The CFO's guide to cutting IT spend without slowing the team down
Most cost-cutting programs fail because they treat IT like overhead. They aren't. The fastest 20–35% of savings come from visibility, not austerity — and you can ship them in a quarter.
If you are a CFO or finance director reading this, you have probably already asked your IT counterpart for "the list of everything we pay for." You probably got back a spreadsheet that was 60% complete, two months out of date, and missing the contracts that auto-renewed last week.
That is normal. It is also fixable. The IT estate of a typical 50–500 person company is fragmented across at least seven systems: an asset register here, a procurement tool there, identity provider, expense reports, a forgotten Notion page, a Slack channel called #it-renewals that no one reads, and the inbox of one overworked IT manager. None of those systems agree with each other.
The CFO's job is not to memorize that mess. It is to install the four cost levers below, in order, and let the data surface what to cut.
The four levers (in priority order)
- Visibility. One canonical source of truth for every asset, license, contract, and renewal date.
- License rightsizing. Stop paying for seats nobody uses.
- Contract discipline. Kill silent auto-renewals and consolidate vendors.
- Hardware lifecycle. Refresh, reassign, retire — on a schedule, not on rumour.
Pulling them in this order matters. Lever 1 is the only one that gives you the data to pull the others. Skip it and you'll cut blindly — usually cancelling the wrong tool and triggering a productivity crater that costs more than the savings.
Rule of thumb
Visibility alone, with no further action, typically surfaces 8–12% of immediate savings (cancelled trials, abandoned tools, duplicate licenses). The remaining 12–23% comes from levers 2–4 over the next 90 days.
Lever 1 — Visibility: kill the spreadsheet
The first month is the hardest because there is nothing to optimise yet. The work is connecting your data sources and reconciling them into a single inventory: people, hardware, software, licenses, contracts, renewal dates, costs, owners. You want to be able to ask, in one place: "What does Maria in Marketing have? What does it cost us? When does it renew?"
What "visibility" actually requires
- Identity sync — every active employee, contractor, and seat-bearing identity from your IdP (Google, Microsoft, Okta).
- Hardware register — every laptop, monitor, phone, and accessory, with its assignee and warranty status.
- License inventory — every SaaS subscription, with seats purchased vs. seats used, last login data, and cost per seat.
- Contract repository — uploaded PDFs with parsed start/end/auto-renew clauses and notice periods.
- Spend mapping — each contract tied to a cost centre, owner, and category.
If you are starting from scratch, give yourself two to three weeks. If you already have a fragmented setup, the consolidation typically takes one week of clean-up and one week of validation. Tools like a modern IT asset management platform exist to compress this work.
Lever 2 — License rightsizing
This is where most of your fast wins live. The mechanics are simple: for every paid SaaS, compare seats purchased to active users in the last 30/60/90 days. Anything below 60% utilisation is a candidate to downsize at the next renewal.
The typical numbers we see in mid-market companies:
| Category | Median over-licensing | Notes |
|---|---|---|
| Productivity (Microsoft 365, Google Workspace) | 9–14% | Tier mismatches more common than empty seats |
| Collaboration (Slack, Zoom, Notion, Asana) | 15–25% | Leftover seats from departures, contractor churn |
| Design / Engineering tools | 20–35% | Trials never deactivated, dual-tooling |
| Sales / Marketing stack | 25–40% | Heaviest waste — pilots, dormant accounts |
| Specialised vertical SaaS | 10–20% | Smaller absolute numbers, but high $/seat |
Three rightsizing tactics that actually work
- Last-login sweeps. 90 days no login → reclaim seat. Automated.
- Tier downgrades. Premium users on tools they only use the basic features of. Audit feature usage, not just login.
- Reassignment. Before buying a new seat, the system suggests reusing a dormant one.
Lever 3 — Contract discipline
Auto-renewal is the most expensive default in software. The vendor designed it that way. Your job is to flip it from "we pay until someone complains" to "we re-evaluate before every renewal."
Three policies, each non-negotiable:
- 90-day pre-renewal alert on every contract, routed to the budget owner and finance.
- Mandatory renewal review for any contract above a threshold ($5k/$10k/$25k depending on company size). Owner has to justify continuation in writing.
- Vendor consolidation review annually. Anywhere you pay two vendors for overlapping capabilities, default to picking one.
The contract repository matters more than the alerts. If your renewal notice arrives in someone's email and the PDF lives on a shared drive, the alert can't actually compare against the terms. Centralise the contracts first; the playbook for contract centralisation is here.
Lever 4 — Hardware lifecycle
Hardware is the most visible cost (everyone sees a laptop) and the worst-tracked. The two failure modes are predictable:
- Ghost hardware. Laptops assigned to people who left, sitting in a drawer, not returned, not retired.
- Premature refresh. Buying new gear because no one knows the warranty status of the existing fleet.
A defensible hardware policy rests on three numbers per device: assignee, warranty end, refresh due date. With those three, you can rotate hardware on a schedule, recover ghost devices automatically at offboarding, and stop emergency-buying laptops because someone "needs one tomorrow."
Detail on this is in the hardware lifecycle guide.
The dashboard the board actually wants to see
Once the four levers are running, a single dashboard usually replaces three quarterly slide decks. It has four numbers:
- Total IT spend, broken down by hardware / software / services, vs. last quarter and headcount-normalised.
- Software utilisation rate — % of paid seats that were active in the last 30 days.
- Renewal pipeline — total $ in contracts renewing in the next 90 days, with owner.
- Asset coverage — % of employees with their full hardware + software entitlement tracked. Below 95% is a red flag.
What this looks like 90 days in
A realistic outcome for a 200-person company starting cold:
- Month 1 — 100% asset/license visibility. Cancelled trials and dormant tools surface 6–10% savings immediately.
- Month 2 — first wave of license rightsizing at the next two renewal events. 5–8% additional savings, locked in for the year.
- Month 3 — vendor consolidation review identifies 2–4 overlapping tools. 3–7% additional savings, plus a simpler stack for the team.
The ceiling is higher with patience: companies that run this discipline continuously for 12 months consistently report 30–35% reductions vs. their baseline, with no measurable productivity drop.
Common failure mode
Treating this as a one-off audit. The savings compound only if visibility is continuous and renewals are managed proactively — not when someone runs the spreadsheet again twice a year.
How InventorIA fits
InventorIA was built for exactly this CFO workflow. One platform that ties hardware, software, licenses, contracts, and renewals to people — with AI that answers questions in plain language ("which seats can I cut before quarter close?") instead of forcing you to learn a query language. Setup takes about an hour for most teams, and the free tier covers up to 10 users so you can validate before committing budget.
See your real IT spend in one dashboard
InventorIA gives finance and IT a single source of truth — assets, licenses, contracts, and renewals. Free to start.
Start free →