The twelve week programme that removes fulfiller and custom table exposure while it is still free to remove. Three knowledge checks along the way, and 3 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. 3 times in the session the frame splits and a senior licensing analyst gives the view from inside real ServiceNow 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 fifteen, and this closes module three. Now, I want to be honest with you about what today is, because there is almost nothing new in it. Sessions eleven through fourteen gave you the fulfiller line, the audit, custom tables, and the true up. Today is those same four things, done early. That is the entire content. And you might reasonably ask why that deserves a session of its own, so let me answer that directly. Because timing is the single largest variable in what any of this costs, larger than negotiating skill, larger than the discount you win, larger than the size of your estate. The identical finding, five hundred over licensed seats, costs you staff time on an ordinary Tuesday and full multi year rate at a signature date. Same seats, same evidence, same people. Only the calendar changed. So today is about the calendar, and about the twelve week programme that keeps you on the cheap end of it. Three checks, homework, let's go.
Five objectives. First, price the timing, so you can say what the same finding costs at each stage and use that curve to argue internally for the work, because you will have to argue for it. Second, run the twelve weeks, the week by week programme that closes fulfiller and table exposure before a renewal window even opens. Third, take the quick wins first, the four categories that convert into saving with no negotiation and no capability loss, taken in the order that builds credibility for the harder conversations. Fourth, fix the structure, the four changes that stop exposure regenerating so that this is the last time you ever run the full programme rather than the first of many. And fifth, build the evidence pack, the document set that turns a cleanup into a defensible position, assembled before anybody asks you for it, because a pack produced on request always looks assembled on request.
The same finding at four different prices. Free, a role removed on an ordinary Tuesday, no negotiation, no counterparty, no deadline, and this is genuinely the cheapest that finding will ever be. Twenty five to forty five percent, the share of claimed exposure that disappeared when cleanup ran before the count was taken, so still cheap, but notice what has changed, you are now on somebody else's clock. Thirty to fifty percent, how far settlements closed below the opening claim when the buyer brought independent data, so recoverable, but now you are paying for something rather than simply removing it. And full price, the same seats discovered at a signature date with a renewal attached and no time to clean anything, which is the expensive end and, in my experience, the most common one. Now sit with the note, because it is the whole argument. The finding does not change across those four columns. The seats are identical. Only the timing changes, and the timing is entirely inside your control right up until the moment it is not. Let's hear that from the field.
Guest analyst clip. I get called at every point on that curve, and the difference in what I can achieve is enormous, which is frustrating because the work is identical. When somebody calls me eighteen months before a renewal, we do genuinely creative things. We reshape the estate, we retire products, we plan a tier migration properly, we build a position over three quarters. When somebody calls me three weeks before a signature date, I am not a strategist any more, I am a negotiator working with whatever exists, and what exists is usually a drifted estate and a deadline. Same client, same problem, wildly different outcome. And here is the part that frustrates me most, because it is so avoidable. The three week client almost always knew. They had been meaning to look at their user counts for a year. It was on somebody's list. It kept getting pushed because there was no deadline, and that is precisely the trap, the work has no natural urgency until it has an expensive one. So the advice I give every client who will listen is to manufacture the deadline artificially. Put the licensing review in the calendar as a recurring commitment with the same status as a financial close. Not because it is urgent, but because the only alternative is doing it when it becomes urgent, and by then it costs a multiple.
The work has no natural urgency until it has an expensive one. Manufacture the deadline. That is the practical answer to why this keeps getting postponed in every organisation, including well run ones, and it is not a discipline problem, it is a structural one. Nothing in your calendar prompts you to look. So put something there. Now, the cost curve as an actual business case.
Here is the same five hundred seats, priced by when you find them. Any quiet quarter, it is an internal cleanup nobody outside the company even knows happened, and it costs staff time only, the seats simply stop counting. Twelve weeks before renewal, it is a rightsizing exercise feeding a renewal position, costing staff time plus the internal debate about contested cases, which is real but manageable. After a claim lands, it is a defence run under a deadline you did not set, costing staff time plus whatever the reconciled exposure settles at. And at the signature date it is a concession made to close a renewal, at full rate and usually multi year, because whatever you concede gets baked into the new term and then uplifts annually for its duration. Now the note, and this is why I made this slide a table rather than a paragraph. This is your business case. When somebody asks why licensing hygiene deserves a day a quarter of a skilled person's time, you do not argue about principles. You show them row one against row four for the same finding, and the conversation ends.
