This manufacturer completes a partial Citrix exit saving 44% case study shows how moving the right half of an estate off the platform, rather than all of it, produced a larger saving than either staying put or leaving entirely. It is an anonymised composite built from real engagements. The client is described by sector and approximate scale only, with no named client or confidential detail disclosed.
Situation
The client was a global industrial manufacturer with roughly 7,500 Citrix users across corporate offices, plants, and distribution sites. Citrix delivered a mix of applications: a small number of heavy engineering and design tools that needed the platform's full delivery and graphics capabilities, and a much larger set of simple line of business and administrative applications that most plant and office staff used. A multi year agreement was approaching renewal, and the vendor had proposed an increase consistent with the repricing widely reported since the 2022 Cloud Software Group acquisition.
The company had treated Citrix as a single, indivisible commitment. Because a few critical workloads genuinely needed the platform, it had renewed the entire estate at the platform price for years, including thousands of users who barely touched what they were paying for.
Challenge
The renewal was framed as all or nothing. The vendor knew the engineering workloads were hard to move and priced the whole estate on the assumption that the company could not leave. The instinct internally was the same: Citrix is critical, so the increase is unavoidable. The challenge was to break the estate apart, prove which users actually needed the platform and which did not, and act on the difference inside the renewal window.
Citrix being critical for some users is not a reason to pay the platform price for all of them.
Approach
1. Measure real usage, not headcount
We measured what each user group actually ran, rather than what they were licensed for. The method mirrors our guidance on independent counter measurement. The data showed that close to half the estate used only simple published applications with no dependency on the platform's advanced delivery features.
2. Segment the estate by dependency
We split the 7,500 users into a critical segment that genuinely needed Citrix, principally the engineering and design users, and a movable segment running workloads an alternative could deliver. The split was based on the usage evidence, not on assumptions or org charts.
3. Cost and execute the movable segment
We built a real migration plan for the movable users, priced the alternative platform, the project, and the run cost, and confirmed the business case held. This was an executable plan, and the company carried it out for that segment. The migrated licenses were then retired from the Citrix estate.
4. Renegotiate the retained estate against the exit
With half the users gone and a proven, repeatable exit method, the remaining critical users were no longer a captive base. We renegotiated the retained agreement against a now credible threat to move more, anchoring on measured quantities rather than the vendor's inflated count.
Outcome
The combined effect was a 44 percent reduction in annual Citrix cost against the proposed renewal, measured as of 2026. Roughly half came from retiring the migrated licenses outright, and the rest from renegotiating the retained estate at a better unit price with capped increases for the next term. The engineering workloads stayed on Citrix, where they belonged, but at a defensible quantity and rate. The migration cost for the movable segment was recovered inside the first year through the licenses no longer renewed.
Lessons for buyers
First, a Citrix estate is rarely a single decision. The fact that some workloads must stay does not justify renewing every user at the platform price, and measuring real usage is what reveals the split. Second, a partial exit can beat both alternatives, capturing most of the saving of a full migration while avoiding the risk of moving workloads that genuinely need the platform. Third, the moved segment creates leverage on the segment that stays, because the vendor has seen you execute and knows you can do it again. Finally, the work has to rest on real usage data, since a partial exit aimed at the wrong segment raises cost rather than cutting it. For the wider principle, see our guidance on negotiating when you cannot leave the platform and our Citrix exit advisory service.
Frequently asked questions
Is this case study real?
It is an anonymised composite based on real engagements. The sector, scale, and outcome are representative, but no named client or confidential detail is disclosed.
What is a partial Citrix exit?
A partial Citrix exit moves a defined set of users or workloads off Citrix to an alternative while keeping the rest on the platform. It captures most of the cost reduction of a full migration without the risk and disruption of moving everything, and it creates leverage on the licenses that remain.
How did the manufacturer save 44 percent?
Roughly half the estate ran simple published applications that an alternative could deliver, so those users were migrated and their licenses retired. The remaining critical users stayed on Citrix but were renegotiated against a now credible exit, producing a combined 44 percent reduction against the proposed renewal as of 2026.
Why not complete a full Citrix exit?
A segment of the estate depended on capabilities the alternative could not match cost effectively. Forcing those workloads off would have raised cost and risk. Keeping them on a renegotiated Citrix agreement, with the moved users gone, was the lowest cost compliant outcome.
Does a partial exit work for any enterprise?
It works wherever part of the estate runs workloads an alternative can serve. Most large Citrix estates contain such a segment. Identifying it accurately, with real usage data, is what separates a credible partial exit from a costly mistake.
For related outcomes, see how a logistics group used exit leverage to cut Citrix 31 percent, and our guidance on Citrix sales tactics.