Most of your Zaps should stay exactly where they are.
Zapier is the best tool in the category for small, low-risk glue between two apps, and replacing a Zap that has worked for a year buys you a migration and nothing else.
The ones that should move are handling revenue-critical data, costing real money per task, or failing without telling anyone. We will tell you which yours are.
Four jobs we would not move off it.
Two-app glue
Form to Slack, calendar to sheet. One trigger, one action, nothing downstream depends on it.
Low volume, low stakes
A few hundred tasks a month where a missed run is an inconvenience, not a revenue event.
Long-tail connectors
An obscure app with a Zapier integration and no API worth writing against. Pragmatism wins.
Prototyping a process
Proving a workflow is worth having before anyone builds it properly. Fastest route to an answer.
When a Zap has outgrown Zapier
If none of these apply, leave it alone. If one does, it is worth an hour of conversation.
Revenue data depends on it
Lead creation, routing, attribution or CRM writes. These need upsert keys, retries and a dead-letter queue — none of which Zapier is designed to give you.
Silent failure would go unnoticed
If a run can fail and nobody finds out for a fortnight, the failure mode is worse than the workflow is valuable. That needs alerting and an error path.
Task cost has become material
Per-task pricing scales linearly with volume while the logic stays identical. At enrichment volumes this is usually the first trigger to fire.
In a typical audit only a small minority of Zaps meet any trigger. Those move to a proper orchestration layer; the rest keep running and we document who owns them. A migration proposal that recommends moving everything is a sales document, not an audit. How we pick the tool.
Zapier, answered.
Should we move off Zapier?
Mostly no. Move the Zaps where revenue data depends on the run, where a silent failure would go unnoticed, or where per-task cost has become material. Everything else is fine as it is.
Is Zapier too expensive at scale?
It can be, because you pay per task while the logic stays the same. Whether that matters depends entirely on your volume — which is why we count tasks before recommending anything.
Can Zapier handle CRM automation?
For simple field updates, yes. For lead creation and routing it lacks the pieces that keep a CRM clean: idempotency, retries with backoff, and a queue for failures. That is the case for moving those specific workflows. CRM implementation.
Do you do Zapier work at all, or only migrations?
We build and fix Zaps too. Often the right answer is a cleaner Zap with a named owner rather than a new platform.
Get a straight answer on which Zaps to leave alone.
Two weeks, fixed scope. Every Zap listed with its task count, owner and verdict — keep, fix or migrate. You keep the document either way.