Your SaaS Stack: The Cost Nobody Measures in Your Company

Open your corporate card statement and count how many software subscriptions show up there. Do not look at the amounts yet, just count. If you passed twelve and got stuck on some line thinking "what is this again?", your SaaS stack has already grown beyond what you can manage. And the problem is not the monthly bill. It is that nobody knows what is still turned on.

I am writing this because a UN report came out saying global warming will pass 1.5 degrees. Not "might pass". Will. The detail that got me was not the number, it was the logic of the document: they already assume the overshoot and now they discuss how to come back down later. It goes over, then it comes down. They call that overshoot.

I kept thinking that this is exactly what happens with software inside a company. Nobody decides to bloat. Each subscription, on its own, makes sense. The overshoot is the sum of small decisions nobody ever added up.

Why the SaaS stack grows without anyone deciding

The pattern is always the same, and I have seen it in a five person client and in a two hundred person client.

Marketing needs to schedule posts, so they subscribe to a 39 dollar tool. Sales wants a better CRM, so they subscribe to another one. Someone in finance tests a reconciliation app on a trial, forgets to cancel, and nine months later it is still being charged. The designer buys a stock photo plan on their own card and asks for reimbursement. None of that is wrong. Each person solved a real problem, right then, the fastest way they could.

But nobody was put in charge of looking at the whole. There is no "what tools do we have" meeting. There is the sales meeting, the product meeting, the month end meeting. The SaaS stack has no owner.

A client of mine, a logistics company with 40 employees, thought they spent around R$ 2,000 a month on software. We mapped everything: R$ 6,400. Three tools did the same thing in different teams. Two had not been opened in over a year. One had 22 paid licenses for 9 people who still worked there.

The cost that does not show up on the invoice

Money is the easy part to see. The other costs are worse.

Your data gets chopped up. Client registered in the CRM, quote in a spreadsheet, contract in a third system, billing in a fourth. When someone asks "how much has this client bought from us", the answer takes two days and comes out crooked. That is not a people problem, it is an architecture problem.

Your team copies and pastes. Every tool that does not talk to the others creates a human in the middle building a bridge. That person who exports a CSV every Monday and imports it somewhere else is not working. They are being an API made of flesh and bone.

Your risk surface grows. Every login is a password that can leak. Every integration is an access that stays active after the employee leaves. I have never seen a company revoke everything properly when someone left. Never.

A tool nobody removed is different from a tool somebody chose.

And there is the energy cost, which is where the story goes back to the UN report. Data centers consume electricity and water. That BI report that runs every hour and nobody has read in eight months burns real energy. It is small per company and it is a lot when you multiply it by every company in the world doing the same thing. I am not going to say that canceling subscriptions saves the planet, that would be dishonest. But stopping processes nobody uses is the easiest thing on this list, and it is free. You save money and the energy is a bonus.

Overshoot applies to software too

The interesting part of the report is the after. Passing 1.5 is not the end, it is a bad window that lasts decades until emissions fall and the number comes back down. It is expensive and it takes time, but you can come down.

Software works the same way, with the huge advantage that the time scale is yours. You do not need thirty years. You need one quarter and the willingness to cancel things somebody chose one day.

The blocker is rarely technical. It is political. Canceling the marketing tool feels like saying marketing got it wrong. That is not it. The tool made sense in 2023, when the company was a different size with a different process. The context changed. Nobody got it wrong.

How to run a SaaS audit in practice

You can do this without a consultant, without a project, without a committee. Block off one afternoon.

1. Pull 12 months of statements from the corporate card and the company account. Filter everything recurring. Include reimbursements, that is where the ghost software lives.
2. Build a dumb table with five columns: tool, monthly cost, who owns it, how many licenses, what it is for in one sentence. If you cannot write the sentence, you just found a cut candidate.
3. Check real usage. Almost every tool has an admin screen with last access per user. A license with no login in 60 days is a dead license.
4. Look for overlap. Two task tools. Three places to store files. Two chat channels. Pick one of each and kill the rest.
5. Cut in two rounds. First the obvious ones, nobody uses them and nobody complains. Then the negotiated ones, where you sit down with the team and decide what to migrate.
6. Create one simple intake rule. A new subscription needs a named owner and a review date. That is it. No five step approval process, otherwise people go back to subscribing quietly on their personal card.

The average cut I see lands between 25% and 40% of the spend, and almost always without anyone missing a thing. In that logistics company, it went from R$ 6,400 to R$ 3,900 in the first month. What was left started talking to each other through integrations, and the woman who exported spreadsheets on Monday went back to serving clients.

"But I need all of them"

Maybe you do. The audit is not a campaign to use less software, it is to know what you use. Some companies finish the mapping and cut two things, that is all. Great result too: now there is a list, with owners and dates.

The most common objection I hear is "migrating is a lot of work". It is. But compare it to the work that already happens every month in silence, that hour someone spends fixing data that came in wrong, the rework of a quote done twice. You already pay that cost, it just is not on any spreadsheet.

The second objection is "what if I need it later". You export before canceling. It takes fifteen minutes. Save the file. In two years doing this with clients, I had to restore exactly one tool.

The point

The UN report is about a system nobody controls alone, where each small decision seemed reasonable and the sum became a problem. Your SaaS stack is the same thing in miniature, with one difference that works in your favor: it fits in a spreadsheet and you are the one in charge of it.

Start with the count. Before optimizing, just knowing the number already changes the conversation.

If you want a hand mapping this out and connecting the tools that are left, that is part of what I do here.

Open your corporate card statement and count how many software subscriptions show up there. Just count. Do not look at the amounts yet.

A logistics client of mine, 40 employees, thought they spent R$ 2,000 a month. It was R$ 6,400.

Three tools did the same thing in different teams. Two had not been opened in over a year. One had 22 paid licenses for 9 people who still worked there.

Nobody messed up. Every subscription made sense the day it was signed. The problem is that nobody was in charge of looking at the whole.

We mapped everything in one afternoon and the spend dropped to R$ 3,900 in the first month. Nobody missed a thing.

If you got stuck on some line of the statement thinking "what is this again?", you already know where to start.

#SaaS #CostManagement #Technology #Automation #Productivity