Why counts go up and rarely come down, and how to reverse it. Three knowledge checks along the way, and 4 clips from a senior licensing analyst.
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. 4 times in the session the frame splits and a senior licensing analyst gives the view from inside real SAP 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 four, shelfware, true ups and the ratchet. Every organisation I have worked with has shelfware, including the ones with strong procurement functions and careful administrators. That tells you it is not a diligence problem, it is a mechanism, and once you can describe the mechanism you can work against it. So today: why the asymmetry exists, the four routes by which counts rise without anybody deciding to grow, measuring shelfware in a way that survives challenge, building a reduction case that is a business position rather than a licensing opinion, and the timing that decides whether any of it counts. Three knowledge checks. Let's begin.
Five objectives. First, explain the ratchet: counts rise easily and fall only through a contractual right you probably do not have. Second, measure shelfware honestly, because assigned, active and meaningfully used are three different numbers. Third, build a reduction case that is named, evidenced and agreed with the business before it ever reaches the vendor. Fourth, time the ask, since a reduction has one window a term and it opens long before the renewal meeting. And fifth, win the right in advance, because a reduction band negotiated at signature is worth more than any argument you can make later.
So, the ratchet, and the asymmetry is structural rather than accidental. Up is easy: a request, an approval, sometimes not even that, and it takes minutes. Down is a right, so if the agreement is silent then reduction at renewal is a request rather than an entitlement. Adds set floors, because a mid term true up becomes the new minimum you renew from. And nobody owns it, since there is no team whose job is to find the licences that are no longer needed. Let me put that in the plainest terms I can.
Guest analyst clip.
So shelfware is not a failure of diligence. It is the predictable output of a system where one direction is frictionless and the other requires a project, and describing it that way matters, because it changes how the conversation goes internally when you start pulling on it.
Right, four routes by which counts rise. The project spike: a programme licenses a wide population for a rollout, the programme ends, and the licences do not. The leaver gap: people leave and accounts stay assigned, with no joiner to trigger a review so nothing prompts removal. The safety margin: extra licences bought to avoid running out, consumed by growth, and then treated as the new baseline. And the mid term true up: a settlement for over-deployment during the term, which becomes the committed minimum at the next renewal. Notice that none of those is a decision to increase spend. Each one is a sensible local action taken by somebody doing their job, and the aggregate is a licence count nobody chose.
First knowledge check. Why do Salesforce licence counts rarely fall? A, because organisations keep growing. B, because adding is frictionless while reducing needs a right, a measurement and a negotiation. C, because unused licences cost nothing. D, because Salesforce prohibits reductions. Pause here and pick an answer before you continue.
B. The asymmetry is in the process rather than in anybody's intentions, which is exactly why the drift happens even in well run organisations with careful procurement functions. A is true for some, and it does not explain the estates that shrank while their licence count rose, and I have seen several of those. C is the assumption that lets it continue, and an unused licence costs precisely the same as a used one. And D is not true: reductions happen regularly, and they happen for buyers who prepared for them rather than for buyers who asked for them.
So, measuring it honestly. Three numbers, and only the third is the truth. Assigned: what you pay for, the easy number, and the one the invoice agrees with. Active: signed in during the period, already smaller, and easily defended. Meaningfully used: did work in the system, so records touched rather than just a login, and that is the honest figure. Set the period deliberately, where ninety days is standard, and some roles genuinely work quarterly so know your exceptions before somebody points them out to you. And expect to be challenged, which is why the definition of used should be agreed before you publish anything at all. Now, building the case.
Guest analyst clip.
Five steps, and four of them are internal. Name every licence: person, role, department, last meaningful use, because aggregates get argued with and names do not. Sort into three piles, remove, downgrade and keep, and most of the value is in the middle pile. Take it to each manager with their own list and their own data.
Then let them correct you, because some of your removals will be wrong, and finding that out early is exactly what makes the rest of the list credible. And then it is a business position rather than a licensing opinion, which is the whole difference between a reduction and a request.
Second knowledge check. When should you first raise a reduction with the vendor? A, as early as possible, to give them time to plan. B, after the internal case is built and agreed with the business. C, in the renewal meeting itself, for maximum impact. D, never, reductions should simply be applied at renewal. Pause here before you continue.
B. An unagreed reduction gets tested, and the way it gets tested is a conversation with the business unit that holds the licences, and if that conversation finds anybody who still wants them then your number collapses and it does not recover. A gives them months to build precisely that counter position. C wastes the preparation, because a number produced in the meeting invites a request for evidence and a follow up where the momentum is gone. And D describes an entitlement most agreements do not grant, and applying a reduction you have no right to apply is how a true up begins.