The programme, twelve weeks, four phases. Weeks one to three, count, the five exports from session twelve plus table counts against the allowance from session thirteen, and note the discipline, raw data only, no interpretation, no conversations with anybody yet. Weeks four to six, classify, the four column pass on users and every custom application tested against the two boundary rules, with reasons recorded and exceptions collected as you go. Weeks seven to nine, remove, machine accounts, dormant roles, approvers, dashboard viewers, retired application tables, uncontested categories first and in that order. And weeks ten to twelve, fix and paper, templates, group gates, custom role review, leaver hooks, and then the evidence pack, dated and signed, ready for whatever comes next. Now read the note, because it is where programmes fail. Phases one and two produce no savings at all and they are not optional. The temptation, especially when somebody senior is asking what you have saved, is to skip to removal. That is how organisations remove the wrong things and then spend three weeks defending the decision, which costs more than the phases you skipped.
Knowledge check one. Your renewal is fourteen months away and the platform team is busy, as platform teams always are. When should the programme start? A, now, because the work is cheapest with no deadline attached. B, at twelve weeks, which is what the programme is designed for. C, when the account team requests a review. Or D, after the renewal, so the next term starts clean. Pause here, and ask which column of the cost curve you are choosing.
The answer is A, now. And answer B is the interesting distractor, because twelve weeks is what I have been teaching, so it feels correct. But twelve weeks is the minimum viable runway, not the target. At fourteen months out you are sitting in column one of the cost curve, where the same finding costs staff time and absolutely nothing else, and there is no reason to wait to enter a more expensive column later. Answer C hands the timing to the counterparty, which is close to a definition of losing leverage, and it is what most organisations do by default rather than by choice. And answer D, wait until after the renewal so the next term starts clean, is the most expensive option in the list wearing the costume of tidiness, because you would renew at inflated volumes and then spend an entire term paying for seats you have already identified as unnecessary. That is paying for knowledge you already had.
The quick wins, four categories, and one property they share that makes them special. Integration and service accounts, machine credentials on named licences, no human objects, nothing breaks, saving is immediate, always start here. Dormant users, no login in ninety days, and build the list with names, departments, and dates so that managers are confirming rather than debating, which is a different meeting. Approvers on fulfiller licences, the six figure reclassification from session eleven, they keep approving and only the licence line changes so there is genuinely nothing to lose. And tables from retired applications, where the project ended but the tables still hold against your allowance, so decommissioning recovers capacity for free. Now the fifth line, which is the point of the slide. All four share one property, nobody loses a capability they actually exercise. That is what makes them uncontested, and it is exactly why they come before anything requiring persuasion. You are not softening people up, you are sequencing by resistance, and resistance is a real constraint in any organisation.
Knowledge check two. A team objects that removing dormant fulfiller roles will slow them down if those people ever return to the queue. What is the response? A, accept it, because the operational risk outweighs the saving. B, grant the role back on request within a day, and remove it now. C, remove half, as a compromise. Or D, escalate to their director for a decision. Pause, and ask what is actually being requested, and what it costs to provide.
The answer is B, and the general principle is worth more than the specific answer. The objection is about future access, and an operational alternative answers it completely, same day restoration costs the platform team minutes and costs the licence line nothing at all while the role is off. Answer A pays a standing premium for a hypothetical, which is precisely what the objection is asking you to do, usually without realising it. Answer C, remove half as a compromise, is the instinct of a peacemaker and it is the worst of both, you keep half the cost and all of the argument, and you have also signalled that your analysis is negotiable by volume rather than by evidence. And answer D escalates a question you can solve yourself, spending sponsorship you should be saving for the genuinely contested cases in week nine. Solve what you can solve. Escalate only what actually needs a decision.
