Workflow automation tools compared: Make, Zapier, n8n and Pabbly
Compare workflow automation tools by how they count usage, then build, secure and troubleshoot scenarios for a small business.
Begin with your workflows, not the tools. List the two or three automations you actually need, how many steps each has, and how often each runs per month. Automation platforms count usage differently (per task, per operation or credit, or per workflow execution), so the same workflow can cost very different amounts depending on how it is counted. Also decide early whether you want a hosted service or are willing to maintain a self-hosted setup.
Read the best-for and head-to-head comparisons first, especially the ones that count usage for an identical workflow, then the free plan and pricing explainers for your shortlist. Review pages and the safety page come next, once you are down to one or two options.
After you build, keep the practical guides close: webhook basics, error handler choices, duplicate prevention, polling and rate-limit budgeting, and the security checklist for connections and API keys. When a scenario or Zap stops running, go to the fix-it page that matches the symptom and follow the vendor's documented steps before rebuilding anything.
Pabbly counts only actions on external apps; Make counts most module runs, including scheduled checks. A worked 200-submission example shows which free plan holds up.
Run one hypothetical three-step workflow through Zapier's task rules and Make's credit rules, using figures from each vendor's official pages. This is a counting comparison, not a quality ranking.
Compare Make's per-operation credits with n8n's per-run executions, the work of self-hosting n8n, integrations and code steps, using each vendor's official pricing and documentation, with a worked monthly example.
A trade-off table organized by the reason you want to leave Make (cost, complexity, self-hosting, app coverage), with billing units and prices taken from each vendor's official pricing page.
Price one multi-step workflow on Zapier, Make, Pabbly Connect and n8n using each vendor's own billing unit and official pricing page, and list what you give up by leaving Zapier.
What Microsoft 365 already lets you automate, when Power Automate needs a $15 Premium license, how Make's credits compare, and which one fits a small team.
A research review of Pabbly Connect from its official pricing page and terms: how tasks are counted, which internal steps are free, app coverage, support and plan limits, with a fit table against Make.
Scheduled checks, extra records, retries and AI modules each use Make credits differently. Audit one run, then fix the biggest bucket. Includes schedule maths from Make's own example.
How Zapier counts tasks according to its help center, what the Free plan leaves out according to its pricing page, and a quick way to estimate whether 100 tasks a month is enough.
Compare n8n Cloud plans with the self-hosted Community edition using n8n's own pricing page, documentation and license text, and decide who should take on hosting.
Google's published Apps Script quotas for consumer and Workspace accounts, what they mean for reminder emails and scheduled jobs, and when a paid automation tool is still the better choice.
What each Make error handler does according to the Make Help Center, including the renamed Ignore (now Skip) and Break (now Retry), and how incomplete executions fit in.
A research review of Make: how credits are counted, what the free plan allows, the scenario builder's strengths, error handling and its learning curve, with worked credit math against Zapier tasks and Pabbly Connect tasks.
A plan-by-plan summary of what Make's official pricing page and help center state about Free, Core, Pro, Teams and Enterprise, what one credit buys, and worked monthly credit totals for three sample workloads.
What Make's official pages say about the Free plan: 1,000 credits a month, storage limits, scheduling, AI options and the points where a free account stops working or needs an upgrade.
What Make AI Agents are according to Make's own documentation, how runs, tool calls and tokens turn into credits, how model choice changes the cost, and a checklist for a safe first agent.
Five official Make templates for lead capture, invoices, social posts, email list sync and reporting, with the apps each needs, an estimated credit count per run and the changes to make before switching on.
Use Make's MailerLite Create/Update a Subscriber module so repeat sign-ups update one record instead of creating copies, with field mapping, consent handling and the credits used per new subscriber.
Split an invoice's line items with Make's Iterator, act on each line, then combine them with an Array aggregator, with credit counts per bundle so you can see where the cost comes from.
What Make's Teams plan adds over the plans below it according to Make's documentation, which team roles a small business actually needs, and how to decide the cost for a 3-person business.
An example reminder workflow from Xero to email, the stop conditions that keep paid or disputed clients from being chased, and monthly Make credit and Pabbly task estimates.
Connect a website form to HubSpot CRM and MailerLite with Make or Pabbly Connect, keep one record per person, respect consent, and estimate monthly credits and tasks.
How instant (webhook) triggers differ from polling triggers, what that means for Make credits and Pabbly tasks, and a safe way to test a webhook before real data flows through it.
A sourced checklist for Make and Pabbly Connect: who can use a connection, how to scope tokens, team roles, 2FA, webhook URLs, and how long run data is kept.
How Make stores failed runs, the documented automatic retry schedule, how much storage you get, how to resolve stored runs, and when a replay can create duplicates.
Sourced facts on who owns Make, which security audits and certifications it publishes, where data is processed, and what a small business should check before connecting client accounts.
A decision table by budget, technical comfort and volume using each vendor's published pricing unit, with a five-step workflow costed on all four and links to the head-to-head guides.
