Nick QuirkAsk an agency owner how many SaaS tools their team uses and you'll usually hear "maybe fifteen?" Then you pull twelve months of card statements and find 38 subscriptions, four of which do roughly the same thing, and one that nobody has logged into since a strategist who left in 2023 set it up.
That's tool fatigue. It isn't just a budget problem — it's a focus problem. And the fix is a proper SaaS stack audit: a structured, once-a-year pass that tells you exactly what you pay for, who actually uses it, and what can go.
What tool fatigue actually costs you
The subscription line is the smallest part of the bill. Three costs matter more:
- Context switching. Research on knowledge work consistently puts the recovery cost of an interruption somewhere north of 10 minutes. If a content strategist touches six tools a day just to find a brief, a deadline, and a comment thread, you're losing 30-45 minutes per person per day to navigation, not work.
- Data fragmentation. When time lives in Harvest, tasks live in Asana, deliverables live in Google Drive, and retainer scope lives in a spreadsheet, nobody can answer "are we profitable on this client?" without an hour of manual assembly. So nobody asks.
- Onboarding drag. Every extra tool adds roughly half a day to a new hire's ramp. At 40 subscriptions with 12 that are client-facing, you're spending a week teaching software instead of process.
A 20-person agency paying an average of $9 per user per month across 30 tools is spending about $5,400 a month, or $65,000 a year. That's a senior hire. But even if the money were free, the attention cost would still justify the audit.
Six signals your stack has fatigue
- Someone asks "where does this live?" more than once a week in Slack.
- You have two tools in the same category (two whiteboards, two doc tools, two time trackers) and can't articulate the split.
- A tool's Slack notifications are muted org-wide — a reliable sign it's been abandoned in practice but not in billing.
- Client reporting requires exporting from three systems into a deck.
- Your last three new tools were bought to solve a process problem, not a capability gap.
- Nobody knows who owns the account for at least one active subscription.
How to run a SaaS stack audit in five steps
1. Build the inventory from money, not memory
Do not start with a team survey. Start with 12 months of statements from every card and PayPal account, plus your accounting export filtered to "software" and "dues and subscriptions." Add anything billed annually — those are the ones people forget.
For each line, capture: tool name, monthly cost, billing cycle, renewal date, seat count, internal owner, and the category it belongs to. A spreadsheet is fine. Expect the list to be 40-60% longer than your mental model.
2. Get real usage data, not opinions
Most SaaS admin panels show last-login dates per seat. Pull them. The pattern you're hunting for is the 15% tool: a subscription where fewer than one in six seats logged in during the past 30 days.
Then run a two-week diary in Slack or a shared doc: every time someone opens a tool to complete a real task, they note the tool and the task. It's crude, but two weeks of that data will show you which tools carry actual workflows and which exist because someone likes them.
3. Score every tool on four axes
Rate each subscription 1-5 on:
- Usage depth — percentage of seats active weekly.
- Uniqueness — is this capability available in something you already pay for?
- Client visibility — do clients log in, receive output, or see it referenced? Client-facing tools have higher switching costs, so weigh them accordingly.
- Data gravity — how much historical data lives here that you'd need to migrate or lose? Time-tracking history and project archives score high; a screenshot tool scores low.
Anything scoring under 8 out of 20 is a cancellation candidate. Anything scoring 16+ is core infrastructure — leave it alone even if it's expensive.
4. Sort into keep, consolidate, downgrade, kill
Kill is easy: zero-login tools, duplicates with a clear winner, and anything a departed employee owned. Cancel before the next renewal and export the data first.
Downgrade is the most overlooked bucket. Agencies routinely pay enterprise tiers for features they use once a quarter. Check seat counts against actual headcount — if you're paying for 24 seats with 18 employees, you're donating $1,300 a year.
Consolidate is where the real gain sits, and where most audits go wrong. The rule: consolidate across the workflow, not the category. If your PM tool, time tracker, and retainer spreadsheet are three separate systems that all describe the same client engagement, merging them removes reconciliation work every single week. That's why we built PeakKR to combine projects, time, and retainer tracking in one place — the three data sets only mean something together. If you're evaluating options, a structured comparison of the major platforms is more useful than another free trial: this roundup of PM tools built for agency work is a reasonable starting point.
5. Migrate with a freeze and a deadline
Pick a 30-day window. During it: no new tool purchases, one named owner per migration, and a hard cutoff date after which the old tool is read-only. Migrations that run "until people are ready" never finish — you end up paying for both systems for eight months and confusing the team.
Export historical data before you cancel. Time-tracking history in particular is worth keeping: it's the raw material for estimating SEO campaigns based on what work actually took, and you can't rebuild it once the account closes.
Where consolidation backfires
Three failure modes worth naming.
Consolidating creative work into generic tools. Designers and video editors have legitimate specialist needs. Forcing them into an all-in-one to save $50 a month buys you resentment and worse output.
Replacing a tool without replacing the habit. If people didn't update statuses in the old system, they won't in the new one. Fix the ritual — a 10-minute Monday review, a Friday time-entry deadline — before you fix the software.
Adding surveillance in the name of visibility. Consolidating time tracking into your PM tool is only a win if the team sees it as a planning input, not a monitoring device. The framing matters more than the feature set; there's good research on why knowledge workers resist time tracking and how to avoid triggering it.
Keeping it lean after the audit
Stacks re-bloat within 18 months unless you install two rules.
A one-page tool request. Anyone proposing a new subscription answers four questions: what workflow does this serve, what tool already in the stack partly does it, what's the annual cost at our headcount, and who owns it. Half the requests die at question two.
A renewal calendar. Put every renew