The four structural fixes, and you have met all of them by now, which is deliberate. Templates default down, new starters get requester and App Engine User, with anything billable granted by exception, with a licence decision and a name attached. Gates on groups and roles, a maintained list of groups carrying billable roles plus sign off on any new custom role granting write on a task table, which closes the session eleven leak and the session thirteen one together. A licensing question in build review, what data does this application touch, is it scoped, how many tables, three questions on the design template answered before code. And joiner mover leaver hooks, so HR events strip billable access automatically, because without that the dormant population rebuilds itself from the top every single year. And then the note, which I want to repeat because it is the most ignored advice in this entire module. Budget for these explicitly. They produce no headline saving of their own, which is exactly why they are first to be cut when a programme gets squeezed, and cutting them is the reason the programme has to repeat in two years.
Guest analyst clip. I want to talk about how to get the structural work funded, because I have watched a lot of people fail at this and the failure is usually about framing rather than about merit. If you go to a steering committee and say, we would like to spend six weeks on templates and access hooks, you will lose, and you should, because that sentence contains no value proposition. What works is presenting it as the maintenance cost of the saving you just delivered. You have found, let us say, a million dollars. Wonderful, everybody is delighted. Now you say, this million recurs annually only if we change four things, and if we do not change them the estate regenerates and we are back here in two years having this conversation again, except we will have spent the credibility. Frame it as protecting a delivered benefit rather than as a new project, and it funds itself, because nobody wants to be the person who let a million dollars leak back out. The other technique that works is attaching it to the same approval as the cleanup, so it is one decision rather than two. If the structural work needs its own business case, written later, competing against other priorities, it will lose, every time, because it genuinely is less exciting than everything else on that list.
Frame it as protecting a delivered benefit, and attach it to the same approval as the cleanup so it is one decision. That is good organisational advice and it generalises well beyond licensing. Anything unglamorous that protects something valuable should be bundled with the valuable thing, never presented separately afterwards. Now, the evidence pack.
The evidence pack, five documents, built before anybody asks. The classification sheet, every named user with holds, does, should hold, and reason, dated and refreshed each quarter so that a series exists rather than a snapshot, and a series is much harder to argue with. The table register, every custom table with its application, its owner, live or retired, counted against the contracted allowance. The removal log, what was removed, when, and why, which is what turns your lower count from an assertion into a documented history, and that distinction matters enormously in a review. The boundary tests, for each custom application whether it reads or writes process data, decided deliberately and recorded rather than discovered later by somebody else. And the reconciled position, entitlement against consumption per line, signed by the licence owner, which is session ten's baseline now carrying module three's evidence underneath it. That is the pack. And notice it is not extra work, every item is a by product of the programme you just ran. The only additional discipline is keeping them, dated, in one place.
Knowledge check three, and it is a question about how you judge success. You finish the programme nine months early and find nothing to negotiate. No claim, no exposure, nothing outstanding. What did the work buy you? A, nothing this cycle, because the value only comes when a claim arrives. B, a lower renewal base, a documented position, and no exposure to settle. C, only the internal benefit of tidier access. Or D, a stronger case for buying more licences. Pause, and ask what the renewal is actually priced against.
The answer is B, and I put this check in deliberately because the framing trap in answer A catches genuinely sophisticated people. Finding nothing to negotiate is the successful outcome, not a wasted effort, and it can feel like the latter if you have been measuring yourself by exposure avoided. Look at what actually happened. The renewal prices against your corrected volumes rather than your drifted ones, so the base is lower. The uplift then compounds on that smaller base for the entire term, which is session four's arithmetic working in your favour for once. And there is no exposure hanging over the price conversation, which is the leverage point from session fourteen. Answer C undersells it substantially, and answer D is simply the wrong direction. And there is a quiet fourth benefit worth naming. A customer who visibly runs their own estate gets treated differently on everything else, pricing, escalations, roadmap conversations. Competence is noticed, and it is priced.
