The return on retiring Agile is mostly avoided cost, and it is spread across budget lines that rarely get added together. Licences, servers, storage, backup, patching, specialist time and the risk of running unsupported software on your own network.
No single budget holder sees the total. Licences sit with procurement, servers with infrastructure, storage with the platform team, effort inside a managed service line, and risk on a register nobody prices.
Added up, a dormant Agile environment is expensive: application and database licences, the servers underneath them, production-grade vault and database storage, the backup and DR footprint behind that, the patch cycles and certificate renewals and access reviews, and the specialist who is the only person who still understands the configuration. All of it recurring, all of it for a system with few active users.
Against that sits the programme cost, which is one-off, and the archive cost, which is recurring but an order of magnitude smaller because storage on cold tiers costs a fraction of production infrastructure and needs no licences.
Then there is the part that is not a cost line until it is. Running unsupported software on the network means no security fixes, credentials nobody has rotated, and an application server that cannot be patched because the platform beneath it cannot be upgraded. That belongs in the case as a quantified risk rather than a footnote.
Gather these and the case mostly writes itself.
Agile application and the database underneath it, including support and maintenance renewals.
Server capacity, virtualisation or hardware, and the environments beyond production that are still maintained.
Vault and database storage at production grade, plus the backup and DR copies.
Days per year across patching, renewals, access reviews and incident handling, costed at a real rate.
What it currently takes to answer a historical question, multiplied by how often it happens.
Quantified as expected cost, not listed as a heading. An unsupported, unpatchable system on the network has a number.
Four weeks, mostly gathering numbers other people already hold.
Pull licence, infrastructure, storage, backup and effort figures from the teams that hold them, into one view.
Migration and archival sized from profiling, so the investment side is measured rather than assumed.
Ongoing archive cost by storage tier and retention period, which is the recurring cost that remains.
Unsupported software exposure expressed as expected cost, so it can sit in the case rather than beside it.
One-off spend against recurring saving, with every assumption written down and each one testable.
The hardest part of the business case is not the analysis, it is collection: the figures sit in six different places and no one person holds them all.
These are the specific numbers to ask for, and who typically holds each one.
And what each one usually overlooks.
Licences, storage, backup, patching and specialist time all recur. Sunk cost is last year's; the spend continues.
A dormant system generates no complaints until it generates an incident, and by then the options are worse.
The people who understand the configuration are the constraint, and every year of delay makes the discovery phase harder.
It is dominated by avoided recurring cost rather than by new revenue. Retirement releases application and database licences, the servers beneath them, production-grade vault and database storage, the backup and DR footprint, and the maintenance effort across patching, renewals and access reviews. Against that sits a one-off programme cost and a much smaller recurring archive cost. For a long-running instance with few active users, payback in the first eighteen to thirty months is typical, though the honest figure depends on your own licence and infrastructure numbers.
Add six lines that usually sit in different budgets: annual licence and support for the application and its database; infrastructure for production and any non-production environments still maintained; storage for the vault and database at production grade plus the backup and DR copies; maintenance effort in days per year at a real rate; the effort currently spent answering historical enquiries; and the quantified risk of running unsupported software. The exercise is mostly collection, and the total is usually larger than any individual owner expected.
Substantially less than the environment it replaces, and the shape of the cost changes. There are no licences, no application or database servers, and no production storage. What remains is object storage charged by the terabyte, with deep history on cold tiers at a fraction of warm-tier pricing, plus modest query costs that scale with actual use. For an instance consulted occasionally, the running cost commonly falls by an order of magnitude.
As a risk input with a date, not as a cliff. What happens after December 2027 is that Agile PLM no longer receives Premier Support, so the software keeps running but stops receiving fixes. The business-case consequence is that the risk line grows over time rather than jumping: an unsupported application and database on the network, with no security patches, becomes progressively harder to defend to an auditor or an insurer.
Some, and it is worth including honestly rather than overselling. Historical enquiries typically get faster, because the archive is indexed on the keys people actually search by. Access control improves, since archive access runs through current identity management rather than a legacy user table full of leavers. And the archive being open-format structured data means product history can join other enterprise data for analysis, which is usually an unplanned benefit rather than a justification.
Two things, consistently. Counting only licence cost, which understates the total by leaving out infrastructure, storage, backup and effort. And omitting the archive cost from the after state, which makes the saving look larger than it is and invites a challenge that undermines the rest of the case. A case that states both sides fully, with the assumptions written down, survives scrutiny better than one that maximises the headline number.
We will help you assemble the current-state cost across licences, infrastructure, storage and effort, size the programme from profiling, and model the archive so the payback stands up to scrutiny.