StackReceipt

Automation / Automation guides

Know what starts the bill.

Three guides for choosing an automation tool: count the work it bills, check the plan it needs, and decide who will fix it when it fails.

By Penny · StackReceipt · Documentation research and worked examples, not hands-on reviews

Start with the failure or billing question closest to your job. The worked examples show their inputs, arithmetic and limits; each guide ends with checks that can rule a tool out.

Zapier

A form-to-Slack handoff: count 500 successful actions, see why a filter changes plan eligibility, and avoid sending a message twice during recovery.

Make

A ten-record batch: follow bundles through two actions, add the empty source checks, and decide whether faster polling is worth the extra credits.

n8n

A document-signing lifecycle: separate requests, completion callbacks and reminders into executions, then decide whether your team can own the deployment.

A shared vocabulary

A record is not a billing unit.

One customer request may cause several actions, many module runs or multiple workflow starts. Draw the work before comparing allowances.

Trigger and action
A trigger starts work; an action does something in a destination. Zapier’s guide separates successful actions from its unresolved trigger-rate wording.
Bundle and credit
A bundle is a package of data passing through Make. The number of bundles reaching an action determines how often it runs; credits are what the platform meters.
Execution
n8n counts workflow runs rather than individual nodes. A separately triggered callback can start another run; a paused workflow resuming is a different case.
Polling and webhook
Polling checks a source on a schedule; a webhook receives an incoming event. Empty-check accounting differs by platform and by trigger type.

These are orientation notes. Use the linked guides’ dated sources and explicit assumptions before estimating usage.

Compare the same job.

Write down the source event, required destination result, acceptable delay and failure owner. Check whether your existing app already does the job before adding another service.

Then test normal, missing-field, duplicate and failed-connection cases with synthetic data. Keep the output IDs and platform usage entries. An allowance is useful only if the required workflow can run on that plan.

Open the workload calculator →