Usage analytics, the gap ledger, expansion resistance, and building the renewal file all year. 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 twenty three of thirty. Last week ended with a question: does the renewal file exist for your biggest subscription? Today is the session that builds it, because the file is not a document you write, it is a residue of how you run the years between signatures. Those years have a structural imbalance worth naming at the start: the account team works your estate all four quarters, adoption reviews, pilots, briefings, quarter end offers, while most customers engage commercially for six weeks before each renewal. Forty six weeks of selling against six weeks of buying, every year. This session is about closing that gap without hiring anyone: the six sources of shelfware, the quarterly usage review, the gap ledger that turns waste into leverage, the art of saying no to mid term expansion professionally, and the deprovisioning ritual, the one licensing lever that needs no negotiation at all. An hour a quarter, and the renewal desk never surprises you again.
Five takeaways. One, the sources: the six ways shelfware gets made, and the slightly deflating truth that most of it is manufactured at signature and merely discovered later, which changes what mid term management is actually for. Two, the rhythm: the quarterly usage review, one hour, four steps, the only evidence machine renewals respect. Three, the ledger: one row per line item, billed against active with cause, action, and date, and the principle that its value is proportional to its age, twelve dated rows outargue any last minute assertion. Four, the resistance: the expansion plays that run all year, and the counters, one channel, a business case gate, and yeses timed to your calendar instead of their urgency. And five, the hygiene: deprovisioning as a licensing control, leavers, movers, service accounts, the meter you directly operate. The session's single rule, which every discipline today serves: never let silence be the signature.
Where shelfware comes from, six sources, and notice as we go how many were covered by earlier sessions, because that is the point. Source one, oversized initial deals: partner workbooks, growth optimism, bundle padding, session eighteen's whole subject, fixed upstream by sizing from identity data and holding the rest. Two, floors and minimums: module floors billing above real populations, session sixteen, challenged at signature or paid forever. Three, failed or partial rollouts: the module deployed to half its plan, or the transformation that got cancelled with its subscription still billing, Tom's pharma client found exactly this, fixed by milestone ramps and swap rights. Four, departed users: leavers who keep seats for years, this session's hygiene ritual. Five, workforce shrinkage: divestitures and reductions that the metric serenely ignores, session nineteen's down lane. And six, M&A overlap: two contracts covering one merged workforce, module six's territory. The pattern across all six: most shelfware is manufactured at signature and merely discovered mid term. Which reframes this session's job precisely: not prevention, the earlier sessions did prevention. Detection, documentation, and conversion, turning the waste you inherited into the leverage you spend. The machine for that is the quarterly review.
The quarterly usage review, one hour, four steps, deliberately boring. Step one, extract: active users per module from the application's own analytics and your identity system, billed quantities from the orders, the same query every quarter so the trend line means something. Automate it if you can; a recurring calendar entry and a spreadsheet are entirely sufficient, the discipline matters more than the tooling. Step two, compare and ratio: billed over active, per line item, and flag everything above one point two, the same threshold we set in session sixteen. Step three, cause: every flagged gap gets a diagnosis, leavers not deprovisioned, rollout stalled, module abandoned, floor artifact, because the cause routes the fix, deprovisioning is an operations ticket, a stalled rollout is an escalation, an abandoned module is a renewal swap candidate. And step four, log and act: one dated row per line in the gap ledger, plus the quarter's actions with owners. Fifteen minutes of writing. Why so little process for something this valuable? Because the renewal desk's entire structural advantage is that most customers cannot produce twelve months of usage evidence on demand, and the desk knows it, prices it, and depends on it. This ritual is how you always can. An hour a quarter buys the only argument the machine cannot dismiss. First check makes the stakes concrete.
First check, and it is the conceptual heart of the session. Beyond its own subscription fees, what does undocumented shelfware actually cost at the renewal table? A, nothing extra, unused seats are just wasted fees. B, it becomes the seller's evidence: paid but unused quantity reads as demonstrated budget and renews as the base, the uplift applies to it, and any late right sizing claim without documentation reads as posture. C, it converts to service credits if you ask. Or D, Oracle proactively flags it and offers a reduction. Pause here. At renewal, who is holding the evidence about your unused seats, and what story does it tell for them?
