SaaS subscription management: a practical guide

SaaS subscription management is the discipline of knowing, for every piece of software your organisation pays for: who the supplier is, what you are contracted to, what it costs, when it renews, how much notice you must give to leave, who owns the relationship internally, and how many of the licences you bought are genuinely being used.

It sits between three teams who usually hold different pieces of the picture. IT knows which applications exist and who has access. Finance knows what is leaving the bank account. Procurement knows what was signed. When those three records never meet, software decisions get made on partial information — and the most common result is a renewal that happens by default rather than by decision.

Most organisations already have several partial lists: a finance export, a licence spreadsheet, an admin console per vendor, and a folder of contract PDFs. Improving any one of them does not fix the problem, because none of them is complete on its own.

The first practical step is to agree a single record where every subscription lives, and to decide which source wins when two sources disagree. A workable rule of thumb: the vendor's own admin data is the authority on seats assigned and edition, finance is the authority on what was actually paid, and the signed contract is the authority on term, notice period and price.

Seats purchased and seats actually used are two different numbers, and the gap between them is where most avoidable software cost sits. The gap usually comes from a small number of repeatable causes: leavers whose licences were never reclaimed, people moved onto a higher edition than their role needs, two tools bought by different teams doing substantially the same job, and seats bought ahead of a hiring plan that changed.

Usage data only becomes useful when you can tie it back to a specific contract line. Knowing that thirty people have not signed in for ninety days matters when you also know those thirty seats sit on a contract that renews in seven weeks and requires thirty days' notice.

Treat any calculated opportunity as a hypothesis to validate against your own usage data before you act on it — not as a saving you have already made.

The date that actually constrains you is not the renewal date — it is the last day you can give notice. A contract renewing on 1 June with sixty days' notice stops being negotiable at the start of April. Teams that track only renewal dates routinely discover this a fortnight too late.

A workable renewal rhythm works backwards from the notice deadline: review usage and internal demand well before it, decide whether to renew, reduce, consolidate or exit, and make contact with the supplier while you still have the option not to renew. Going into a renewal conversation with your own usage evidence changes the discussion from price-taking to an informed one.

Software arrives on the ledger in inconsistent shapes: monthly and annual billing, per-user and flat fees, multiple currencies, invoices via a reseller, and a long tail bought on company cards. Until those are normalised to a common basis — typically an annualised figure per supplier and per product — spend cannot be compared or trended honestly.

Once it is comparable, the useful questions become answerable: which suppliers account for the bulk of software spend, where is spend growing faster than headcount, and which categories contain more than one tool doing the same job.

Governance here does not mean a certification. It means that each subscription has a named owner, that access changes are recorded, that renewals are decided rather than defaulted, and that you can show the evidence behind a decision months later.

That evidence base is also what makes audits and internal reviews straightforward: a complete inventory, entitlement figures with a stated source, and a history of changes.

You do not need a complete dataset to get value. Start with the suppliers that represent the largest share of spend and the contracts renewing soonest — that is usually a short list.