PE-backed portfolio companies accumulate software the same way a garage accumulates tools. Do not “cut SaaS.” Build the inventory, classify every app by business purpose, owner, users, cost, contract date, utilization, overlap, and integration dependency. Then classify: keep, renegotiate, consolidate, eliminate, or investigate. A first pass usually eliminates 12 to 20 percent of spend inside 90 days without breaking anything.
Every PE-backed portfolio company I have walked into has a software mess. Sometimes it is inherited from three acquisitions in six years. Sometimes it is department-level buying with no central approval. Sometimes it is a founder-era portfolio of “we bought that in 2019 and nobody remembers why.” Sponsors know the mess exists. They want a CFO who can quantify it and take it down without breaking operations. This is how.
Why software rationalization is different from other cost work
Cutting software is not like cutting temp labor or renegotiating parcel. Every app touches an operational workflow. Every kill decision has a switching cost. And every “we only pay $18K a year for that” is easy to ignore individually and worth $2M in aggregate.
The AICPA’s 2025 Finance Function Benchmarking Survey noted that mid-market companies with revenue between $50M and $500M average 47 to 92 distinct SaaS applications. Of those, 15 to 22 percent have functional overlap with another app in the same company, and 8 to 12 percent are barely used at all. That is the shape of the opportunity.
Build the inventory first
You cannot rationalize what you cannot see. Step one is a complete app inventory. Not “the list IT knows about.” The full list, including departmental shadow IT.
Sources to combine:
- AP vendor master, filtered for software vendors (12 months)
- Credit card feed, filtered by MCC 5734, 5817, 7372 (software categories)
- SSO logs (Okta, Azure AD, Google Workspace) for anything auth’d
- IT asset management tool if you have one
- Department heads asked directly: “list every subscription your team pays for”
You will find 30 to 50 percent more apps than IT thinks exist. That is normal. That is the point.
The app inventory template
| Field | What it captures |
|---|---|
| App name | Vendor, product name |
| Business purpose | One sentence. What breaks if this goes away. |
| Business owner | Named person, not “IT” |
| Users (provisioned) | Seats paid for |
| Users (active last 30d) | Actual usage, from SSO or admin console |
| Annual cost | Total contracted spend, all-in |
| Contract end date | Auto-renew flag if any |
| Utilization | Active / provisioned as percent |
| Overlapping function | Any other app in the inventory doing the same job |
| Integration dependencies | What upstream / downstream systems depend on it |
| Switching difficulty | Low / Medium / High |
| Strategic importance | Core / Support / Nice-to-have |
This is not a two-week exercise. Budget 4 to 6 weeks for a clean inventory at a mid-market company. Assign one person, do not commit-tee it.
The classification matrix
Once the inventory is clean, every app gets a decision:
| Classification | Trigger | Action |
|---|---|---|
| Keep | Core function, high utilization, no better alternative | Renew at negotiated rate |
| Renegotiate | Kept, but overpaying, or seats overprovisioned | Right-size seats + renegotiate rate at renewal |
| Consolidate | Functional overlap with another app already owned | Sunset one, standardize on the other |
| Eliminate | Low utilization, no critical dependency, no business owner defends it | Cancel at contract end |
| Investigate | Cannot determine what it does, or who uses it | Assign owner, 30-day investigation, revisit |
The “Investigate” bucket is the one everyone forgets. About 5 to 8 percent of the app list ends up here on the first pass. Investigate first, then classify. Do not eliminate what you do not understand.
The 5 categories where you will always find 20 percent savings
1. Video conferencing and collaboration
Nearly every company owns Zoom, Teams, Google Meet, and often Webex from a legacy division. Consolidate to two at most. One primary, one for external calls with clients on the other platform.
2. Document management and e-signature
DocuSign, Adobe Sign, HelloSign, and PandaDoc all show up in the same inventory. Storage split between SharePoint, Google Drive, Box, and Dropbox. Pick one signature vendor, pick one storage vendor.
3. Project management and productivity
Asana, Monday, Jira, Trello, ClickUp, Airtable. Every department picks their favorite. Consolidate by function: one for engineering, one for the rest of the company. Two, not six.
4. Marketing tech stack
Marketing accumulates faster than any other function. Email tool, CRM, analytics, landing page builder, ABM tool, event tool, SEO tool, social scheduling. Some of these overlap materially. Marketing ops rarely wants to give any up. Ask them to defend each in writing.
5. Individual developer and analyst tools
A single engineer with a corporate credit card and no policy will have 8 individual subscriptions in a year. Sometimes duplicated across teammates. Consolidate to team licenses, or eliminate.
How to actually kill a subscription
The failure mode is not the decision, it is the execution. To eliminate an app:
- Set a sunset date at least 45 days out
- Notify the business owner and every provisioned user
- Migrate data or workflow to the replacement app
- Update SSO to block access on sunset date
- Cancel with the vendor in writing before the auto-renewal window
- Confirm on the credit card feed 60 days later that the charge stopped
Step 6 is the one most companies skip. Look at the actual bill. Cancellations get lost. Confirm the money stopped moving.
Where AI tools fit
The AI tool stack is the newest software mess. Every seat of ChatGPT, Copilot, Claude, and Gemini, plus a departmental Custom GPT here, a Copilot Agent there, a Claude Project somewhere else. The same rationalization framework applies. Pick a primary AI assistant, consolidate seats, and set an enterprise agreement instead of individual credit-card buys. See the piece on Custom GPT vs Claude Project vs Copilot Agent for the decision framework.
How this shows up in the board pack
Software rationalization is a specific line item in the operational value creation section of the board pack. Report contracted savings, realized savings, and vendor count reduction. Sponsors like this because it is unambiguous, it hits the P&L, and it survives QoE at exit. See the PE-backed CFO board reporting package for the format.
What survives at exit
Real software savings, documented against a controlled baseline, with signed vendor cancellations, survives buy-side QoE. Aspirational software savings (“we plan to consolidate to Vendor X in year 2”) does not. See add-backs and QoE for what actually holds up.
Push back on this.
Every operator’s situation is a little different. If you run this differently, disagree with the methodology, or think we got something wrong, tell us. We publish the best counter-approaches on our Reader Contributions page, credited or anonymous, your call. Email hello@thepragmaticcfo.com.
FAQ
Who should own the app inventory?
IT does the mechanics. Finance owns the outcome. IT knows what is provisioned. Finance knows what is paid for. Both are needed.
What about tools bought on personal credit cards and reimbursed?
Ban the practice. Move everything to a corporate card with vendor coding, or a central procurement request. Reimbursed personal-card software is where shadow IT hides.
How do I convince a department head to give up a tool?
Show utilization data. If 3 people use it out of 25 provisioned seats, the conversation writes itself. If utilization is genuinely high and the tool is defended, keep it. Not everything is a cut.
Do we need a SaaS management platform?
Not for the first pass. A clean spreadsheet gets you 80 percent of the value. If the app list is over 150 or you are running M&A integrations quarterly, a SaaS management platform pays back.
What is the biggest failure mode?
Killing a tool that had a critical downstream dependency nobody documented. The integration dependencies field in the inventory exists to prevent this. Do not skip it.
Sources
- AICPA and CIMA, “Finance Function Benchmarking Survey,” 2025
- Grant Thornton, “Middle Market Report: Technology Spend,” 2025
- AlixPartners, “Digital Transformation in PE Portfolio Companies,” 2025
- Bain & Company, “Global Private Equity Report,” 2026
Written by The Pragmatic CFO. 15+ years running FP&A and finance operations across PE-backed portfolio companies.
The Pragmatic CFO 