So, timing, working backwards. Twelve months out, start measuring, because you need a trend rather than a snapshot to survive challenge. Nine months out, build the named list and take it to managers, one department at a time. Six months out, the business position is agreed and written down, and now it is a company number rather than yours. Before the notice date, table it, because after the notice window closes everything becomes a favour. And not in the last month, because a reduction raised late reads as a negotiating tactic and it will be treated as one. Let me explain why that sequence is not negotiable.
Guest analyst clip.
After the notice window closes, everything becomes a favour. That is the sentence to remember, and it applies to every ask in this course, not only to reductions.
Now the rights that make it possible, and you ask for these at signature because later is genuinely too late. A reduction band: a stated percentage you may reduce at renewal without reopening the agreement. True ups that do not ratchet, so mid term additions priced without becoming the permanent floor. Swap rights, letting you convert between licence types as roles change rather than buying the new one and keeping the old. And a shorter term on volatile lines, because where the population is genuinely uncertain you should not lock it for three years. None of those is exotic. All of them are refused if unasked and frequently granted if asked early enough to be part of the shape of the deal.
Five failures. The aggregate number: we think about two hundred are unused, which is an opinion and will be treated as one. Announced before agreed, so the business hears about it from the account team and then defends every single licence. Login treated as usage, because a weekly login to a dashboard is not a working user, and the reverse is also true and catches people out. Raised after the notice date, meaning correct analysis delivered when there is no longer a mechanism to act on it. And no reduction right, where the whole case is sound and the agreement gives you no way to apply it.
Last knowledge check. Which single contract term does most to prevent shelfware accumulating? A, a capped uplift. B, a stated reduction band at each renewal. C, a longer term for a better rate. D, co-termination of all lines. Pause here and pick an answer before you continue.
B. The band converts reduction from a negotiation you might lose into a right you can exercise, and it changes behaviour internally as well, because a reduction that will definitely be applied is worth the effort of measuring for. A protects the rate on whatever quantity you hold, which is valuable and does nothing about the quantity itself. C usually makes the ratchet worse, by extending the period during which nothing can be corrected. And D concentrates leverage, which helps you use a reduction right, and is not one by itself. Let me set out the rhythm.
Guest analyst clip.
A rhythm, so this is never a project again. Quarterly, the three numbers: assigned, active, meaningfully used, per licence type, plotted over time so you have a trend when you need one. Quarterly, the leaver check against the identity system, and this category needs no negotiation at all, it is simply yours. Twice a year, the named list reviewed with managers, so the position is always roughly current rather than being reconstructed under pressure. Annually, the rights check, meaning what your agreement actually permits, because that decides what all this work is worth. And at every renewal, table it, even when the number is small, because a reduction never asked for is never granted.
Three sentences. Licence counts rise because adding is frictionless and reducing requires a contractual right, a defensible measurement and a business agreement, so shelfware is the predictable output of the process rather than a failure of anybody's diligence. A reduction case is built internally: named licences with last meaningful use, sorted into remove, downgrade and keep, walked through with each manager and corrected by them, so what reaches the vendor is a company position. And timing decides whether any of it counts, because a reduction tabled after the notice window is a favour, and the term that matters most is a reduction band negotiated at signature.
Homework before session twenty five, about two hours. One, pull assigned against active by licence type for the last ninety days, and write down the gap per type. Two, define meaningfully used in one sentence before you measure it, then measure it and compare all three numbers. Three, run the leaver check against the identity system, because this is the free reduction and it needs no case at all. Four, find your reduction right by reading the agreement, and if there is no band then that is your most valuable renewal ask. And five, draft one department's list, named, with last use, and do not send it, just see how it feels to hold real evidence rather than an estimate.
Five guides, all on redresscompliance dot com. Salesforce shelfware covers finding wasted seats and what to do once you have found them. Salesforce minimums and true ups explains how the floor works and how a true up becomes permanent. Salesforce licence count audit covers measuring the estate properly, in a way that survives challenge. Salesforce licence optimization walks the reduction process end to end including the internal steps. And the Salesforce renewal negotiation playbook shows where the reduction sits inside the renewal and when to table it.
That is session twenty four. The thing to take away is that the ratchet is a mechanism rather than bad luck, so the counter is also mechanical: measure quarterly, name every licence, agree it internally, and win the reduction right at signature rather than arguing for it at renewal. Next time we close module five with licensing events: acquisitions, divestitures, org consolidation and org splits, where corporate activity turns into licensing exposure. See you then.