Skip to content
Efidis
Efidis

Costs and decisionsUpdated 9 min read

Software subscription audit: what you're overpaying for

3D illustration of membership cards, a long invoice and lilac scissors cutting one of the cards.
In this article

An illustrative example. A 12-person company adds up what it pays for software every month and gets $1,005. When it checks who actually logs in to each tool, it finds it's paying for 43 seats for 12 people, there's a form builder nobody has opened in a year, and customer data lives in three different places. Nobody planned it that way: each tool came in to solve a specific problem at the time.

This audit is a way to see that whole list in one afternoon and decide with numbers what stays. You don't need a license management platform: a card statement and a spreadsheet are enough.

Why the total is always higher than you think

Subscriptions grow a little at a time. A $15-per-seat tool never goes through a budget meeting, gets paid on the card of whoever needs it and renews on its own. Two years later, nobody remembers who asked for it.

Data from the companies that track this points the same way. Zylo, which manages more than 40 million licenses for its customers, found that on average only 54% of purchased licenses get used, and that median SaaS spend is $9,455 per employee per year (Zylo, 2026). Business units, not IT, control 81% of that spend.

Smaller companies show the same pattern. Cledara, which analyzed spending at tech companies with fewer than 200 employees, estimates that for every 10 tools a company thinks it's using, 14 are actually running, and that companies with 100 to 200 employees waste 34% of their software budget (Cledara, 2024).

Both sources are vendors that sell subscription management software, and their customers are larger than a typical small business. Don't use their percentages as your number; use them as a reason to run your own.

One home per piece of data

The usual question is "which tool can we drop?" It's hard to answer, because every tool has someone defending it. There's a more useful question: where does each piece of data live?

Here's the rule: every piece of operational data should have one home. Customers live in one place, products and prices in another (or the same one), orders in one, inventory in one, invoices in one. When data has two homes, someone copies it from one to the other, and the two versions drift apart.

Apply that rule and three kinds of tools stand out:

  • Tools that hold data that already has a home. A spreadsheet of customers when you already pay for a CRM. A report that's just a copy of sales and inventory.
  • Bridge tools: they hold no data of their own and only exist to move data between two other tools. A connector that pushes orders from the spreadsheet into inventory, a weekly import. Every bridge is a sign that two homes should be one.
  • Tools holding data nobody looks at. Forms from an old campaign, a dashboard built for one meeting.

Tools that hold unique data that people use (invoicing, email, accounting) are rarely the problem. For those, the audit only checks how many seats you're paying for.

The audit in one afternoon

One afternoon is enough for the first version if you have the card statement at hand and admin access to the main tools. Pinning down real usage can take another week, but the big decision is already visible in the first draft.

  1. Pull the list from statements, not from memoryGo through 12 months of the company card, the cards of anyone who buys on their own and vendor invoices. Twelve months, because annual plans only show up once. Write down every recurring charge.
  2. Fill in the first columnsTool, monthly cost (annual plans divided by 12) and paid seats. If a tool bills by usage, use the average of the last three months.
  3. Measure real usageIn the admin panel, count who logged in during the last 30 days. If the tool doesn't show that, ask the team who opened it this week. What matters is the number of active people, not the accounts created.
  4. Write down what data it holdsOne or two words: customers, orders, inventory, invoices, tasks, documents. If it holds nothing of its own and only moves data, write "bridge." If it holds the same thing as another row, flag it.
  5. Decide using the four exitsFor each row, pick one of four decisions: keep, cut, merge or cancel. The criteria for each are below.

The template, with an example

This is the illustrative example of the 12-person company, with round numbers. Copy the columns and replace the rows with your own.