The answer is B, and it is the insight this whole session stands on: shelfware has two costs, and the invisible one is evidentiary. The fees you can see. The evidence works like this: a quantity you paid for without complaint, year after year, is the strongest possible signal of what your budget tolerates, and the renewal machine reads it exactly that way, the full billed quantity is the base, the uplift applies to every seat including the empty ones, and last session's flat if you add play anchors against the whole number. Meanwhile, a right sizing claim raised for the first time at T minus sixty, with no paper trail, reads as negotiating posture, and here is the uncomfortable part, that reading is fair, because an undocumented last minute claim is exactly what posture looks like. Now watch the gap ledger invert it: the same unused seats, documented quarterly, with causes diagnosed and actions taken, become your case, twelve dated rows proving the gap was real, persistent, and managed. Same shelfware, opposite weapon, and the only difference is who wrote it down first. C is a myth, subscriptions do not convert to credits, we buried this in session sixteen and it keeps climbing out. D misreads the vendor's own incentives: customer success and adoption teams exist to drive usage up ahead of renewal, because usage defends the base. Nobody on that side of the table is paid to shrink your bill. The ledger exists because that job is entirely, structurally, yours.
The gap ledger itself, five columns, one row per line item, every quarter. Billed and active: the order quantity against measured monthly actives, the gap stated in the product's own numbers, which is what makes it undismissable. Ratio and trend: this quarter's ratio against last quarter's, because a widening gap flags itself now, while there is time to act, instead of at T minus sixty when there is only time to argue. Cause: leavers, stalled rollout, abandoned module, floor artifact, the diagnosis that routes the action and, later, defends the claim, a gap with a documented cause is a managed position, a gap without one is just untidiness. Action and owner: deprovision, escalate, mark for swap, or, legitimately, accept, some gaps are headroom you chose, the ledger records that choice too. And date and author: when measured, by whom, because the whole edifice rests on the timestamps. Which brings us to the ledger's core principle, worth stating as a law: its renewal value is exactly proportional to its age. Four dated quarters of a documented, managed gap is a right size case the desk cannot wave away. The identical numbers, assembled the week before renewal, are an opinion. Our guest analyst has seen both versions walk into the same room, and the comparison is the best argument for this session I know. Let's hear it.
Guest analyst I want to tell you about two clients with the same problem and opposite outcomes, because together they make the case for the most boring discipline I teach. Client one, a retail group, discovered at renewal minus sixty that roughly thirty percent of their Fusion seats had been dormant for over a year. Real waste, honestly measured, we pulled the data ourselves. We brought it to the renewal, and the desk's response was polite and devastating: you have paid for these seats without comment for three years, they are your growth capacity, and the renewal base stands. And with nothing but a fresh spreadsheet against three years of paid invoices, that argument held. The uplift applied to everything. Client two, an engineering firm, had been running a quarterly gap ledger for two years, not because they were sophisticated, because a junior analyst had set a calendar reminder after a course like this one. Eight dated quarters: billed, active, cause, action. Leavers deprovisioned monthly, one stalled module escalated twice and then marked for swap, every row timestamped. Same conversation, same vendor, same thirty percent gap. But this time the desk was not looking at a spreadsheet, it was looking at a record, and the record said: this customer measured the gap, managed it, warned about it, and will defend it. They right sized twenty eight percent at flat rates, and the meeting took forty minutes. Here is the lesson as plainly as I can put it: the seats were equally empty in both companies. What differed was who owned the story of the emptiness. Evidence is not what you know. It is what you wrote down, dated, before it mattered.
Same gap, same vendor, opposite outcomes: growth capacity for the client with a fresh spreadsheet, a twenty eight percent reduction at flat rates for the client with eight dated quarters. Evidence is what you wrote down, dated, before it mattered. Second check.
Check two, the conversion offer. Mid term, your ledger shows six hundred unused seats. The account team offers to convert them into a new AI module of equivalent value: no new spend, better utilization, everybody wins. The contract has no swap rights. The right response: A, accept, converting waste into something useful is a win. B, recognize it as a new purchase in disguise: the seats' value is already spent either way, the module adds a new obligation that renews forever, so it faces the business case gate and the session twenty one clauses like any purchase, and the six hundred seats stay in the ledger as renewal evidence. C, accept if the module lists at more than the seats' value. Or D, decline and terminate the six hundred seats immediately. Pause here. What happens to the seats' renewal leverage if they convert today? And what does the new module cost at the next renewal?
