Contents
Key takeawaysThe discount bandRenewal upliftCloud Pak VPC countsILMT and the true upTerms that beat the discountWhat the engagements showedWhen to startWhat to do nextFAQThere is no published IBM ELA discount. The band below comes from closed agreements, and the number that decides where you land in it is the ILMT reconciliation you bring, not the percentage you ask for.
- Median 27 percent. Across 30 IBM ELA and audit defense engagements in 2024 and 2025, the median discount achieved from IBM's opening best and final offer was 27 percent, inside a 19 to 34 percent band at enterprise scale.
- The top of the band is earned, not asked for. Buyers above 30 percent carried a credible Red Hat OpenShift carve out or a documented exit path on the legacy middleware base.
- Uplift is the quiet loss. Renewal uplift ran 18 to 32 percent above the prior contract value unless a cap was anchored at the original signing.
- VPC counts over commit. IBM proposed Cloud Pak VPC counts sat 22 to 41 percent above actual deployment on most estates.
- ILMT sets the baseline. The median PVU finding was 3.2 times the customer's own estimate, almost always because ILMT data was missing or unreconciled.
How much discount should you get on an IBM ELA?
A prepared enterprise buyer closes 19 to 34 percent below IBM's opening best and final offer, with a median of 27 percent across the 30 engagements in our 2024 to 2025 file. The percentage is measured against IBM's own opening position, not against a published price list, because IBM does not publish one for the ELA.
| Position in the band | Discount from the opening best and final | What it took |
|---|---|---|
| Bottom | 19 to 22 percent | Accepted the proposed Cloud Pak VPC count and the bundled Red Hat line; negotiated on percentage alone. |
| Median | 27 percent | ILMT history reconciled to entitlement before pricing; metrics translated where the usage supported it; uplift cap requested. |
| Top | 30 to 34 percent | Red Hat OpenShift carved out or priced separately, a documented exit path on legacy middleware, and swap, drop and true down rights written in. |
Why the opening best and final is not final
IBM's first proposal is built on its own view of your deployment, which is usually larger than yours once ILMT is reconciled. The "best and final" label describes the approval level the account team has used so far, not the limit of what IBM will sign. Every one of the 30 engagements moved from it.
What the enterprise scale band assumes
The band applies to agreements that bundle WebSphere, MQ, Db2 or the Cloud Pak equivalents with Red Hat and a renewal term of three years or more. A single product renewal under 1 million dollars behaves differently: the deal desk does not engage and the practical ceiling is lower.
- Scale: multi product estates with Red Hat or Cloud Paks in the bundle.
- Term: three years or longer, usually with an annual true up.
- Evidence: ILMT reports the buyer has reconciled, not IBM's audit output.
What does an IBM ELA renewal uplift look like without a cap?
Renewal uplift averaged 18 to 32 percent above the prior contract value in our file unless the buyer had anchored a cap at the original signing. The uplift arrives as a repriced bundle, not as a line called uplift, which is why most buyers only see it once the renewal quote is set beside the expiring order.
| Renewal regime on a 10 million dollar three year ELA | Renewal value | Difference against the prior term |
|---|---|---|
| Uncapped, bottom of the observed range | 11.8 million | 18 percent, or 1.8 million |
| Uncapped, top of the observed range | 13.2 million | 32 percent, or 3.2 million |
| Cap of 3 percent written at the original signing | 10.3 million | 3 percent, or 0.3 million |
The arithmetic is illustrative on a round base; the percentages are the observed range. The cap is worth more than any discount point you will win at renewal, and it can only be bought at the signing before.
Where the uplift hides: Cloud Pak repricing and sub capacity
Two mechanisms carry most of the increase. The first is the move from PVU priced products to Cloud Pak VPCs, where IBM's conversion ratio sets a new base. The second is the loss of sub capacity eligibility when ILMT reporting has lapsed, which reprices virtualized estates at full capacity.
How far do IBM Cloud Pak VPC counts over commit?
