Ramps, carryover, overage, and the true up, turned into contract language before you sign. Three knowledge checks along the way, and 1 clip from a senior cloud advisor.
This is a taught session, not a talking head. The instructor works through analyst grade slides, and three times the video stops on a question with four options on screen. Pause, commit to an answer, and the next slide explains which option is right and why each of the others is wrong. Once in the session the frame splits and a senior cloud advisor gives the view from inside real Oracle negotiations, and the instructor picks the clip apart when the slides return.
The full narration of this session, section by section, for reading and reference. Guest analyst clips are marked.
Welcome back, session seven of thirty. Last time you learned the meter: the models, the units, the bill, and the number I asked you to compute for homework, your distance to commit. Today that number goes to work, because this is the session where sizing becomes negotiating, and where every abstract warning about forfeits from sessions one, three, and six turns into specific contract language you can put in front of a deal desk. Here's the sentence the whole session hangs on: the forfeit is designed in the sizing meeting, not discovered at the renewal. By the time credits expire unconsumed, the mistake is a year old. So today we do the sizing meeting properly: the method, the structures, the clauses, the true up, and, because the world is imperfect, the rescue options when the commitment is already wrong. Let's start with your number.
Five takeaways. One, the method: a four step sizing calculation, baseline, committed growth, haircut, structure, that starts from the meter and ends in a number you can defend in front of a CFO, a steering committee, or an Oracle rep, which are three different kinds of hostile audience. Two, the structures: flat, ramped, floor plus overage, and the multicloud pool, and how to match the shape to the estate. Three, the clauses: carryover, ramp language, overage rates, and tier locks as actual order text, because verbal assurances have the shelf life of the rep who gave them. Four, the true up: growing a commitment mid term as a fresh negotiation instead of a bigger number on the same bad paper. And five, the rescue: the honest menu when the commit is already wrong, in either direction, and the one popular rescue that makes everything worse.
Your distance to commit lands you in one of four rooms, and each room has a different conversation in it. Room one, on track: trajectory within about ten percent of the commitment. Congratulations, your renewal conversation is about rate and protections, not size, and that's the strongest position in this module, bank it. Room two, forfeit ahead: trajectory materially under the commit, every unconsumed dollar donated at period end. The rescue slide later is written for you, and the most important thing you can do is start early, because rescue options decay with the calendar. Room three, overage ahead: trajectory exceeding the commit. That's good news wearing a price problem, the business is adopting faster than procurement guessed, and whether that growth bills at your discount or drifts toward list depends entirely on overage terms we'll cover today. And room four, the honest one: no number. Nobody can actually produce the trajectory. If that's your estate, the rule is simple and non negotiable: no commitment gets signed or renewed until someone can, because session six's discipline isn't optional homework, it's the prerequisite for every dollar decision in this module. The motto, now operational: commit to what you measure, structure for what you hope.
The sizing method, four steps, and the discipline is that they happen in this order. Step one, baseline: trailing six to twelve months of actual consumption, monthly, by service family. Not the plan, not the sizing sheet from the migration project, the meter. If the meter and the plan disagree, the meter is telling the truth. Step two, committed growth only: add workloads that are funded, scheduled, and owned, and count them from engineering's dates plus realistic slippage, never from the steering committee's dates, which are aspirations wearing a calendar. The test is brutal and useful: a migration without a named owner and a spent budget is hope, and hope gets structured, not committed. Step three, haircut the total: take ten to fifteen percent off the sum, and here's the asymmetry that justifies it. If you undersize, you pay overage on the gap, and if you negotiated overage properly, that's your committed rate or close to it. If you oversize, you forfeit the gap at one hundred cents on the dollar. Overage on a well negotiated deal is the cheaper mistake, every time, so the haircut leans you deliberately toward it. And step four, structure the rest: the distance between your haircut number and the optimistic case gets handled by ramp steps, carryover, and overage terms. Structure is how you stay honest about hope without paying for it in advance.
