Start with the usage report before you touch the contract
The number that actually moves a multi-year SaaS renewal is not the discount in the vendor's first email. It is the usage number your own team pulls before anyone from the vendor side gets on a call. Before I let a client's IT director or VP of Revenue Operations even open the renewal quote, I ask for three things: the current license count by type, an active-user report going back six months, and every amendment signed since the last renewal. That last one matters more than people expect, because most SAP and Salesforce contracts pick up riders over three years that nobody remembers signing.
For Salesforce, the active-user pull is straightforward: export the login history object, filter by last login date, and cross it against the named-user license count in Setup. For SAP, it takes more work. Named-user measurement runs through USMM and gets consolidated in the License Administration Workbench, and if your Basis team hasn't run that consolidation in a year, the numbers you get during renewal week are already stale. Either way, do this pull yourself. Do not accept the vendor's own usage dashboard as your baseline, because their counting rules for what counts as active tend to run generous in exactly the direction that keeps your bill high.
The other document to have open before the first call is your own contract's true-up and true-down language. Most multi-year enterprise agreements let you add seats mid-term at the contracted rate, called a true-up, but say nothing at all about removing seats, meaning you can grow into the deal but never shrink out of it. Reading that clause now, not during the renewal call, is what a licensing change checklist is for. Know before you sit down whether you are even structurally able to negotiate a lower seat count, or whether the contract already answered that question for you three years ago.
A third of the seats had gone quiet, and that number is the real starting point
Here is the scenario, and it is closer to typical than most procurement teams want to admit. A client came to me six weeks out from a Sales Cloud renewal covering 1,020 named-user licenses. We pulled the login history and found 340 of those licenses had not logged in in six months. Zero logins, not slow adoption. Sales operations had onboarded three regional teams two years earlier, two of those teams had since moved to a different CRM workflow, and nobody had gone back to remove the seats.
That number, 340 out of 1,020, is worth more in the room than any discount percentage the account executive offers first. It reframes the entire conversation. The vendor's opening position assumes you are renewing your current footprint and probably growing it. A third of your seats sitting idle changes what the vendor is negotiating against. Instead of a renewal risk to manage, you look like a customer about to right-size downward, and the only real question left is how much of that reduction shows up in their number this quarter. That is the real cost of an unused license made concrete instead of theoretical, and a login report is much harder for an account team to argue against than a general complaint about price.
What that number is not worth is a straight dollar-for-dollar credit. Vendors will not simply subtract the cost of 340 unused seats from your quote, because license count and price per seat are two separate levers on their side, and they would rather move price a little than cut volume a lot. The way to use the number is to walk in already having decided your target seat count, in this case 680, and let the vendor counter from there. Anchoring on your own usage data, not their proposal, is the entire point of doing the pull in the first place.
What the renewal timeline pressure actually looks like
Vendor renewal timelines are built to work against you, and once you've seen the pattern a few times it stops feeling personal. The first email usually lands ninety days out, friendly and informational. Around day sixty, a specific discount appears, tied to a specific signature date, usually the vendor's own quarter end rather than your contract's actual expiration date. Around day thirty, the account executive gets joined by a renewal specialist or a regional director, and the tone shifts from partnership language to consequences: price protection expiring, support tiers reverting. Sometimes there's also a note about auto-renewal clauses that will kick in if nothing gets signed.
That auto-renewal clause is worth reading in January, not in week eleven of a twelve-week countdown. Most enterprise SaaS contracts require sixty or ninety days of written notice to prevent an automatic renewal at list price, and vendors are not shy about letting that window close on customers who are still negotiating in good faith. I've seen a SAP renewal play out exactly this way: the client's negotiating team was still going back and forth on year-two pricing when the notice deadline passed, and the fallback position in the contract auto-renewed the full named-user count at the prior rate with no discount attached. The same pricing pressure that shows up right after go-live, the kind covered in any honest account of SAP support after go-live, doesn't disappear once the system is stable. It just moves to renewal week.
The other pressure tactic to watch for is bundling. A renewal conversation is also a sales conversation, and account teams get measured on expansion, not just retention. Expect a new module, a Data Cloud add-on, an AI feature tier, or an SAP RISE conversion to get folded into the renewal quote as a package deal, with the discount on your existing seats made conditional on adopting the new piece. Separate those two conversations on purpose. Decide what you actually want to buy on its own terms, and don't let a discount on seats you already own depend on a purchase you hadn't planned to make this year.
The trade a multi-year deal is actually offering you
A three-year commitment buys you something real: a capped annual increase, usually in the low single digits, instead of whatever the vendor's list price does when it resets every twelve months. It can also buy price protection against edition changes, meaning if the vendor restructures its packaging the way Salesforce has done more than once, you keep your current feature set at your current rate instead of getting pushed into a new edition boundary mid-term. Both of those are worth having, and I'm not telling clients to avoid multi-year terms. I'm telling them to know exactly what they're giving up for it.
What you give up is the ability to change your mind. A one-year term lets you shrink your footprint every twelve months as usage data comes in. A three-year term locks in a seat count, or a floor below which you cannot drop, for the entire term, and if the same pattern that produced those 340 idle licenses keeps happening, you'll spend three more years paying for a number that was already wrong at signature. Look past the sticker price in year one and work out the total cost picture across a Salesforce edition before you sign anything.
The other lock-in is structural, not financial. On the Salesforce side, moving between editions mid-contract can trigger a full renegotiation rather than a simple change order, and that's exactly the kind of buying boundary between Salesforce editions a renewal conversation should force you to name explicitly. On the SAP side, a multi-year named-user agreement signed alongside a RISE or clean-core commitment ties your licensing to an architecture decision that outlives the person who approved it. Go into a multi-year term having asked, out loud, in the room, what happens if this number needs to change in year two.
Bring people into the room who can say no
The renewal call goes worse when procurement is the only person from your side in the room, because procurement can negotiate price but usually cannot answer whether a module actually gets used, and the vendor's account team knows the difference immediately. Bring the actual business owner, the VP of Sales Operations or the SAP Basis lead who ran the usage report, into the internal prep meeting even if they don't join the vendor call itself. They're the ones who can say, on the spot, whether cutting a hundred seats breaks something real or just deletes shelfware.
Bring finance too, and specifically someone who can model total cost across the full term, not just year one. A discount that looks generous over three years can still cost more than a shorter term with no discount once you factor in a seat count you already know is inflated. Line up an executive sponsor in advance as well, so the vendor can't escalate past your negotiator to someone in your building who hasn't seen the usage data and might agree to something your prep team already ruled out.
Leave the vendor's renewal specialist without a soft target inside your organization. If everyone on your side has seen the same login report and agrees on the same target seat count, there's no one left for a well-timed phone call to talk into a worse deal than the one your team already agreed to walk in with.
Set the walk-away number before the vendor gets on the call
Before the first real negotiating call, write down the seat count you actually need based on your usage data, and the maximum per-seat price you will accept at that count. Then write down what you would actually do if the deal fell through: run month-to-month for a quarter, or move the workload onto a module you already own. Most teams never write that second part down, and it's exactly the part that keeps a vendor's deadline pressure from working on you.
My client with the 1,020 licenses ended up signing a three-year renewal at 700 seats, not the 680 the usage data alone suggested, because two teams had a hiring plan that justified the buffer. The price per seat came in lower than their previous agreement, and the true-up clause capped annual growth at ten percent with no matching true-down option, a trade-off they decided they could live with given the rest of the deal.
Pull your own login data before your next renewal cycle starts, not during it. The report costs an afternoon. Walking into a multi-year renewal without it costs three more years of paying for a number nobody ever actually checked.