The answer is B, and the method is to trace what actually changes hands. Start with the seats: their fees are committed to end of term regardless, conversion recovers nothing, that money was spent at signature and no arrangement of SKUs unspends it. So what does the conversion actually do? Three things, all for the seller. It retires your best renewal argument, the documented six hundred seat gap, the exact asset Tom's engineering firm spent two years building. It installs a fresh module whose adoption clock starts at zero, next year's shelfware candidate. And, the clause nobody reads aloud, that module renews, forever, at whatever the machine asks, having entered the estate with no business case, no negotiation, no clauses, no ramp. That is not utilization, it is a purchase that skipped procurement by dressing as housekeeping. B routes it correctly: if the module is genuinely wanted, it faces the gate and gets the session twenty one treatment, and the natural venue is the renewal itself, where real swap rights can be exercised if you have them or written if you do not, with the six hundred seats as the leverage that pays for them. C prices the favor in list price, the one currency that means nothing. D proposes the impossible: quantities do not terminate mid term, that asymmetry is half this module's curriculum, and asking for it tells the account team the ledger is not understood on your side of the table. The general rule, and it covers every variant of this offer: anything that converts documented waste into fresh obligation is trading your evidence for their revenue.
Expansion resistance, five disciplines, and the framing matters: none of this is hostility, the account team selling all year is their job, done well. The problem is structural, forty six weeks of selling against six weeks of buying, and these disciplines rebalance it. One, know the plays: adoption reviews that end in module pitches, free pilots that auto convert, executive briefings engineered to produce commitments in the room, and quarter end urgency applied to things you never asked for. All legitimate, all designed to route spend around your gate. Two, one channel: every commercial conversation goes through the named owner from session twenty two, and, the hard half, your own people know it, because the architect who tells the account team over coffee that the demo was amazing has just priced the next quote, enthusiasm expressed to a seller is a term sheet concession you did not mean to make. Three, the gate: mid term spend passes a business case or waits for the renewal, and the renewal venue is not bureaucracy, it is where additions arrive with clauses, ramps, and caps instead of stub terms at list. Four, time the yes: real needs get bought deliberately, bundled into the renewal where the leverage lives, or at the vendor's quarter end with the clauses attached, never at their urgency, always on your calendar, session twenty one's endgame applied to small purchases. And five, log the pressure: every pitch, pilot, and urgency play goes in the ledger too, because a year of documented selling pressure is context that makes the renewal escalation meeting remarkably clarifying.
Deprovisioning discipline, the meter you actually control, three practices. Leavers, monthly: the joiner mover leaver process ends with license deprovisioning, run monthly against the HR feed, because a leaver's seat billing for a year is the purest shelfware in existence and the easiest to prevent, it requires no negotiation, no escalation, no vendor conversation, just an operations ticket that fires on schedule. Movers and roles, quarterly: the role change that no longer needs the module releases the seat, and here is the efficiency, your security team already runs access reviews for exactly this population, borrow their process, the access review is a licensing control wearing a compliance badge. And service accounts and tests: integration users, test accounts, training shells, they accumulate silently, they count like humans on most named user metrics, and nobody owns them, so inventory them quarterly and, per session eighteen, name the legitimate integration accounts on the order so their status is settled paper rather than renewal debate. One honest boundary on all this: Hosted Employee metrics do not respond to deprovisioning, the workforce counts what the definition says it counts, that is session nineteen's territory and the down lane clause. But every named user line in the estate is directly operable, today, by you. It is the rare licensing lever that requires no negotiation at all, which is exactly why it is the one most estates never pull. Last check.
Last check, the free pilot. The account team offers a six month pilot of a new AI module: no charge, genuinely useful capability, and the paper says it auto converts to a paid annual subscription at month six unless cancelled in writing. Legal confirms the terms are standard. The right handling: A, accept, free is free and the team wants to try it. B, accept only after rewriting the conversion: explicit opt in required to convert, no auto renewal on the resulting subscription, and the end date diaried with an owner; otherwise the pilot is a purchase with a delay, priced at your future forgetfulness. C, decline all pilots on principle. Or D, accept and rely on remembering to cancel at month five. Pause here. Who profits from the auto convert clause, and what behavior, exactly, is it pricing?