First check, the method with real numbers. Measured run rate: seventy thousand a month, flat. One funded migration, named owner, spent budget, adds an estimated forty thousand a month starting Q3. The sizing sheet on the table proposes committing one point six million for the year. The method says: A, one point six is right, it matches the optimistic full year case. B, about one million: the eight forty baseline plus the migration's half year at engineering dates, haircut applied, with the upside handled by ramp and overage terms. C, eight hundred forty thousand, never commit to anything not yet consumed. Or D, one point three million as a compromise between the camps. Pause here. Baseline, committed growth at real dates, haircut, structure.
The answer is B, and let's run it. Seventy a month is eight hundred forty a year of measured baseline. The migration is genuinely committed, funded, owned, dated, so it counts, but from Q3, which is half a year: forty a month for six months is two hundred forty. Total, one point zero eight million. Haircut ten percent, land at roughly a million. Now the wrong answers, because each is a real meeting. A, one point six, prices the migration as if it consumes from January the first, and no funded migration in recorded history has consumed from January the first. That sheet is a forfeit with a cover page. C, the pure caution answer, refuses to count growth that meets every test we set, funded, scheduled, owned, and would pay overage all year on workloads everyone knew were coming; caution has a price too. And D, the compromise, is the worst of them, because one point three is sized by the politics of the meeting rather than any arithmetic, and numbers born that way can't be defended later, to anyone. If the migration beats its dates, wonderful: overage at your negotiated rate catches it, or a mid year true up does, and both of those are exactly what structure is for. The commit is for what you'd bet on. Everything else is terms.
The four structures, and matching the shape to the estate. Flat annual: the same commitment every period, the simplest paper. Right when the estate is steady, trajectory within ten percent, and nothing big is planned. Ramped: periods that step up with the migration, six hundred, nine hundred, one point two million, and, the refinement you now know to demand, the discount tier priced at the full run rate from day one. Right for growing estates with funded, dated moves. Floor plus overage: commit only the reliable floor, and let peaks bill as overage at rates you negotiated in advance. Right for spiky, seasonal, or genuinely uncertain estates, and criminally underused, because reps don't propose it, buyers have to ask. And fourth, the multicloud pool, the newest shape: one credit pool spendable across OCI and the Database at services inside AWS, Azure, and Google. Right for estates running Oracle in multiple clouds, and module three will make that case fully. Here's what I want you to notice across all four: the shape changes, the paper doesn't. Every structure ends in the same three clauses, carryover, overage rate, tier protection, and the shape you pick just determines how hard each clause works. Let's hear how this plays out when a real migration meets a real calendar.
Guest analyst The best commitment I ever negotiated was for a manufacturer moving their Oracle estate to OCI over two years, and what made it good was not the discount. We sized their ramp off the engineering plan, then we assumed the plan would slip, because every plan slips, and we negotiated three things for that specific scenario. First, carryover: unused credits rolled one period forward, capped at twenty percent of the period. Second, the ramp steps could shift right once, by up to two quarters, on written notice, no penalty. And third, overage at the committed rate, so if the migration somehow ran early, that cost nothing extra either. Then reality arrived. A factory acquisition ate the integration team, and the migration slipped seven months, which on a normal contract would have forfeited about four hundred thousand dollars across two periods. Instead the carryover absorbed the first period, the ramp shift absorbed the second, and the total cost of a seven month slip was zero. Here is the part I want buyers to hear: none of those three clauses was expensive. Oracle agreed to all of them in one round, at signature, because at signature they wanted the deal. The identical asks, made mid slip, would have been a restructure negotiation with a churn threat behind it. Insurance is cheap before the storm. That is the whole lesson.
A seven month slip that cost zero, because three cheap clauses were bought at signature, when Oracle wanted the deal. Carryover, a ramp shift right, and overage at the committed rate. Write those three down; they're the difference between his story and the four hundred thousand dollar version of it.
Check two, the ramp pricing trap. A ramped deal steps six hundred K, nine hundred K, one point two million across three years. Sales prices the discount tier per period: shallow in year one because the commitment is small, deeper each year as it grows. The counter is: A, accept, smaller commitments naturally earn smaller discounts. B, price the tier at the full ramp run rate from day one, because the three year relationship is two point seven million and year one's rate should reflect the deal Oracle is actually getting. C, flatten the ramp to nine hundred K a year to average the tiers. Or D, drop the ramp entirely and commit one point two million from year one for the best tier. Pause here. What is Oracle's actual three year revenue in this deal?
