Small teams often end up with automations in three places. A few Zaps someone made two years ago, one Make scenario a contractor built, and some CRM workflows. When something breaks, nobody knows where to look. The tool matters less than having a rule for which job goes where. This guide gives you that rule, a way to estimate the monthly bill before you build, and a short checklist to run against what you already have.
What you need
- A list of the automations you want, written as “when X happens in app A, do Y in app B.”
- Your CRM plan. Native workflows are often gated. HubSpot workflows, for example, need a Professional or Enterprise subscription. Pipedrive Automations need Growth or higher. Close Workflows need Growth or Scale. Zoho CRM has workflow rules on every edition, with different limits.
- Current pricing, which changes. At the time of writing:
- Zapier: Free has 100 tasks/month and two-step Zaps only. Professional starts at $19.99/month billed annually ($29.99 monthly) for 750 tasks. Team starts at $69/month annually ($103.50 monthly) for 2,000 tasks and up to 25 users.
- Make: Free has 1,000 credits/month, 2 active scenarios and a 15-minute minimum interval. Core is listed at $12/month for 10,000 credits, Pro at $21 and Teams at $38. Make advertises savings of 15% or more on yearly billing.
Step by step
1. Sort every job into one of three buckets
Go down your list and ask one question per row: does it stay inside the CRM?
- Stays inside the CRM (assign owner, create task, change stage, send internal alert, enroll in sequence): use native CRM workflows. They run on the same data and don’t count against a task allowance. The person who looks after the CRM can also see them.
- Moves data between two common apps in a straight line (form to CRM, CRM deal won to accounting tool, new contact to email list): use Zapier. It has the widest app catalog and the gentlest learning curve.
- Needs branching, loops, many records at once or data reshaping (process every line item on an order, merge two sources, call an API with no ready-made connector): use Make. Its visual scenarios handle that shape of job, and credits cost less per step.
If a job fits the first bucket, it goes there, even if you already pay for Zapier. Every Zap that only touches your CRM is a second place things can break.
2. Estimate the monthly volume before you choose a plan
The two tools count differently, and that’s where the bills surprise people.
Zapier counts a task each time an action step succeeds. Triggers, polling and built-in tools such as Formatter and Paths don’t count. Zapier also notes that some actions, such as AI steps, can cost more than one task.
Make counts a credit for each module action: read a row, search, create, update, transform.
Worked example. Say you get 200 web leads a month, and each lead runs: find contact in CRM, create or update contact, create deal, post to Slack.
- In Zapier, the trigger is free and the 4 actions cost 4 tasks, so 200 × 4 = 800 tasks. That’s just over Professional’s 750-task entry tier.
- In Make, the trigger module plus 4 actions is about 5 credits, so 200 × 5 = 1,000 credits. Core’s 10,000-credit tier has a lot of headroom.
Now add the jobs that should be native, such as “create a call task” and “assign owner.” In Zapier those would be 2 more tasks per lead, 400 a month, that you don’t need to pay for if the CRM does them.
3. Check how fast it needs to run
Zapier says triggers labelled Instant fire immediately on any plan. Other triggers poll: every 15 minutes on Free, 2 minutes on Professional, 1 minute on Team. Make’s minimum scheduling interval is 15 minutes on Free and 1 minute on Core and above.
For anything speed-to-lead related, you want an instant trigger (a webhook or app event), or the CRM’s own workflow. A 15-minute poll on a web form means some leads wait 15 minutes before anyone is even told. Check the trigger label before you build.
4. Decide who will maintain it
This matters more than the price.
- Native workflows: whoever administers the CRM.
- Zapier: anyone comfortable filling in forms. Most office managers can maintain a straight-line Zap.
- Make: someone who’s happy reading a flowchart with routers and filters. If that person leaves, a Make scenario is harder for the next person to pick up.
If nobody on staff fits the Make profile, either keep jobs in Zapier and native, or have whoever builds it in Make write a one-paragraph note on each scenario.
5. Audit what you already have
Spend 30 minutes on this:
- Export or list every active Zap and scenario. Note the trigger app, the action apps and the monthly runs.
- Mark any that only touch the CRM. Move those to native workflows when you next have a quiet afternoon.
- Mark any two automations that do the same thing. This is common: a Zap and a CRM workflow both creating the first task.
- Turn on error notifications in whichever tool you keep, and send them to a shared inbox, not one person.
- Write a one-line “where things live” note in your team docs: “CRM-only jobs live in the CRM. App-to-app jobs live in Zapier. Bulk/data jobs live in Make.”
6. Use one tool for app-to-app work if you can
Running both Zapier and Make is fine for a 50-person company with an ops person. For a 5-person team it usually means two bills and two sets of logins. Pick the one that fits most of your bucket-2 and bucket-3 jobs, and accept a slightly awkward build for the odd one out.
Common mistakes
- Using Zapier for CRM-internal jobs. It burns tasks, and the logic is split from the CRM where people expect to find it.
- Ignoring polling delays. A 15-minute free-plan poll on a lead form undoes your speed-to-lead work.
- Underestimating volume. Count actions per run, not runs. A 6-step Zap on 300 leads is 1,800 tasks.
- Nobody gets the error emails. Automations fail quietly when a password changes or a field is renamed.
- Upgrading the CRM just for workflows without pricing it. For one or two jobs, a Zap can be cheaper than moving every seat up a tier. For ten jobs, the CRM upgrade usually wins.
When it’s worth getting help
If you have under ten automations and they’re mostly straight lines, the rule in step 1 and an afternoon’s audit will get you tidy. It’s worth getting help when your automations have grown in all three places and nobody’s sure which one does what. Or when a job needs a CRM API rather than a ready-made connector. My B2B CRM page covers how I untangle that. The cost calculator helps you weigh the hours.