ToolMonthly costPaid seatsReal usage (30 days)Data it holdsDecision
CRM$300104Customers and dealsCut to 4 seats
Order spreadsheet$0 (included)66Orders and customers (duplicate)Merge
Inventory app$15033Products and stock levelsMerge
Invoicing$8022InvoicesKeep
Automation connector$7011None (bridge: spreadsheet → inventory)Merge
Form builder$4020Orders from an old campaignCancel
Task manager$120125TasksCut to 5 seats
E-signature$4531ContractsKeep
Reporting tool$20042Copy of sales and inventoryMerge
Total$1,0054312 people

Just by canceling and cutting, without changing anything about how the business runs, the example saves $290 a month: $40 from the form builder, $180 from the CRM (10 down to 4 seats at $30) and $70 from the task manager (12 down to 5 seats at $10). That's $3,480 a year for one afternoon of work.

The interesting part is the rows marked "Merge." Orders and customers live in both the spreadsheet and the CRM; inventory lives in its app and gets copied into the report; the connector only exists to move data between the spreadsheet and the inventory app. That's $420 a month in tools, plus the time of whoever does the copying, all around a single flow: order, inventory and report.

The four exits

  • Keep: it holds data that doesn't live anywhere else, people use it and paid seats match active users.
  • Cut: it gets used, but you're paying for more seats or a higher plan than you need. Drop seats down to active users and check whether anyone uses the higher plan's features.
  • Merge: it holds data that already has another home, or it's a bridge. Today's decision is choosing which of the two homes keeps the data; canceling comes later.
  • Cancel: nobody used it in 30 days and it holds nothing you need to look up. Export whatever it has before you cancel.

Work out the real cost per person

Each tool's per-seat price is misleading, because one person uses several. The number that's useful for comparisons is total software cost per active person:

StepWhat to calculateIn the example
AMonthly total for every tool$1,005
BPeople who use at least one tool12
Cost per person per monthA ÷ B$83.75
Cost per person per yearA ÷ B × 12$1,005
After canceling and cutting(A − savings) ÷ B$59.58 a month

With that number you can compare any alternative on the same basis: what it costs per person per month. If someone offers to replace five tools, compare their per-person price against what you pay today for those five, divided by the same people. Also add the hours spent copying data between them; they often outweigh the licenses. The software cost calculator does that sum for you, and the formula for the hours is in what manual data entry costs.

When merging tools isn't worth it

Fewer tools isn't always better. Don't merge if:

  • It's your accounting or tax and invoicing software. Those follow their own rules and change with regulations; the usual move is to leave them in their own system and connect them to operations.
  • The tool is great at something very specific and shares no data with the rest: design, email, video calls. There's nothing to merge.
  • Switching costs more than you'd save in two years, counting the migration and the team's time learning something new.
  • You just renewed an annual contract. Note the renewal date and decide a month before; canceling mid-contract rarely gets you money back.
  • Your team is three people and total spend is low. Cut what's extra and look again once you grow.

The goal of the audit is for every piece of data to have one home and for every paid seat to be a person who works with it. Having fewer tools is a side effect.

What to do with the results

Start with what carries no risk: cancel what nobody uses and cut seats. Put renewal dates on a calendar with a reminder a month ahead. That step alone usually pays for the afternoon several times over.

For the rows marked "Merge" there are three paths, from simplest to most complete. You can pick which of the two tools keeps the data and stop using the other one for it. You can connect the tools that stay, so data flows on its own and nobody copies it. Or you can bring the core of the operation (customers, orders, inventory, receivables and reporting) into a single custom app, and leave out what makes sense to keep separate, like accounting. The third path fits when several "Merge" rows revolve around the same flow, as in the example: what goes away are the bridges.

Repeat the audit every six months. The second time takes much less, because the table already exists and you only review what's new.

Sources

  1. Zylo, "70+ SaaS Statistics for 2026," with data from the 2026 SaaS Management Index (opens in a new tab)
  2. Zylo, "Announcing Zylo's 2026 SaaS Management Index," January 29, 2026 (opens in a new tab)
  3. Cledara, "The 2025 Software Spend Report," published 2024 (opens in a new tab)

Frequently asked questions

Pay only for what your team uses

Efidis builds a custom app for the part of your operation that passes data between tools today, and connects it to the ones that should stay. On the call we'll go over your list and tell you what to merge and what to leave.