The Martech Stack Audit Most Teams Run Backwards (Start at Renewal Dates)

A martech stack audit is a tool-by-tool review of what your marketing team pays for, who actually logs in, and what would break if each tool disappeared. Gartner's 2025 survey puts martech utilization at 49% of capabilities, but the bigger savings usually sit in idle seats and duplicate tools, and they can only be captured in the 60 days before each renewal date.

So the order matters. Most audit templates start with a giant spreadsheet of features and end with a slide nobody acts on. I think that's backwards. The money in a martech audit is locked behind contract dates, and if you discover a dead tool three weeks after it auto-renewed for another year, you found the waste but you can't touch it until next year.

This guide is the version I'd actually run: renewal calendar first, seats second, overlap third, and features last (if at all).

First, two "utilization" numbers that measure different things

You'll see two headline stats quoted in almost every martech audit article, often in the same paragraph, as if they were the same thing. They aren't.

The first is Gartner's. According to Braze's summary of the 2025 Gartner Marketing Technology Survey, marketers actively use about 49% of their martech capabilities, and martech eats roughly 22% of the total marketing budget. Gartner's own martech page frames that 49% as a risk that's forcing CMOs into hard decisions. Worth knowing: this is self-reported by marketing leaders. It's a feature-depth number. "We own the platform and use half of what it can do."

The second is seat utilization. Productiv measured how many SaaS seat licenses had any usage within a 90-day window and landed on 47% in 2023, as Scott Brinker covered on chiefmartec. Zylo's 2026 SaaS Management Index lands in the same neighborhood: the average organization uses 54% of its SaaS licenses, and even companies with 1 to 500 employees waste an estimated $3.8M a year on unused ones. That's a login number. "We pay for this seat and nobody opened the tool in three months."

Why does the distinction matter? Because the fixes are completely different. Low capability utilization is a training and process problem, and cutting the tool often makes it worse. Low seat utilization is a procurement problem, and it's almost pure cash you can get back at renewal. If you mix them up, you end up canceling the platform your team half-uses and keeping the 14 idle seats on the tool nobody touches.

To be fair to the vendors, Brinker makes a good counterpoint in the same piece: enterprise licenses are often priced as a discounted bundle that already assumes partial usage, so "unused seat" doesn't always equal "wasted dollar." That's true at enterprise scale. At a 20-person marketing team buying seats on a credit card, it mostly isn't.

Step one is a calendar, and it takes about an hour

Before you touch a feature list, build a renewal calendar. Every tool, one row:

• Tool name and the job it does (in five words or fewer)
• Annual cost, including add-ons and overage fees
• Billing owner (the person whose card or PO it's on)
• Renewal date and notice period (30, 60, or 90 days is common on annual contracts)
• Seats purchased vs seats with a login in the last 90 days

Pull the list from the finance ledger and corporate card exports, not from memory. MarTech's stack checklist makes the same point about including shadow tools, and notes that this inventory step alone tends to surprise most leaders. From what I've seen in published audit write-ups, the surprise is rarely one giant tool. It's the long tail of $49/month subscriptions that three different people expensed for one project in 2024.

Then sort the sheet by notice deadline, not by cost. Anything with a cancellation deadline in the next 60 days goes to the top of the review. A $30K platform renewing in nine months can wait. A $6K tool whose notice window closes in 18 days cannot.

Benchmark to aim for: Zylo's data suggests the best-run organizations get license utilization to 90% or higher, and its advice is to start with anything under 50% before touching the rest. For a marketing team, I'd use those exact thresholds on the seat column. Under 50%: rightsize at the next renewal, no debate. Between 50% and 90%: ask the tool owner to justify each idle seat by name.

The 14-day freeze beats every scoring rubric

Once seats are sorted, you'll have a list of tools that are technically in use but feel optional. This is where most audits stall, because every tool has one champion who swears it's critical.

The MarTech piece suggests a simple test: for each tool, ask what would actually break if you canceled it tomorrow, and if nothing breaks, you don't need it. I like the question. I don't trust the answers, though, because people consistently overestimate how much they depend on their own tools.

So test it instead. Pick the three to five tools where nobody can name a concrete breakage, remove access (or pause the integration) for 14 days, and keep a shared doc where anyone can log what they couldn't do. You're looking for one of three outcomes:

1. Silence. Nobody noticed. Cancel at the next notice deadline.
2. Annoyance. Someone wanted a report and pulled it from somewhere else in 10 minutes. That's overlap. Cancel and document the replacement path.
3. Real breakage. A form stopped syncing, a nurture stopped firing, a client report went blank. Restore it immediately and move the tool to the "keep" column with an owner attached.

