Your Zapier/Make Automations Are Probably Broken (Here's How to Check)

The business owner who never checks their automation logs is paying full price for a system that's failing silently. Here's the exact audit process — and what to look for.

You built the automation. It worked the first time you tested it. You moved on. That was 8 months ago.

In those 8 months, an API connection expired. A third-party app updated its schema. A filter condition stopped being met. A webhook URL changed. And your automation has been failing silently — or worse, completing with wrong data — while you assumed it was running fine.

This isn't an edge case. It's the default state of most automation stacks that haven't been audited.

How Widespread Is This Problem?

When we audit a client's automation stack, we typically find:

• 15–30% of Zaps/scenarios have recent error history that nobody noticed
• Another 20–30% run "successfully" but produce empty or incorrect outputs — tasks completed with no real action taken
• 10–20% are turned off — often because something broke and it was easier to disable than fix
• Only 40–60% are actually working correctly end-to-end

The bill keeps charging for 100% of the capacity. The actual value delivered is often half that.

Step 1: The Error Log Audit (20 Minutes)

This is the fastest way to find obvious failures.

In Zapier:

1. Go to Zap History in your dashboard
2. Filter by status: select Error and Stopped
3. Set date range to last 30 days
4. Download or scroll through everything that failed
5. Group failures by Zap name — some will have recurring patterns

Any Zap with more than 2–3 errors in 30 days needs attention. Any Zap with error rates above 10% of total runs is broken in a meaningful way.

In Make (formerly Integromat):

1. Open each active scenario
2. Check the History tab for error executions (shown in red)
3. Look at the execution count vs. error count ratio
4. Make's error panel shows which module failed and what the error message was

Make's error messages are generally more detailed than Zapier's — use them. A "connection failed" message usually means an API key expired or permissions changed. A "field not found" message means an app schema changed.

Step 2: The Silent Failure Audit (45 Minutes)

This one takes more work because the automations look fine — no errors — but aren't delivering value.

Pick your 10 most business-critical automations. For each one:

1. Identify the expected output. What should happen when this Zap fires? Be specific: "A contact record is created in HubSpot with first name, last name, email, and source set to 'Website Form.'"
2. Find a recent execution in the history. Click into it and look at the actual data that was passed through each step.
3. Verify the output actually exists. Check the destination system. Does the CRM record exist? Does the spreadsheet row have data? Was the email actually delivered?

This process often reveals:

• Automations that fire correctly but write to the wrong destination (old Google Sheet, deleted HubSpot pipeline)
• Automations that pass empty fields because a form field was renamed
• Automations that duplicate records instead of updating them
• Automations where the final action is correct but an intermediate step is silently skipped

Step 3: The Connection Health Check

Every automation platform uses OAuth connections or API keys to talk to third-party apps. These connections expire, get revoked, or break when you change passwords.

In Zapier:

1. Go to Connected Accounts in your settings
2. Look for any accounts with warning indicators
3. For each critical app (Gmail, HubSpot, Slack, Airtable, etc.), click the connection and test it
4. Re-authenticate anything that fails the test

In Make:

1. Open Connections in the left sidebar
2. Look for connections with error states (red indicator)
3. Click each connection and verify it's still authorized

Stale connections are the #1 cause of sudden automation failures. Re-authenticating connections should be part of your quarterly maintenance routine.

Step 4: The Zombie Automation Cleanup

Zombie automations are: turned off, firing but doing nothing meaningful, or connected to systems you no longer use. They cost you in three ways:

1. Task/operation counts. Even automations that do nothing useful consume your monthly task quota.
2. Mental overhead. A list of 47 automations is harder to manage than a list of 20.
3. Confusion. When something breaks, a bloated automation stack makes root-cause diagnosis much harder.

Go through every automation and ask: "Is this connected to a workflow that still exists?" Archive or delete anything that isn't. A clean list of working automations is worth more than a long list of broken ones.

Step 5: Consolidate Where You're Running Both

If you're on both Zapier and Make, you're probably paying twice for overlapping capability. Pick one as your primary platform. Migrate critical workflows from the secondary to the primary. Cancel or downgrade the secondary.

The scenarios where you genuinely need both are rare. Usually it's historical — someone built some Zaps years ago, someone else started using Make, and now you have two platforms and nobody knows which is authoritative.

What to Do After the Audit

You'll end up with three lists:

• Fix now: Automations that are broken and business-critical
• Fix later / deprioritize: Automations that are broken but low-impact
• Archive/delete: Automations that shouldn't exist anymore

Work the "fix now" list first. Most fixes are straightforward: reconnect an OAuth, remap a field, update a filter condition. The ones that require rebuilding from scratch are typically the ones that were poorly designed to begin with — and those deserve a rebuild with proper testing.

If this feels like more than you want to dig into yourself, the AI Operator Audit covers your full automation stack — not just the subscriptions, but the actual workflows, what's working, what's broken, and what to prioritize.

Related: How Much Is Your AI Stack Wasting? · The 10-Point AI Audit Checklist · Free AI SOUL.md