The answer is B. Oracle isn't signing a six hundred thousand dollar deal, they're signing a two point seven million dollar relationship with a payment schedule, and everyone in the room knows it, including the deal desk that will approve the tier. Pricing the tier at the full run rate from day one is standard for ramped deals and granted routinely, to buyers who ask, which makes per period tiering exactly what it looks like: an opening position that costs money only when nobody pushes back. A accepts that opening position as physics. C, flattening the ramp, fixes a pricing problem with a structural sacrifice, reintroducing year one forfeit risk that the ramp existed to remove, when one sentence of tier language fixes it for free. And D is our old friend from session six, the full commitment from day one, donating the year one gap to buy a tier that the ask in answer B would have gotten anyway. The pattern across this module, and honestly across this course: when a money problem and a structure problem look tangled, check whether a sentence of language untangles them before you sacrifice the structure. It usually does.
Overage and the true up, the growth mechanics, five items, all negotiated before you need them. One, the overage rate: the default drifts toward rate card, and the ask, made at signature, is that consumption beyond the commit bills at your committed discount. The argument is one sentence: success should not be the scenario the contract punishes. Two, the cadence: know when overage actually invoices, monthly as you exceed, or trued at period end, because surprise timing breaks budgets even when the rate is right, and your finance team will thank you for the one email it takes to find out. Three, the true up trigger: when trajectory structurally exceeds the commit, upsizing captures a better tier, and you want that conversation to happen on your data, not the rep's forecast. So define your own trigger, three consecutive months above the commit's run rate is a good default, and put it in the operating rhythm from session six. Four, the true up as a negotiation, and this one's mandatory: a bigger commitment is a new deal. Fresh tier on the new total, a fresh look at the term, and a fresh pass at the entire special terms list, carryover, locks, notice, all of it. Never, ever sign an upsize that only changes the number, because you're handing over the one moment of leverage the term gives you for nothing. And five, the paper trail: every item above lives as order language. A rep's assurance about how overage is usually handled is a lovely sentiment with the shelf life of that rep's tenure on your account.
Now the session for the imperfect world: the commitment is already wrong, and the clauses that would have saved you aren't in the order. The honest menu, three items. If forfeit is ahead, option one: consume deliberately. Pull forward work that was genuinely planned: the DR environment scheduled for next year, the performance testing platform, the data warehouse build, training sandboxes. Consuming credits on real future value beats forfeiting them by exactly the value created. The boundary matters though: this is bounded by real projects with real owners. Spending for spending's sake is just the forfeit with extra steps, and we'll test that in the check. Option two, still forfeit ahead: negotiate the restructure. Oracle sometimes converts a failing commitment into a longer smaller one, or bridges it with migration funding, and the reason is cold arithmetic on their side: a customer who forfeits, resents it, and churns at renewal is worth less than one who restructures and stays. The ask costs nothing, and your leverage is the renewal sitting on the horizon, which is why timing is everything, six months early this is a negotiation, at renewal week it's a foregone conclusion. And if overage is ahead, option three: bring the true up forward, on your trigger, on your data. Running hot is leverage, you're the proof the platform works, so trade the upsize Oracle wants for the tier and terms you want. All three options share a final step: document why the commitment was wrong. That document is the evidence that prices the next one.
Last check. Six months into a one point two million annual commitment, trajectory says seven hundred thousand. Someone proposes spinning up idle compute instances to burn credits before period end, consumed beats forfeited, they say. The disciplined call: A, do it, consumed does beat forfeited. B, pull forward real planned work instead, and open the restructure conversation with Oracle now, six months before renewal, while the churn risk is their problem too. C, do nothing, take the forfeit quietly, size better next year. Or D, stop all new OCI adoption in protest of the bad sizing. Pause here. Idle compute creates what, exactly?