The IBM proposed Cloud Pak VPC count over committed against actual deployment by 22 to 41 percent on most of the estates we rebuilt. The count is usually derived from the PVU entitlement being converted, not from what is running, so it carries every unused entitlement forward into a subscription.
How to size VPCs from deployment
- Pull the ILMT deployment report for the products being converted, by environment.
- Remove decommissioned, duplicate and non production instances that the entitlement count still carries.
- Convert the deployed capacity to VPCs with IBM's published ratios, product by product.
- Add a growth allowance you can defend from the project pipeline, not from IBM's forecast.
- Put the resulting count on the table with the report behind it.
Buyers who arrived with this count rather than IBM's negotiated the discount on a base 22 to 41 percent smaller, which is worth more than the discount itself on most estates.
Why does ILMT decide the number before IBM does?
The median PVU finding in our audit defense engagements was 3.2 times the customer's own estimate, almost always because ILMT data was missing or had never been reconciled to entitlement. IBM's renewal proposal is built on that finding, so the buyer who has not done the reconciliation negotiates on IBM's number.
In the true up estimates we reviewed, IBM's demand ran 25 to 60 percent above the defensible figure once deployment was reconciled to entitlement and shelfware was removed. The true up is a negotiation, and IBM defaults to the figure that recovers its day one discount.
The reconciliation sequence
- Entitlement first: Passport Advantage reports, proofs of entitlement and the license information documents for every product in scope.
- Deployment second: ILMT reports with the agent coverage gaps closed, by quarter, for the full sub capacity history.
- Net position third: deployment against entitlement, with shelfware netted and metrics translated where usage supports it.
Which IBM ELA terms move more money than the discount?
Three clauses moved more money than the headline percentage in our file, and none of them sat in IBM's default template. Swap, drop and termination for non use rights were negotiable in roughly seven of nine ELAs when requested explicitly, and appeared by default in fewer than one in four.
| Clause | IBM default template | Buyer target |
|---|---|---|
| Renewal uplift | Silent; renewal reprices to the then current bundle | Written cap of 3 percent or CPI, whichever is lower, for the renewal term |
| Swap rights | Absent | Exchange unused entitlements for other products of equal value at each anniversary |
| Drop and true down | Absent; quantities fixed for the term | Reduce committed quantities at anniversary within a stated band |
| Termination for non use | Absent | Remove products with no deployment after twelve months without penalty |
| Price hold | Quote valid to quarter end | Unit rates held for defined additions across the term |
| Red Hat | Bundled at a blended rate | Priced separately with its own discount and exit terms |
Why Red Hat is the lever
Red Hat OpenShift and RHEL carry their own competitive market, which the rest of the IBM bundle does not. Buyers who carved Red Hat out, or priced it separately with a documented alternative, reached the top of the discount band on the whole agreement, because IBM prices the bundle against the line it might lose.
What have recent IBM ELA renewals shown?
One renewal shows how the pieces combine. IBM's proposal came in at 22 million dollars across three years. The ILMT gap analysis removed the capacity IBM had assumed, metrics were translated where the usage supported it, and Red Hat came out of the bundle. The renewal closed at 14.7 million, 33 percent below the proposal.
The renewal price follows the evidence you bring. Without ILMT history and your own sizing, IBM's proposal is the only model in the room.
Where the common advice on IBM ELA discounts is wrong
The standard advice is to push the discount percentage as hard as possible before signing, because the ELA locks it for three years. We disagree. Across the 30 IBM ELA and audit defense engagements Morten Andersen ran in 2024 and 2025, the percentage moved the price less than the base it was applied to. The proposed Cloud Pak VPC count over committed by 22 to 41 percent, and the uncapped renewal uplift of 18 to 32 percent returned most of the discount at the next term. Reconcile ILMT to entitlement first, size the VPCs from deployment, carve Red Hat out, and write the cap and the swap rights in. Then negotiate the percentage on the base you have proved.
What the numbers do not cover
The band is drawn from upper enterprise estates in North America and Europe. Public sector framework pricing, mainframe MLC and single product renewals under 1 million dollars sit outside it. A quote that lands inside the band is a starting point for the structural terms, not a finished deal.