Make's documented reasons for switching a scenario off, how to find the failed run, the fix for each cause, and how to restart without creating duplicate records.
Make's documented steps to verify, reauthorize, edit or replace a connection, which scenarios share it, why Google and Microsoft tokens stop working, and how to rerun failed work without duplicates.
Make's notice on ending Integromat support in 2023, what changed for anyone following old Integromat tutorials, and how operations now map to Make credits, with a checklist and a worked cost example.
Why Zapier turns a Zap off or holds its runs, from a 95% error rate to task limits, plan changes, disconnected apps and flood protection, and the steps Zapier documents to fix the cause and replay what was missed.
About 30 Make terms defined from Make's own documentation, with where each one appears in the editor, whether it uses credits, and the closest Zapier or Pabbly Connect name for people switching tools.
A form-to-sheet-to-MailerLite scenario on Make's free plan: run it once, read the bundle bubble, count the credits per run, set a schedule the free plan can afford, and fix the five settings beginners miss.
Every Make schedule option explained, from regular intervals and fixed days to on-demand and instant triggers, with the documented minimum intervals, how the scenario time zone works, and a table of checks per month for each interval.
How Make data stores, data structures and record modules work, how much storage each plan gets, and a decision table for data store versus Google Sheets versus Airtable, with a worked example that stops duplicate order emails.
How Make routers order and run their routes, what a fallback route does, which filter operators exist, how AND and OR rules combine, and why filtered-out bundles cost nothing, with a VIP, standard and spam lead example and a test checklist.
Configure Make's HTTP Make a request module against a documented public API, GitHub's Get a repository endpoint: method, headers, authentication, query string versus body, Parse response, status-code handling, Parse JSON with a data structure, and what 401, 404 and 429 mean.
A cheat sheet for Make's documented date functions and tokens, with worked examples for US-format CSV dates, a reminder 24 hours before an event, and the off-by-one-day bug caused by parsing and formatting in different time zones.
The technical limits Make documents for each plan and credit level, including maximum file size, data store storage, webhook queue size, webhook rate and log retention, with the symptom you see when you hit each one and the cheapest workaround.
How Make's Watch New Rows trigger decides what is new, why blank rows and edited sheets cause misses or repeats, and how to use Search Rows with Update a Row or Add a Row to keep one row per record, with credit costs and a setup checklist.
What changed with Make's Gmail V4 app, when you still need your own Google Cloud OAuth client, why connections expire after 7 days or 6 months, and the Gmail sending limits that apply to anything a scenario sends.
The beehiiv triggers, actions and searches in Make, what each beehiiv plan includes for the API and webhooks, and a form-to-beehiiv scenario that sends custom fields and a UTM source, with the API fallback when a module lacks a field.
The Kit (formerly ConvertKit) triggers and actions in Make, the API Secret connection and what Kit's API docs say about API keys versus OAuth, Kit plan requirements, and a worked Stripe purchase scenario that tags a buyer and adds them to a sequence.
A Make scenario that turns a completed Stripe Checkout into a MailerLite subscriber with a product field and buyer group, respects promotional-email consent, removes refunded buyers from the group and stays safe when Stripe resends an event.
How a Pabbly Connect workflow is built from a trigger, actions, mapping, filters and routers, which steps Pabbly counts as tasks and which internal steps are free, what the free plan allows and how auto re-execution helps with failed steps.
How to generate the MailerLite API token Pabbly Connect needs, which MailerLite events Pabbly documents, how custom fields and groups behave when you map them, and the 120 requests per minute limit that governs bulk runs.
A plain-language explainer of what you hand over when you connect an app to Make or Pabbly Connect with OAuth or an API key: scopes, expiry, the risk of an unscoped key, where to revoke access in Google, MailerLite, Kit and beehiiv, and a least-privilege checklist.
A cross-vendor table from MailerLite, Kit and beehiiv developer docs, plus Moosend's automation route: which subscriber, unsubscribe, bounce, click and purchase events each one pushes, the plan required, retry schedules and signature checks, and which workflows must still poll.
The documented request limits, batch endpoints and 429 behavior for MailerLite, Kit and beehiiv, a worked example of how long a 20,000-contact field update takes on each, and how to pace it in Make or Pabbly Connect.
A fill-in runbook for any automation: purpose, trigger, owner, connections, data touched, failure alerts, rollback and a monthly credit budget, demonstrated on a Make scenario and kept in Make's scenario notes.
A pre-launch checklist for Make and Pabbly Connect: run once with test data, empty and malformed inputs, duplicate triggers, the error-handler path, time zone edge cases, a credit-per-run estimate and the alert recipient, with the feature to use for each check.
A handover checklist built on Make's documentation for organizations, teams, roles, connections and blueprints: transfer ownership, move scenarios, replace connections tied to a departing person's logins, rotate API keys and confirm webhooks still fire.
What Kit Visual Automation entry points, actions, events and conditions each do, how statuses affect subscribers, which plans include them, and a worked course launch funnel showing exactly where subscribers exit.
Research guide based on Kit documentation; not a hands-on testRead guide