If you asked a founder "what's the highest-ROI 90 minutes you can spend this quarter?" you'd get a long list of plausible answers. Customer calls. A pricing experiment. A content sprint. A hiring decision. Almost nobody would say "auditing your SaaS stack." Almost everyone is wrong.

A proper quarterly audit recovers, on average, 25-35% of SaaS spend - immediately, in the first pass. For a team spending £3,000/month on tools, that's £9,000-12,000 per year, reclaimed in 90 minutes of work. The hourly rate on that one habit is £6,000. I don't know a founder whose other work hits that rate.

And yet nobody does it. Or rather: founders say they do it, mean to do it, put it on the quarterly list, and quietly let it slip every quarter until a funding round forces a finance review and the exercise happens in panic mode. This essay is about why that's irrational, and why treating the audit as a ritual rather than a project is how you actually make it stick.

Why audits don't happen

It's not that founders don't know the audit is useful. They do. The objections I hear are almost all variations on the same pattern:

"It's boring." Correct. Auditing subscriptions is tedious. The work is clerical, not strategic. Most founders optimise their schedule around high-agency, creative work and push clerical tasks into the margins, where they get dropped. Understandable, but expensive.

"We don't have time." Usually the person saying this is the same one who will spend three hours next week researching a new AI tool they haven't decided to buy. The time exists. The priority doesn't.

"We'll do it at the next board review." Board reviews are the worst moment to audit. You're defending numbers, not questioning them. You'll skim the line items, not interrogate them. By the time a board-level person is asking why your SaaS spend is £X, you're already past the moment where a clean audit would have been useful.

"We'd rather grow than cut." False binary. The money you recover from the audit goes straight back into growth - specifically, into the one or two experiments you haven't been able to afford. Cutting £12k/year from bloat is equivalent to one more month of runway, or a part-time contractor, or three months of a paid-ads experiment.

The deeper truth: audits don't happen because they feel like housekeeping, and founders don't think of themselves as housekeepers. That self-image is expensive to maintain.

30%
The typical spend cut recovered in the first audit of a mid-stage startup's stack, without touching any tool the team actively uses.

Ritual vs project

The most common failure mode is treating the audit as a one-off project. "We'll do a big stack review, get things in order, and then we're done." You do the review. You cancel the obvious waste. You feel virtuous. Six months later, you've added a dozen new tools, forgotten three trials that auto-renewed, and the stack is back to baseline.

The audit isn't a project. It's a ritual. The distinction matters.

A project has a completion date. Once you're done, you move on, and entropy takes over. A ritual has a recurring schedule. It re-runs at intervals. It keeps the system in a narrow band around the steady state you want. The difference in outcomes, over years, is enormous.

Every stack I've audited multiple times shows the same pattern: the first audit finds 30% cuttable, the second finds 10-15%, the third finds 5-10%, and from there it stays stable - as long as the audits continue. Stop auditing, and within two years you're back to 30%. The ritual isn't doing new work each time; it's maintaining the discipline that keeps the stack lean in the first place.

Audits don't clean the stack. The expectation of audits does.

The right cadence

Quarterly is the right cadence for most teams. The logic:

Monthly is too often. Tools need at least six weeks of real use to evaluate whether they're earning their place. Monthly audits push decisions you can't actually make yet, and create audit fatigue. The team tunes it out.

Annually is too rare. Auto-renewals hit yearly, trials extend, shadow-IT accumulates. A year is long enough for the stack to drift past the point where the audit is comfortable. Annual audits become "clean up the mess" exercises, not "maintain steady state" exercises.

Quarterly hits the sweet spot. Enough time to evaluate whether a new tool is earning its place. Short enough that drift is limited. Tied to natural business rhythms (quarterly reviews, planning cycles). Four times a year is enough to be a ritual, not so much it becomes a chore.

Calendar it explicitly. "First Monday of the quarter, 10am-11:30am, SaaS audit, company credit card + password manager + billing." Recurring meeting. Non-negotiable. Treat it like payroll or tax filing - operationally boring, strategically non-optional.

What a good audit actually produces

A good audit isn't just a list of cancellations. It produces four outputs, and the cancellations are the smallest.

A decision log. For every tool kept, a one-line note on why. "Linear - core engineering tracker, load-bearing, owner: Alex." For every tool killed, a one-line reason. "GA4 - replaced by PostHog + Plausible, not opened in 90 days." The log is the continuity between audits. Next quarter, you read last quarter's decisions and see whether they still hold.

A renegotiation list. Tools worth keeping but worth renegotiating at renewal. Enterprise SaaS almost always has 10-20% discount room if you ask. The audit is when you build the list so the renewal date doesn't sneak up on you.

A consolidation queue. Categories where you're paying for two tools doing the same job. The tool you're keeping, the tool you're migrating off, and the owner of the consolidation work. Our switch guides cover the common moves.

A red-flag watch list. Tools you're worried about - a stagnant product, a team that's been acquired, pricing moving in the wrong direction. These don't get cancelled yet. They get monitored, and moved to the kill list if the signals keep pointing the wrong way. See Dead Pool for tools we're watching broadly.

The compounding effect

One audit saves 30%. Two audits keep you at 35-40% under baseline. Five audits, spread across 18 months, and you're in a different operating reality entirely.

More importantly: teams that audit quarterly start thinking differently about new tool adoption. Every potential signup is evaluated against "will I be comfortable defending this in the next audit?" The question alone stops most bad additions. New tools have to justify their seat at the table before they get a credit card.

This is the shift that separates lean operators from bloated ones. Not an "initial stack decision" that was smarter. A running process that keeps asking. The stack of a founder who audits quarterly, over time, looks fundamentally different from the stack of a founder who doesn't - not because one bought better tools, but because one has a working feedback loop and the other doesn't.

The counter-argument

The most common pushback I hear: "I don't want my team thinking about cost all the time. I want them thinking about output."

Understandable, and wrong. The alternative to thinking about cost deliberately once a quarter is thinking about cost anxiously, constantly, whenever a credit-card statement shows up. Quarterly audits bound the cognitive load; they don't increase it. Teams that audit quarterly talk about cost less, not more, because they know it's handled. Teams that don't audit spend every board meeting defending line items.

The meta-principle: deliberate, scheduled, bounded work on a thing prevents diffuse, anxious, unscheduled work on the same thing. This is true for audits. It's true for hiring. It's true for strategy. Schedule the hard questions and they stop hijacking your week.

How to start

If you haven't run an audit in the last quarter - or ever - here's the entry path:

  1. Block 90 minutes on the calendar this week.
  2. Read our 90-minute audit framework.
  3. Run it exactly as written. Don't modify, don't skip steps.
  4. Save the decision log. Create a recurring quarterly calendar event.
  5. Don't miss the next one.

That's it. The first audit is uncomfortable because the stack is messier than you thought. The second is satisfying because it confirms you've held the line. By the fourth, the ritual is doing the work and you barely notice.

Want the audit run for you?

If the stack is big, the data is scattered, or you just don't want to do it yourself - I'll run a thorough audit, produce the decision log, and hand back a concrete cut list.

Get a stack audit →
Clinton Feyisitan
Clinton Feyisitan Founder, Fewer Tools. Audits, consolidates, and rebuilds bloated SaaS stacks for early-stage companies.

Further reading