Software Asset Management Process: The SAM Lifecycle, Step by Step

The software asset management process is the routine that keeps three records in step: the software you run, the software you are entitled to run, and the gap between them. It is a loop of six steps: discover, collect entitlements, normalize, reconcile, optimize and report. Run on a schedule, it turns an audit letter into a check of figures you already hold.

This guide walks through each step the way we run it in our managed SAM engagements, and shows where the process usually breaks.

In short

• SAM is a process before it is a tool. A tool counts installations. The process turns that count into a license position a publisher will accept.
• The six steps repeat: discover what you run, collect what you own, normalize the data, reconcile it into an effective license position, optimize and remediate, then report and govern.
• Each publisher’s rules decide the hardest step. Oracle counts processors with its core factor, IBM expects ILMT reports at least every quarter, and a Microsoft position depends on the agreement and product use rights that apply.
• The process pays only if it repeats. A position built once goes stale with the next project, new hire or virtual machine.

The SAM lifecycle at a glance

Step 1: Discover what you run

Discovery scans every server, desktop and virtual host in scope and records the software on each. Agentless scanning reaches most of a network over standard protocols: WMI for Windows machines, SSH for Linux and Unix, and SNMP for network devices. An agent on Windows machines adds what a network scan cannot see, such as which installed software has actually been used recently.

For licensing, three things matter most and are the easiest to miss: Oracle Database options and packs, middleware features, and Java. A scan that reports “Oracle Database 19c” but not the options enabled underneath it will understate what an auditor later finds.

Step 2: Collect what you own

Entitlements are the other half of the record. They live in contracts, ordering documents, purchase orders and each publisher’s license portal, usually spread across procurement, finance and IT. Some are missing altogether: licenses bought by a company you acquired, or carried on an old support contract, are real but often absent from the records.

Record each entitlement with the product, the metric, the quantity, the agreement it falls under and any limits on where it may be used. Without this register, the reconciliation in step 4 compares your deployment against a guess. Our guide to software license entitlements lists the fields to record and where to find each one.

Step 3: Recognize and normalize the data

Raw scan data is untidy. One product appears under several names, versions and editions, and a scan of a virtual machine rarely says which physical hardware sits underneath it. Normalization turns this into one organized dataset, with each installation matched to a real product and edition and each server matched to its processor model and core count.

That second match decides the count for any product priced per processor or per core. In our own platform, Euladox, it runs against a hardware catalogue that covers every server built since 2001 and every major hypervisor and operating system version.

Step 4: Reconcile the data into an effective license position

Reconciliation applies each publisher’s rules to the clean data and compares the result, product by product, with what you own. The output is the effective license position: a surplus or a shortfall for every product.

The rules are where the work is, and they differ by publisher. Oracle counts processors with the core factor and sets Named User Plus minimums per processor. IBM’s lower sub-capacity pricing applies only when ILMT reports prove the conditions were met. A Microsoft position depends on which agreement and product use rights apply, and on whether licenses carry Software Assurance. A tool’s default report is a starting point here, not a position.

Step 5: Optimize and remediate

With a position in hand, act on it. Uninstall what nobody uses, move licenses to where they are needed, switch off options that were enabled by default, and check whether a different license metric would cost less for the same use. Fix any shortfall on your own timetable, before a renewal or an audit prices it for you.

Shelfware deserves particular attention because its cost repeats. On Oracle, annual support is 22% of what you paid for the license, every year, whether anyone uses it or not. Our guide on how to reduce software license costs sets out the five levers in the order we pull them.

Step 6: Report, govern and repeat

Report the position against the targets you set for cost, compliance risk and renewal dates. Then govern it: approval rules for buying and installing software, written into a software asset management policy, and an alert whenever a new deployment opens a gap. Then run the loop again.

How often depends on the publisher and on how fast your estate changes, and some publishers set the pace for you. IBM requires ILMT reports at least quarterly, kept for two years, for any product licensed by sub-capacity. Whatever the cadence, the position should be fresh before every renewal, true-up or ULA certification.

Mistakes we see in SAM processes

• Treating the first position as the finish line. The estate starts drifting from it the day after the report.
• Counting installations instead of license requirements. Forty database installations on a VMware cluster can need far more, or far fewer, licenses than forty.
• Leaving entitlements with procurement. If the people reconciling cannot see the contracts, they are reconciling against memory.
• Starting when the audit letter arrives. By then the publisher sets the timetable, and data released in a hurry is hard to correct later.

Who should run the SAM process

Most organizations choose between an in-house team with a SAM tool and a managed service. A SAM tool can find, count and reconcile your software. Turning that count into a position a publisher accepts takes people who know each publisher’s rules and who run the work on a schedule, not only when an audit letter arrives.

That is what our software asset management services do, year-round, on Euladox. If you would rather start with a single view of where you stand, the SAVE Review is an independent software license review of the whole software and cloud estate.

Questions we are asked about the software asset management process

What are the steps of the software asset management process?

Six steps, repeated: discover the software you run, collect the entitlements you own, normalize the data, reconcile it into an effective license position, optimize and remediate, then report and govern. The loop then starts again, because the estate keeps changing.

How often should the SAM process run?

Discovery and governance run continuously. A validated license position for each major publisher should be produced on a set schedule and refreshed before every renewal, true-up or ULA certification. For IBM products licensed by sub-capacity, IBM sets a floor of quarterly ILMT reports.

Is software asset management the same as IT asset management?

No. IT asset management tracks every IT asset, hardware included, across its life. Software asset management deals with software licenses, entitlements and compliance, and it needs publisher licensing knowledge that general asset tracking does not provide.

Can a SAM tool run the process on its own?

A tool automates discovery, normalization and much of the reconciliation. It cannot read your contracts the way a publisher’s auditor will, and it cannot decide how to remediate a gap. That judgment is the part of the process that saves money.