So what does module three leave you holding? Four things. A defensible user count, every named user classified on activity evidence with reasons recorded and quarterly history behind the number. A governed build, tables counted against the allowance, boundary tests recorded, and a licensing question sitting inside the design review where it belongs. Machinery that holds, templates, gates, build review, and leaver hooks, so the count does not quietly regenerate while your attention is elsewhere. And a pack you can hand over, evidence ready before it is requested, which is exactly what turns a compliance conversation into a formality. Five sessions on one line, which I justified at the start of session eleven by pointing at the four to six times multiple running through every person in your organisation, and I hope by now that allocation looks well spent. Module four moves outward to the rest of the estate, ITSM in depth, then ITOM, ITAM, HRSD and CSM, and SecOps, each with its own counting unit and, you will find, its own version of exactly these problems. One more clip.
Guest analyst clip. Let me leave you with the thing that surprised me most over years of doing this work, because it is not what I expected when I started. I assumed the customers who did best against large software vendors would be the biggest ones, or the toughest negotiators, or the ones with the most sophisticated procurement functions. And those things help, at the margin. But the strongest predictor, by a distance, is something much duller. It is whether somebody in the organisation maintains an accurate picture of what they have, continuously, when nothing is happening. That is it. Not a strategy, not a tactic, a maintained document and a person who owns it. And the reason it matters so much is that every commercial conversation with a vendor is ultimately a conversation about facts, and in almost every one of those conversations only one side has bothered to establish them. When both sides have the facts, the negotiation becomes narrow and technical and quite calm, and it lands close to fair. When only the vendor has them, the negotiation is about persuasion, and persuasion favours the party who does this for a living. So all of this work, the counting and the classifying and the registers, is really just about making sure the conversation is about facts. Everything else follows from that.
Making sure the conversation is about facts. That is module three, and honestly it is the whole course compressed into a sentence. Every document we have built, the user file, the inventory, the baseline, the classification sheet, the table register, exists to ensure both sides arrive with the same facts, because the side without facts negotiates on persuasion and loses. Let's recap and close the module.
Mitigation, in three sentences. The same finding costs staff time on a quiet Tuesday and full multi year rate at a signature date, and the only variable between those two prices is when you looked, which makes timing the largest lever in this entire discipline. Twelve weeks, four phases, count, classify, remove, then fix and paper, with the uncontested categories taken first because nobody loses a capability they actually exercise, and sequencing by resistance is how the programme survives contact with an organisation. And the structural fixes carry no headline saving of their own, which is exactly why they are the reason the programme never has to run in full again, so budget for them before somebody squeezes the plan. That closes module three. Next session opens module four with the biggest single line in most estates, ITSM itself, what each tier actually buys, and the renewal levers that apply to it specifically.
Homework, about an hour, and this week it is a plan rather than an analysis. One, place yourself on the curve, how many months to your renewal and which column of the cost table are you currently sitting in, and write that date down somewhere visible. Two, price the quick wins, machine accounts, dormant users, approvers, and retired application tables, multiplied by your net rates, producing one number you can take to a sponsor, because sponsors fund numbers rather than activities. Three, draft the twelve week plan, four phases against real calendar dates with a name against each phase, and I will be blunt, a plan without names is a wish. Four, pick the structural owner, who owns templates, gates, build review, and leaver hooks, and one name for all four is ideal because it means somebody is accountable for the machinery as a whole. And five, start the pack, create the folder, put the classification sheet and table register in it, and date them, because the pack begins existing the moment you make it and not before.
Further reading, five guides. The rightsizing playbook is today's twelve week programme in written form, with the data pull list and the cuts that survive a CFO review. The true up surprises guide covers the self monitoring discipline that keeps you permanently in column one of the cost curve, which is the goal state. The license audit guide shows what the same evidence pack does when a review arrives anyway, because sometimes it does regardless of your preparation, and it explains why prepared estates settle so much lower. The App Engine explainer covers the table and boundary half of the exposure with the governance that keeps custom build inside its allowance. And the pharmaceutical case study is this entire module executed at enterprise scale, and specifically what the programme produced when it ran early rather than late. That is session fifteen, and that is module three complete. Draft your twelve weeks, name your owner, and I will see you in session sixteen where we start on ITSM itself.