A note on this story: The 67% number comes from an aggregate of SaaS audits we've run with companies between 50 and 500 employees. The pattern is shockingly consistent. Your number might be 55%, might be 78%. It is almost never below 40%.
Picture your SaaS stack as a parking garage you pay for every month.
You think you have 40 cars in there.
When you actually walk through it, you find:
- 8 cars that have been parked since 2023 (active subscriptions for tools nobody uses anymore)
- 5 duplicate cars (Notion + Confluence + Coda all paid for, doing 80% the same job)
- 12 cars where the keys are still in the ignition (terminated employees still on the seat list)
- 3 abandoned cars (free trials that auto-converted to paid, no one noticed)
- 6 cars that one person used once last quarter ("seasonal" usage you're paying for year-round)
- 6 cars actually being used daily
You're paying for 40 cars. You're using 6.
That's the SaaS waste pattern. And it's not a 5% problem. It's a 67% problem.
Why is it 67%? Three structural reasons.
Reason 1: Seats are bought in bulk; usage is per-individual
When IT buys "Slack Standard, 200 seats", they pay $200 × $7.25 = $1,450/month. But the actual user count fluctuates with hires, terminations, contractors, vendors, agency creatives — and nobody decrements the seat count when those people leave. A 200-person company often has 140 active Slack users and 210 paid seats because of scope creep.
Reason 2: There's no "natural decommission" event
When you offboard an employee, IT revokes their Okta access. But Okta is one of 40 tools. The 39 tools where the offboarded employee had a direct login (because the team signed up to it before SSO was enforced) keep billing for that ghost user. Forever.
Reason 3: "Free trial → paid" is a one-way ratchet
A team signs up for Notion in March. Pays for "Plus" 5 months later when they hit the page limit. Migrates to Confluence in October. Nobody cancels Notion because it's still useful for that one wiki page nobody can be bothered to migrate. Notion bills $8/user/month. Forever.
These three patterns layer on top of each other. The result is the 67% waste figure.
The 5-step audit system
Here's the exact playbook we run with customers. Total time: 2 days for a 100-person company, 5 days for 500.
Step 1: Inventory every paid SaaS subscription
Pull the list from:
- Your accounting system (search vendor names: Notion, Slack, Figma, Linear, Atlassian, etc.)
- Your IT spend dashboard (if you have one)
- Your SSO provider (Okta, Azure AD app catalogue)
- Your shadow IT scan (Vendr, Productiv, BetterCloud, or just the credit-card statement)
Goal: a single CSV with vendor, plan, seats_paid, monthly_cost, contract_renewal_date.
A 200-person company typically lands on 35–50 line items. The first surprise: 8–12 of them you didn't know existed.
Step 2: Find the inactive users (the easy 25% win)
For each subscription, pull the user activity report from the SaaS admin panel. Most tools have one. The columns you care about:
- Last active date (not "last login" — last action)
- Total actions in last 30 days
The rule we apply: if a user has < 3 actions in the last 30 days, they're not actively using it.
| Tool | Inactive users (typical) | Monthly recovery |
|---|
| Slack Standard | 18% | $200–$500 |
| Linear | 22% | $150–$300 |
| Notion Team | 28% | $300–$600 |
| Figma Pro | 35% (designers rotate) | $400–$800 |
| Confluence | 40% (documentation tools always over-seat) | $500–$1,000 |
| LucidChart | 48% (used quarterly, not monthly) | $300–$600 |
Pull the inactive-user list, run it past the team owner, downgrade or remove. Typical recovery for a 200-person company: $2K–$4K/month.
Step 3: Find the duplicate tools (the 20% win)
Make a 2-column matrix:
| Job-to-be-done | Tools we pay for |
|---|
| Wiki / docs | Notion, Confluence, Coda |
| Diagramming | LucidChart, Miro, FigJam, Whimsical |
| Project mgmt | Linear, Asana, Jira, ClickUp |
| Time tracking | Harvest, Toggl, Clockify |
| Forms | Typeform, Google Forms, Tally, Jotform |
| Customer feedback | Hotjar, FullStory, LogRocket |
If a single row has more than one paid entry, you have a duplicate-tool problem. Either:
- Pick one, sunset the others (3-month migration, then cancel)
- Right-size each (the unused tool drops to free tier, the active one keeps paid)
The catch: people will scream. The way to defuse it: "We're not removing Notion — we're moving it to free tier. The 5 people who use it daily can still use it."
Typical recovery for a 200-person company: $1K–$3K/month.
Step 4: Find the zombie subscriptions (the 15% win)
These are subscriptions where:
- Usage is zero for 60+ days
- The "owner" listed in your accounting system has left the company
- The renewal is on auto-pay with no review
Pull a list of every SaaS line item where last activity was > 60 days ago. Cross-reference with your terminated-employee list. You will find 4–8 of these in any company over 100 people.
These are pure recovery. Typical recovery for a 200-person company: $500–$1,500/month.
Step 5: Cap the bleed — the 3 governance changes
Even if you recover $4K/month today, the waste rebuilds. Here are the 3 rules to install:
- Quarterly seat review — every tool > $500/month gets a 15-min review with the owner. Inactive users get downgraded.
- No new SaaS without an SSO requirement — every new tool must integrate with Okta/Azure AD. If it doesn't, it doesn't get bought. This eliminates the "ghost employee" problem at the source.
- Single point of accountability per category — one product owner per "job to be done" (wiki, project mgmt, etc.). They're responsible for picking ONE tool and migrating away from the others.
Without these three rules, you'll be doing this audit again in 12 months and finding the same 67%.
What does this look like in dollars?
Here's the typical recovery for a 200-person company spending $45K/month on SaaS (the median we see at this size):
| Source | Typical recovery |
|---|
| Inactive users | $3,000/mo |
| Duplicate tools (consolidation) | $1,500/mo |
| Zombie subscriptions | $700/mo |
| Right-sizing plans | $400/mo |
| Total | $5,600/mo (~$67K/year) |
That's 12.4% of total SaaS spend. Done in 5 days. With nothing more than CSV exports, an admin login, and a willingness to have a few uncomfortable conversations about Notion vs Confluence.
The hardest part isn't finding the waste
The hardest part is getting people to admit they didn't know about it.
Every SaaS audit I've run has at least one moment where the head of Engineering or Marketing says some version of: "How did we end up with three diagram tools? Who approved that?"
The answer is always the same: nobody approved it. Nobody disapproved it. It just happened.
That's not a failure of the team. It's a failure of the system. SaaS waste isn't a leadership problem; it's a default-on, no-friction, auto-renew infrastructure problem.
The fix is either:
- Manual: the 5-step audit above, run quarterly
- Automated: a tool that monitors usage, flags duplicates, and surfaces inactive users continuously
CARTIE AI's SaaS Detective does this automatically — connect your SaaS admin tokens, get the inactive-user list, the duplicate-category report, and the zombie-subscription scan in 10 minutes. The first scan typically pays for the whole year of CARTIE AI in the first 30 days.
But honestly? Even if you don't use a tool — just run the 5 steps above. Block off a Friday afternoon. You will find at least $3K/month of waste in any company over 100 people. Promise.
Now go reclaim your parking spots. 🥃