The answer is B, and let's be precise about why A fails, because consumed beats forfeited sounds so reasonable. Idle compute converts a forfeit into a bill for nothing, same money gone, plus two liabilities the forfeit didn't have: an inflated consumption baseline that will be Exhibit A in the rep's case for a bigger commitment next year, and a paper trail showing your organization burns credits on nothing, which is not the negotiating reputation you want. It's the forfeit with a paper trail against you. The real move has two parts running in parallel: pull forward genuinely planned work, the DR build, the test platform, things with owners and future value, and open the restructure conversation now, at month six, when Oracle still has a renewal to protect and your churn risk is leverage rather than history. C, the quiet forfeit, donates the money and skips the lesson, the worst of both. And D, freezing adoption in protest, punishes the business for procurement's sizing error, which is exactly backwards, the platform didn't fail, the spreadsheet did. Fix the spreadsheet, document the miss, and let the documentation price next year's commit.
Everything today, compressed into the contract language checklist, seven lines that go from this session into your next order. One, commitment and periods: the haircut number, with period boundaries you chose deliberately, because when the periods end determines when forfeits can happen. Two, the ramp schedule: steps matched to engineering dates, written into the order, tier priced at the full run rate. Three, carryover: unused credits roll forward, even partially, even capped at twenty percent like the advisor's deal, any carryover beats none, and the ask is routine. Four, overage rate and cadence: beyond commit bills at the committed discount, invoiced on a schedule finance has seen. Five, rate locks: unit rates for your top services locked for the term, because a discount percentage is hollow if the rate card moves underneath it. Six, true up terms: upsizing recalculates the tier on the whole relationship and reopens the special terms, written down now, while it's abstract and easy. And seven, the session two set: notice windows, non renewal by simple notice, co termination with the rest of your estate, because credit orders are orders, and everything module one taught applies to them. Seven lines. They fit on the same one page the advisor keeps bringing to signings, and now you know why he always has it with him.
Session seven, three sentences. One: size from the meter, baseline, committed growth at engineering dates, a ten to fifteen percent haircut, and structure for everything hopeful, because overage on a good deal is cheaper than forfeit on any deal. Two: pick the structure that matches the estate, flat, ramped, floor plus overage, or the multicloud pool, price ramped tiers at the full run rate from day one, and remember the shape changes but the three clauses don't. Three: overage rates, carryover, and true up terms are order language bought cheaply at signature, and a commitment that's already wrong gets rescued with real work pulled forward and a restructure opened early, never with idle compute. Next session, the lever that changes OCI's price more than any discount: BYOL. Bring your own license, the conversion ratios, when license included actually wins, and how to keep entitlements compliant once they're running in the cloud, because the toggle from session six was just the surface. See you there.
Homework, about an hour, and this week you produce the document that runs your next commitment negotiation. One, run the method: baseline, committed growth, haircut, for your estate's next period, with the working attached, the way session three taught you to keep workings. Two, name the structure: flat, ramped, floor plus overage, or pool, which one fits your estate, in two sentences you could say out loud in a meeting. Three, read your actual overage terms: the rate and the cadence, from the order, not from memory. If the honest answer is rate card and surprise, both go on the renewal ask list. Four, check for carryover: is there one clause anywhere in your current order that lets a single credit survive period end? If not, that's the cheapest line on the ask list. And five, define your true up trigger: write down the consumption threshold that opens the upsize conversation on your data, three consecutive months above run rate if you want the default. An hour of work, and you walk into the next commitment conversation as the best prepared person in the room, on either side of the table.
Five reads before next session, all free on redress compliance dot com. First, the OCI licensing and cost guide, the credit machinery that every clause in today's checklist attaches to. Second, Oracle cloud negotiations, the wider negotiation frame around the commitment conversation. Third, Multicloud Universal Credits, the fourth structure in full detail, one pool across OCI and the Database at services. Fourth, MUC versus Universal Credits, the decision between the classic commit and the multicloud pool, which more estates face every quarter. And fifth, Oracle cloud contracts and credits for CIOs, the executive briefing version of today's checklist, useful for getting your leadership aligned before the negotiation starts. That's session seven. You can size a commitment, structure it, and paper it. Next time, BYOL, the biggest single lever on the OCI price. See you there.