One caveat that seems to catch teams out: check what's quietly wired into the tool before you pause it. Landing pages built on an old page builder, webhook triggers, tracking scripts that still fire on live pages. Pausing access is reversible. Letting a page builder lapse while 40 live landing pages still sit on its domain is a much worse afternoon.

Anyway, the freeze takes two weeks and replaces about three meetings of "but what if we need it." Worth it.

Overlap is where consolidation pays, and where it bites

After seats and the freeze, group what's left by job, not by vendor category. Email sending, social scheduling, reporting dashboards, landing pages, enrichment, call recording. Brinker's data shows the average marketing org runs around 75 apps and that SaaS stacks overall grew 32% over two years despite all the consolidation talk, so you'll almost certainly find two or three tools doing each of a few jobs.

The obvious move is to merge them into whichever platform already has the most data. The less obvious part: consolidation trades a sprawl problem for a lock-in problem. Every tool you fold into one vendor makes that vendor's renewal harder to walk away from. We wrote about this when Salesforce paid $5.66B for Braze and thousands of customers suddenly had their exit costs reshaped by someone else's M&A strategy.

My rule of thumb, which I'm sure some ops leads will disagree with: consolidate the jobs that share customer data (email, SMS, segmentation, CDP-ish stuff), and keep the jobs that don't (design, scheduling, call recording) on cheap, swappable point tools. You want integration where data has to flow, and you want the option to leave everywhere else.

Braze's own consolidation advice has one line I'd underline even though they obviously sell a platform: map identity and data before you migrate anything, roll out one team or channel at a time, and keep the old tool alive until the new one is proven. Skipping that last step is how teams end up with duplicate contact records and a nurture sequence that emails the same person twice.

Features last, and only for the tools you're keeping

Here's where Gartner's 49% number finally becomes useful. Once you know which tools survive, look at capability depth on the five or six most expensive ones. MarTech's checklist treats using under 30% of the features you pay for as a red flag, which seems like a reasonable floor.

But the fix for low feature usage on a tool you're keeping is almost never "buy another tool." It's usually one of these:

• A feature you're paying for that duplicates a separate subscription (lots of email platforms now include landing pages, lots of CRMs include basic attribution). Turn it on, then cancel the separate tool at renewal.
• A tier you outgrew downward. Plenty of teams sit on an enterprise plan for one feature they used during a 2023 launch.
• Training debt. If onboarding a new hire onto your stack takes two weeks or more, MarTech flags that as its own warning sign. An hour-long internal walkthrough on the three features people keep asking about tends to do more than any new platform.

A quick sanity check on measurement tools specifically: an analytics or attribution tool you're "keeping" should be the one that answers the question finance actually asks. If three tools each report a different ROAS, the thing to fix first is platform ROAS accuracy, ideally before you renew any of them. Same goes for the newer AI visibility trackers; at least one study found ChatGPT's API misses most of the pages ChatGPT actually cites, so it's worth checking whether yours measures the real thing before signing another annual contract.

Keep the stack from growing back

The awkward truth about martech audits is that the stack regrows. Brinker's 32% two-year growth figure happened during a period when everyone said they were consolidating. One audit fixes this year. A purchase rule fixes next year.

The one I'd put in place is borrowed almost directly from the MarTech piece: no new tool without a written problem statement, a check that nothing you already pay for does the job, a small pilot, and a 90-day check-in on the calendar before the trial converts to annual. Add one line of my own: every tool gets a named owner who is on the hook at renewal. Tools without an owner are the ones that renew silently.

And run the whole thing on a schedule. MarTech suggests quarterly, or twice a year at minimum. Honestly, quarterly feels like a lot for a small team. Twice a year, timed so each review lands 60 to 90 days before your biggest renewals, is probably the realistic version.

Martech stack audit FAQ

How often should you audit your martech stack?

Twice a year at minimum, quarterly if you have more than 30 tools or a dedicated marketing ops person. Time each review 60 to 90 days ahead of your largest renewal dates so you can still give notice.

What is a good martech utilization rate?

For seats, Zylo's benchmark for top performers is 90% or higher, with anything under 50% flagged for immediate rightsizing. For capabilities, Gartner's 2025 average is 49%, and using under 30% of a tool's features is a common red flag.

Should you consolidate into one marketing platform?

Partly. Consolidate tools that share customer data so records stay consistent. Keep tools that don't share data (design, scheduling, recording) on cheaper point solutions so you can still credibly walk away at renewal.

Prediction, for what it's worth: I'd bet that at least one in five tools on the average marketing team's renewal calendar fails the 14-day freeze with total silence. The audit isn't hard. The hard part is doing it before the renewal date instead of after.