The largest savings came from reclamation rather than discount, because cutting the quantity you buy at all beats shaving the unit price on quantity you do not need
A 20 percent discount on licenses you do not need still costs more than zero. Cut the quantity first, then negotiate on what defensibly remains.
Prepared by Redress Compliance · August 19, 2026 · Multi vendor estates optimized. 40 to 60 estates, 2024 to 2025.
Executive summary
Dormant and duplicate licenses ran 10 to 25 percent of paid entitlement before reclamation. That is quantity, and no discount touches it.
Renewals negotiated inside the final 30 days moved 5 to 10 percent less than those started a year out, because there was no time to reclaim or rebaseline.
Estates without a usage baseline accepted vendor counts that overstated need by 8 to 20 percent. With no baseline, the vendor's number is the only number.
Reclamation and right editioning moved two to three times more money than the discount. Discount is the last lever, not the first.
What does optimization actually mean here?
Matching what you pay for to what you use, on a metric the contract recognizes. Discount is the last lever rather than the first.
Three layers, and the gap between them funds everything
- Entitlement: what you own, reconciled to contracts and the vendor metric.
- Deployment: what is installed, measured by inventory tooling.
- Usage: what is actually used, which is the gap that funds the savings.
The metric governs, not the install count
Reclamation has to follow the licensing metric, whether processor, user or capacity. The discipline draws on recognized practice including the asset management standard and cost frameworks such as the published cloud cost framework, and vendor guidance like the published license management resources shows exactly why the metric governs exposure.
What is the sequence, and why that order?
Six moves, run in order. The first three reclaim, the next two reprice, and the last hardens the position. Each move funds or de risks the next.
| Move | Action | Typical lever | What it depends on |
|---|---|---|---|
| Reclaim | Remove dormant and duplicate seats | 10 to 25% of entitlement | A usage baseline |
| Reharvest | Reassign before repurchasing | Avoids net new spend | The reclaim having happened |
| Right edition | Downgrade over provisioned tiers | 20 to 40% per seat | Feature usage data |
| Reprice | Counter the renewal uplift | 5 to 15% on volume | A smaller defensible quantity |
| Reterm | Align co terminated end dates | Leverage consolidation | Calendar visibility |
| Harden | Lock the metric and price protection | Audit and renewal defense | Everything above it |
Why repricing comes fifth
Because a discount applies to whatever quantity is on the order. Running the reprice before the reclaim applies your best percentage to your worst count.
The baseline is the artifact everything hangs on
A usage baseline reconciled to the contract metric is what turns optimization from a discount ask into a quantity decision. Without it the vendor's count is the only count in the room.
Build the baseline before the renewal needs it
Reconcile entitlement, deployment and usage across the estate while there is still time to act on the gap.
Open the check →What 40 to 60 multi vendor estates showed
Across roughly 40 to 60 multi vendor estates Fredrik Filipsson and the team optimized between 2024 and 2025, the largest savings came from reclamation, not discount. Three patterns recur.
- Dormant and duplicate licenses ran 10 to 25 percent of paid entitlement before reclamation.
- Renewals negotiated inside the final 30 days moved 5 to 10 percent less than those started a year out.
- Estates without a usage baseline accepted vendor counts that overstated need by 8 to 20 percent.
A discount on licenses you do not need still costs more than zero. Cut the quantity first.
- Instant percentile standing: market low, median and high for deals like yours
- Renewal uplift exposure modeled over the full term, with the cap to ask for
- A ranked savings queue with dollar values, not license counts
When should the work actually start?
Nine to twelve months out. The reclamation and baseline work takes a quarter, and the negotiation needs runway to use the evidence it produces.
A renewal opened in the final month hands over the leverage
There is no time to reclaim, rebaseline or build an alternative. Renewals negotiated inside the last 30 days moved 5 to 10 percent less than those started a year out, and the difference was the preparation rather than the argument.
Stagger the renewals so each one has runway
Multi vendor sequencing exists so that every renewal gets its own preparation window rather than three arriving in the same quarter with one team between them.
How does compliance constrain the reclaim?
By deciding what a reclaim actually is. The licensing metric governs, so removing an install does not always remove a liability.
Processor, user and capacity behave differently
Reclaiming a named user removes a unit. Reclaiming an install on a processor licensed product may remove nothing at all unless the capacity position changes with it.
Reharvest before you repurchase
Reassigning a reclaimed license to the next requester avoids net new spend entirely, and it is the move most often skipped because purchasing is easier than administration.
Harden the position once the quantity is right
Lock the metric definition and the price protection into the agreement, so the work does not have to be repeated from scratch at the next renewal. The vendor licensing references such as the published licensing library are where the metric definitions to lock against are written down.
Where the common advice on optimization is wrong
The standard advice is to focus optimization on negotiating a deeper discount at renewal, so buyers put their energy into the price column. We disagree.
Reclamation moved two to three times more than the discount
Across the estates optimized, reclamation and right editioning cut the quantity you buy at all rather than shaving the unit price, and that moved multiples more money.
The buyer side move is to reclaim, reharvest and right edition first, then negotiate price on the smaller, defensible quantity that remains. The practice view sits in cost optimization services.
What the estates measured, 2024 to 2025
Two cuts of the optimization file, and the second explains why the first is usually missed.
Before reclamation, across multi vendor estates where nobody had reconciled entitlement against actual usage.
In estates with no usage baseline of their own, where the vendor's number was the only number available in the room.
The second figure is why the first survives. Without a baseline there is nothing to argue with, so the count stands.
Your first five moves
- Build the usage baseline reconciled to the contract metric, because estates without one accepted counts that overstated need by 8 to 20 percent.
- Reclaim the dormant and duplicate seats first, which ran 10 to 25 percent of paid entitlement and need no vendor agreement.
- Reharvest before repurchasing, reassigning what you reclaimed to the next requester rather than buying net new.
- Right edition the over provisioned tiers, worth 20 to 40 percent per seat and invisible without feature usage data.
- Only then negotiate price, and start 9 to 12 months out. The optimization practice runs the reclamation before the renewal, which is the only order that pays.
Frequently asked questions
What does optimization mean here?
Matching what you pay for to what you use, on a metric the contract recognizes. Discount is the last lever rather than the first.
What moves the most money?
Reclamation and right editioning, which moved two to three times more than the discount because they cut the quantity rather than shaving the unit price.
How much entitlement is dormant?
Between 10 and 25 percent of paid entitlement before reclamation, across the multi vendor estates optimized.
Why does the baseline matter so much?
Because without one the vendor's count is the only count in the room. Estates without a baseline accepted numbers overstating need by 8 to 20 percent.
What are the three layers?
Entitlement, what you own reconciled to the metric. Deployment, what is installed. Usage, what is actually used, which is the gap that funds the savings.
Why is repricing fifth in the sequence?
Because a discount applies to whatever quantity is on the order. Repricing before reclaiming applies your best percentage to your worst count.
When should the work start?
Nine to twelve months out. The reclamation and baseline work takes a quarter and the negotiation needs runway to use the evidence.
What does a last month renewal cost?
Between 5 and 10 percent against one started a year out, because there is no time to reclaim, rebaseline or build an alternative.
Does removing an install always reclaim a licence?
No. The metric governs. Removing a named user removes a unit; removing an install on a processor licensed product may remove nothing unless capacity changes.
What is reharvesting?
Reassigning a reclaimed licence to the next requester instead of buying net new. It is the most often skipped move because purchasing is easier than administration.