The 2026 SaaS spend reset: 14 experts reveal where businesses waste money

SaaS waste is rarely caused by one enormous subscription. It is usually created by dozens of small decisions that nobody owns.

Buying software has become easier than managing it. Sales, marketing, support, analytics, automation and AI can each carry their own bill, and each bill looks small on its own.

To find out where the money actually leaks, we asked founders, CEOs, operators and technology leaders what they see inside real businesses. Fourteen of them shared what they found, what they cut, and the test they use before a renewal.

At a glance

Waste usually starts with duplicate tools

Edward Tian of GPTZero says the first thing an audit finds is redundancy. Two teams pay for two tools that do the same job, because each team solved its own problem in its own month. His fix is a rule rather than a report: a new subscription should replace something, not sit next to it.

Lily Stoyanov, a former CFO who now runs Transformify, sees the same pattern in seats. A company buys licences for a whole team and a handful of people log in. The renewal then arrives quietly, and nobody reviews it. Her test is short: if you switched the tool off today, would anyone notice inside a week?

Matthew Attanasio of Sign Customiser pushes that further. Review software by the workflow it supports, not one vendor at a time. A cheaper tool is not the better tool if nobody can say what job it does.

Siim Kostabi of Pageloot points at team-led signups. Someone adds a tool to fix one problem, the problem goes away, and the billing carries on for years. He treats software as a variable cost with a quarterly review, and gives every tool one owner who decides whether it stays.

The owner matters more than the invoice

David Salamon of eCopier Solutions says most small businesses simply have no central owner. Different people sign up with company cards, so no single person can see the whole bill. His fix does not need software at first. One person keeping one honest list will surface subscriptions nobody remembered.

Pavan Kumar Kamat of Panto AI calls the same problem unmanaged tool sprawl, and treats the stack like an inventory ledger: every tool gets an owner, an active seat count, one primary outcome and a renewal date. In one scale-up he advises, a quarterly review found 18 duplicate products and reclaimed 22% of spend through consolidation and seat cuts.

David Hunt, COO of Versys Media, reviews tools on usage, overlap, renewal date and business impact. One review turned up overlapping project, form and email products costing more than $18,000 a year. No single subscription looked outrageous, which is exactly why the waste survived so long.

Unused seats are only part of the problem

Rob Illidge of Vulse says companies keep paying for people who have left, because seats are never reconciled against the current staff list. He also warns about the opposite case: the tool everyone uses but has outgrown, sitting on a plan far more expensive than the work needs.

Abhishek of Testlify checks real usage before every renewal. A cheap invoice can still be waste if a small share of the seats are active.

• How many people actually opened it in the last 30 days?
• Which features do they use, and how often?
• What business outcome does that usage produce?
• Does the seat count match the current staff list?

AI is adding a new layer of spend

AI came up in almost every response, but not as a saving. Vendors are adding AI features to tools you already pay for and raising the price, while teams also buy standalone AI products. The result is paying twice for capability nobody asked for.

Cheap trials and overlapping assistants create spend before anyone has checked whether the work got better. Tim Akdemir of Aslan Intelligence gives every AI subscription a defined use case, an owner and a measurable result, which matters more as AI pricing moves from flat monthly fees to usage.

Usage-based billing is a different kind of risk

Jacob Perks of Thunkle gave the sharpest example in the set. His worry is not the unused subscription. It is uncontrolled usage. He reviewed an application where paid AI services were being called straight from the visitor browser using the company credentials, with no spending limit attached. The app looked fine, so the problem could have run unnoticed.

He also measured his own AI coding tool and found the underlying consumption would have been worth thousands of dollars at published rates. A flat subscription can hide enormous swings in what you actually consume.

• Put a hard spending cap on every usage-billed account.
• Keep an inventory of tools, accounts and credentials, including keys.
• Never expose paid API credentials to a browser client.
• Alert on the bill, not only on the renewal date.

The best test is surprisingly simple

From very different angles, the contributors kept landing on the same question: what would happen if you switched this tool off tomorrow?

Akdemir asks what output the tool produces, how often anyone uses that output, and what would break without it. If nobody can answer, pause the tool for 30 days and watch. Todd Saunders of Dalton Mills names the workflow the tool owns, the person accountable for it, and the system that depends on its data. Salamon puts it plainly: ask what breaks when you turn it off.

The point of the test is that it moves the conversation away from whether people like a product, and toward whether the company depends on it.

Audit workflows, not only software

Justin Lapier of Synctivate argues most teams look in the wrong place. Two platforms can do genuinely different jobs and still force staff to move data between them by hand. The subscriptions are visible. The hours spent exporting, pasting and reconciling are not.

Auditing the workflow can therefore beat cancelling a licence. Consolidation or a single integration often returns more than the seat you were arguing about.

That is also why cutting tool count is the wrong goal. A company can make itself slower by killing a tool that earns its keep. The goal is to remove software with no clear purpose, no owner and no measurable return.

The quarterly reset is becoming a habit

One recommendation repeated more than any other: stop treating software review as an occasional clean-up. A quarterly zero-based review asks every subscription to justify itself again, with usage checked well before the renewal and renewal dates sitting in a shared calendar with reminders.

Bogdan Nocholas of Vinspector AI now runs exactly that: each quarter, departments justify the tools they keep. The shared idea is simple. Do not wait until the software bill becomes a problem.

• Month 1: list every tool, owner, seat count and renewal date.
• Month 2: pull usage, flag overlap and plans you have outgrown.
• Month 3: cancel, downsize or renegotiate before the renewal lands.
• Every month: cap usage-billed accounts and review new signups.

What this article does not claim

These are practitioner accounts, not a statistical study. The savings quoted belong to the companies that reported them, and your stack will not behave like theirs. We have not audited the underlying invoices.

It is also not a case for a smaller stack. Some businesses legitimately run hundreds of tools. The test is whether each one has a job, an owner and a measurable return.

Finally, no tracker cancels software for you. Cancellation still happens in the vendor account, and a tracker only makes sure you get there before the renewal does.

The reset, in one line

The phrase "SaaS waste" is slightly misleading. The problem is not too much software. It is too many tools without clear ownership, measurable outcomes or regular review.

The approach these 14 people converge on is not "cancel subscriptions". It is: assign an owner, measure usage, understand the workflow, find the overlap, review before renewal, cap usage-based spend, then decide whether the tool deserves to stay.

The cheapest tool is not the one with the lowest monthly price. It is the one that can explain why it still exists.

The 14 contributors

Questions people ask

Keep reading

Recommended next