The answer is B, and start by reading the instrument honestly: the vendor carries six months of hosting cost in exchange for a signature that becomes an annual subscription through inaction. The auto convert clause is pricing a specific, measurable, extremely reliable behavior, organizational forgetfulness, the same behavior auto renewal prices, the same behavior every fuse in this session prices. Six months is two reorganizations and one budget cycle in most companies; the sponsor who accepted the pilot may not even hold the same role when the fuse burns down. So the fix is the same as it always is: conversion requires an explicit yes, the subscription that results carries no auto renewal of its own, and the end date gets a diary entry and a named owner anyway, belt and braces, because clauses protect you on paper and calendars protect you in practice. With those changes, note, the pilot becomes genuinely valuable: six months of real usage data on your workloads is precisely the input the business case gate wants, which is why C overcorrects, the pilot is not the trap, the default is. D accepts the trap on the honor system. A signs a purchase while pronouncing it free. And now step back, because the session closes here: auto renewal, auto conversion, quantities that true up but never down, undocumented gaps that become demonstrated budget, every mechanism we have studied today monetizes the same thing, your silence. The ledger, the diary, the owner, the channel, the gate, five tools, one rule: never let silence be the signature.
The all year renewal file, closing the loop with session twenty two. Last week I asked you to open the file; this session showed where its contents come from, and the answer is: from the rhythm, as residue. From the quarterly review: four to twelve dated ledger rows per line item, billed, active, ratio, cause, action, and the right size table assembles itself from those rows, with its age as its authority, Tom's eight quarters against the fresh spreadsheet. From the resistance log: the year's pitches, pilots, and pressure documented, plus every mid term addition that passed the gate with its clauses, and every one routed to the renewal instead, arriving there with your leverage intact. From the hygiene ritual: deprovisioning statistics showing the named user counts are already clean, which means the renewal conversation starts from real usage rather than from a padded number the desk hopes you have not checked. Notice what happened across these two sessions: the renewal stopped being an event and became a byproduct. The estate that runs the rhythm walks into T minus twelve with the file already written, and the six week scramble that most customers call a renewal becomes, for you, the boring execution of a position you built all year. One capability remains to complete the lifecycle's leverage story: the exit, the thing that prices every renewal even if you never use it. That is next session.
Session twenty three, three sentences. One: shelfware is manufactured at signature and discovered mid term, and its deepest cost is evidentiary, undocumented it is the seller's proof of your budget, documented quarterly it is your proof of the gap, same seats, opposite weapon, decided by who wrote it down first. Two: the gap ledger costs an hour a quarter, one dated row per line with billed, active, cause, and action, and its value is proportional to its age, eight quarters right sized twenty eight percent at flat rates while a fresh spreadsheet renewed everything with the uplift. Three: the account team sells all year, so spend passes the gate or waits for the renewal, pilots convert only by explicit opt in, every fuse gets a diary entry and an owner, and the whole session is one rule, never let silence be the signature. Next week: the exit, termination mechanics, data retrieval windows, and the credibility that prices every renewal whether or not you ever leave. See you there.
Homework, about an hour, and it is quarter one of the ritual, run for real. One, extract once: active users per module for your biggest SaaS contract, from the identity system or the application's own analytics, however rough, this is the baseline every later quarter compares against. Two, build the ledger: one row per line item, billed, active, ratio, cause, action, date, the template is today's table, an hour to build once, fifteen minutes a quarter to maintain. Three, deprovision the leavers: cross the active list against the HR leaver feed and release the seats, then count what you freed, because that number, seats times rate, is the ritual's business case, and it funds the habit forever, most estates find the first pass pays for years of the discipline. Four, diary the fuses: every pilot end date, auto renewal notice window, and auto convert clause across the estate, each with a named owner, silence spends money and calendars stop it. And five, route the channel: confirm who owns vendor commercial conversations, tell the account team, and, the part that actually matters, tell your own architects and project leads, one channel, starting now. An hour of work, and the forty six week imbalance starts closing this quarter.
Five reads before next session, all free on redress compliance dot com. First, the enterprise software shelfware report, the scale of this problem across the industry, useful for the executive who thinks your estate is the exception, it is not. Second, the software shelfware audit guide, the gap ledger method expanded into a full audit when you want to go deeper than the quarterly hour. Third, the Oracle software asset management guide, the SAM practice these rituals belong inside, because the ledger works best as part of a standing function rather than a heroic one off. Fourth, optimizing your Oracle license footprint before renewal, the bridge from this session's ledger to last session's right size table, the two halves of one method. And fifth, the FinOps guide for SaaS licensing and software spend, the operating model that makes the rhythm stick across the whole estate, and a preview of module six, where governance takes over from discipline. That's session twenty three. An hour a quarter, one dated row at a time, and the story of the emptiness belongs to you. Next week, the exit, and the leverage it prints even when you stay. See you there.