When should the IBM ELA work start?
Eighteen months before expiry. The ILMT reconciliation alone takes two quarters on a large estate, and IBM's fourth quarter carries the heaviest software bookings, so a December close gets the most approval attention only if the evidence is already on the table.
| Months before expiry | What to do | Why it matters |
|---|---|---|
| 18 | Close ILMT agent gaps and start the entitlement reconciliation. | Every later number depends on it. |
| 12 | Size Cloud Pak VPCs from deployment; decide which products convert, carry forward or drop. | Sets the base the discount applies to. |
| 9 | Cost the Red Hat alternative and the legacy middleware exit path. | Buys the top of the band. |
| 6 | Send the clause set: uplift cap, swap, drop, termination for non use, price hold. | Terms move more than the percentage. |
| 3 | Receive the proposal; test it against the band and the reconciled base. | Finds the room in the best and final. |
| 1 | Close inside IBM's quarter with signature authority in the room. | Approval attention without a compressed calendar. |
Primary sources for this guide
- IBM Passport Advantage: the agreement framework the ELA is written under.
- IBM License Metric Tool: the ILMT reporting that decides sub capacity eligibility.
- IBM Cloud Paks: the VPC bundles most renewal proposals are built on.
- IBM software licensing: the metric definitions and license information documents.
Source: Redress Compliance advisory engagement file, 2024 to 2025.
What to do next
- Eighteen months out. Close the ILMT coverage gaps and start the entitlement reconciliation.
- Size the base. Convert deployed capacity to VPCs yourself and remove the shelfware before IBM proposes.
- Carve Red Hat out. Price OpenShift and RHEL separately with a costed alternative behind them.
- Send the clause set early. Uplift cap, swap, drop, termination for non use and a price hold, before the first proposal.
- Benchmark the proposal. Test IBM's best and final against the 19 to 34 percent band on VendorBenchmark, free and without signup.
- Close on your calendar. Our IBM ELA renewal team runs the renewal with you, fixed fee or 25 percent of what we save.
Frequently asked questions
How much discount should you get on an IBM ELA?
A prepared enterprise buyer closes 19 to 34 percent below IBM's opening best and final, with a median of 27 percent in our 2024 to 2025 file. The top of the band went to buyers with a Red Hat carve out or a documented exit path on legacy middleware.
Is IBM's best and final offer really final?
No. The label describes the approval level the account team has used so far. Every one of the 30 engagements in our file moved from it once the buyer brought a reconciled ILMT position and a smaller VPC count.
What renewal uplift does IBM apply to an ELA?
Without a cap written at the original signing, renewal uplift ran 18 to 32 percent above the prior contract value in our file. A cap of 3 percent or CPI, whichever is lower, is the buyer target and is worth more than any discount point won at renewal.
How accurate are IBM's Cloud Pak VPC proposals?
They over commit. The IBM proposed VPC count sat 22 to 41 percent above actual deployment on most estates we rebuilt, because it is derived from the PVU entitlement being converted rather than from what is running.
Why does ILMT matter for the ELA price?
Because IBM's proposal is built on its audit view of your deployment. The median PVU finding was 3.2 times the customer's own estimate where ILMT data was missing or unreconciled, and that finding becomes the base the renewal is priced on.
Can you get swap and drop rights in an IBM ELA?
Yes, when you ask. Swap, drop and termination for non use rights were negotiable in roughly seven of nine ELAs we ran when requested explicitly. They appeared in IBM's default template on fewer than one in four.
Should Red Hat stay inside the IBM ELA bundle?
Usually not at a blended rate. Red Hat has a competitive market the rest of the bundle lacks, so buyers who priced OpenShift and RHEL separately, with an alternative costed, reached the top of the discount band on the whole agreement.
Where can I check an IBM proposal against the band?
Run the proposal through the free VendorBenchmark decoder, which normalizes it against closed IBM agreements, then bring the result to our IBM ELA renewal team. Compare against the reconciled base, not the quote as sent.