Oracle offers a free six month extension of the ULA, to give you more time to count properly. Take it? A, yes, free time is free, and more care never hurts. B, yes, because refusing would look uncooperative right before the review. C, no: the count is ready, the extension mostly extends the pressure window, and certifying on schedule ends the game. Or D, no, and extensions are always traps in every circumstance. Pause here, and ask the question under the question: who benefits from six more months?
The answer is C. Run the who benefits test honestly. This company gains nothing: the count is done, the evidence is bundled, the letter is drafted. Six more months buys Oracle two more quarters of renewal conversation, two more chances for a corporate event or a cloud migration to complicate a finished count, occasionally an extended annuity clock depending on terms, and, subtlest of all, a precedent that the exit date is soft, that deadlines in this relationship are opening positions. An estate that is ready certifies on schedule, because the declaration is the one move that ends the game entirely: once delivered, there is nothing left to pressure, no renewal to sell against, no clock to run. The friendliest offers at exits are the ones that keep the game going. A prices free wrong; the cost is denominated in leverage, not dollars, and free leverage transfers are still transfers. B is the atmosphere instinct again, third time this module, and it retires the same way: contractual behavior on a contractual timeline needs no apology, ever. And D swings past the truth: an estate that genuinely isn't ready, mid merger, mid migration, records honestly incomplete, might rationally buy time. But that estate negotiates its extension deliberately, terms in writing, cost understood, which is session fifteen's decision discipline, not a reflex accepted on a phone call at T minus one. This estate is ready. Sign the letter. End the game. Let's see what it certified.
The certification letter, product by product, the module's whole story in one table. Database Enterprise Edition: two hundred ten processors certified, standing on the register, the quarterly scans, and the cloud tags counting under the clause negotiated at entry, that sentence, earning its keep at last. RAC: one hundred sixty eight processors, cluster configurations reconciled to the register with per node evidence, session two's counting rules visible in the workings. Partitioning: two hundred ten, matched to the EE estate with feature usage output, session three's tools doing exit duty. The Diagnostics and Tuning packs: two hundred ten each, pack usage reports and license views bundled per product. And the total position: perpetual, support unchanged at six hundred thirty eight thousand a year, the fee's annuity exactly as set at entry, with an effective cost per processor at roughly half of list. Now the note under the table, because it's the point: this estate entered the ULA forecasting ninety six processors and certified two hundred ten, the acquisition folded in under the inclusion clause, the T minus six acceleration deployed properly inside the term. The letter took an afternoon. The afternoon took three years of records. That's the module's arithmetic, and it only ever runs in that order.
Session fourteen in three sentences. One, certification converts three years of deployment into permanent entitlement at zero marginal cost, which makes the count date the most valuable afternoon in the agreement's life, and undercounting it the purest donation in software. Two, the exit is your process, run on your records, per the clause you negotiated at entry, and every move in Oracle's playbook, the renewal call, the review warning, the scripts, the findings, the slow walk, is an attempt to make it their process instead; the calm, contractual answer to each was written this session. Three, real deployments accelerated into the term are the paid for right being used, extensions offered at the finish line mostly extend the pressure, and an on schedule declaration is the one move that ends the game completely. Next session closes module three: the end of term decision itself, renew, certify, or restructure, priced as the deliberate choice it should be, plus the perpetual ULA examined properly, and how to run the last twelve months when the answer genuinely is another term. The decision framework, not the fear. See you in session fifteen.
Homework, about an hour, and this week it's exit engineering. One, build the timeline: if your estate has a ULA, lay the T minus twelve to T zero runway against the real term end date, milestones in the calendar, owners named per milestone. An exit with dates and names is a project; without them it's a hope. Two, test the counting rules: today's five category table against your own estate. Which rows are unambiguous, and which depend on clause language you should reread tonight? The depends rows are where exits are won and lost. Three, draft the letter skeleton: structure, products, evidence sources per product, counts left blank. A blank letter that exists removes half the exit fear, because the remaining work becomes visible and finite. Four, brief the signer: identify the executive whose signature the clause requires, and put thirty minutes in their calendar, before the pressure season, to walk through what certification is and how the number gets built. Officers who understand sign calmly, and calm signatures end games. And five, rehearse the declines: two sentences, written down, one declining the scripts, one declining the extension. Pressure spends prepared words far better than improvised ones. That's the hour. See you in session fifteen, where module three closes.
Five reads, all free on redress compliance dot com. First, the Oracle ULA negotiation playbook, the certification chapter of the module's standing reference, today's timeline in checklist form. Second, Oracle ULA renewal negotiation tactics, the exit pressure campaign and its counters in full depth, the playbook slide expanded to article length. Third, challenging Oracle audit findings, how contested counts actually get argued and resolved, useful preparation even for exits that never become contests. Fourth, conducting internal Oracle license audits, the counting methodology the whole declaration stands on, session thirteen's companion. And fifth, the case study, a Fortune 500 retailer's ULA, an exit run well, from the inside, with the numbers attached. That's session fourteen. A twelve month runway, five counting rules, five playbook moves with five calm answers, one letter with evidence standing silently behind it, and a game that ends the moment you deliver on schedule. One session left in module three: the decision. See you there.