Zapier, Make or n8n for lead handling: when a no code flow is enough

A no code flow is enough when lead handling means moving a record from where it arrives to where it is worked, with a notification and a few rules. The three tools differ mainly in how they count usage: Zapier counts tasks, n8n counts whole workflow executions, and Make counts credits. For 1,000 leads a month through a four step flow, that is roughly 4,000 Zapier tasks against 1,000 n8n executions. A flow stops being enough when the job needs a conversation, a decision about what to say, or per account sending limits.

Lead handling is where most small teams first reach for automation: a form fills, a message arrives, and somebody has to do something quickly. This post compares the three most common tools on the points that matter for that job, using their own pricing pages, and ends with the signs that you have outgrown a flow.

What does lead handling mean here?

For this comparison it is the first few minutes of an inbound lead's life:

1. A lead arrives (a form, an ad, a message, a calendar booking).
2. It is checked for duplicates and cleaned.
3. It is stored in one place, and routed to the right person.
4. The person is told, and the lead gets an acknowledgement.

Everything after that, the actual conversation, is a separate question. Flows are good at steps one to four and weak at what follows.

How do the three tools count usage?

Sources are Zapier's, Make's and n8n's pricing pages, read on 19 October 2026. Prices change, so treat the table as a snapshot and check the pages before you buy.

What does 1,000 leads a month cost?

Take a flow with a trigger and four action steps: look up the email, check the CRM for a duplicate, create the record, post a message to the team channel. Process 1,000 leads a month.

• Zapier. Each action step that runs is a task, and the trigger is not counted. That is about 4,000 tasks a month. On the tiers listed, the 5,000 task level is $89 a month billed annually or $133.50 billed monthly.
• n8n. Each run is one execution. 1,000 runs is 1,000 executions, which fits inside Starter's 2,500.
• Make. It counts credits, and the paid plan prices were not in the page text we read. Count your operations and check the plan table on its site.

This is an illustration with an assumed flow, not a recommendation. The point is the shape: tools that charge per step punish long flows, and a tool that charges per run does not. If your flow is short, the difference shrinks. If it has twelve steps, it grows.

What do no code flows do well?

• Connecting systems that have no integration with each other.
• Routing by simple rules: country, deal size, source.
• Acknowledging a lead with a templated message.
• Notifying a person fast.
• Syncing between a sheet, a CRM and a calendar.

If this is all you need, a flow is enough and an agent is overkill. Keep it simple, give it one owner, and make every step safe to run twice. The previous post, n8n workflows for outbound, lists what breaks.

Where do flows stop being enough?

Four signs.

1. The reply depends on what the person wrote. A price question, a request for a date, a complaint. A flow can send a template. It cannot read, decide and answer.
2. The conversation continues. A flow has no memory of what was said last Tuesday unless you build one.
3. Time matters across channels. Answering on WhatsApp, Instagram, Messenger and live chat within minutes needs one inbox and one set of rules.
4. A person must step in at the right moment. The flow needs an alert, an owner and a clock.

The Hotel Pirin case study is a good picture of the second sign. A reservations agent answered guests on email, Messenger and WhatsApp with a median first reply of 41 seconds, quoting from live availability, in the guest's language. A flow could have acknowledged those guests. It could not have quoted, asked for dates and handed over when the question needed the owner. For the longer discussion, read speed to lead on inbound messages.

How do the tools connect to Sellinger?

None of them has a native Sellinger app, and we claim none. Sellinger offers a REST API with webhooks from the Full Stack bundle up, and a hosted connector for Claude and ChatGPT. A Zapier webhook step, a Make HTTP module or an n8n HTTP Request node can call the API. A realistic split is to let the flow do the plumbing, such as creating a lead when a form is submitted, and let the agent do the conversation.

How do you set up a safe lead intake in half an hour?

1. Pick one source to start with, such as a website form. Add others later.
2. Choose the key. Use the email address, lowercased, as the identity of a lead.
3. Add a lookup first. Search the CRM for that email. If it exists, update it and stop. If not, continue.
4. Create the record with the source and a timestamp, so you can see where each lead came from.
5. Notify one named person, not a shared channel nobody reads, and include the lead's own words.
6. Add an error path that messages you when any step fails, with the lead's data, so nothing is lost.
7. Test with three cases: a new lead, a duplicate and a lead with a missing field. Only then connect the real form.

Thirty minutes is enough for this, and it prevents the two most common faults: the lead who appears twice and the lead who never appears.

How should you choose?

• If your flows are short and you want the fastest start, pick the tool your team already knows.
• If your flows are long, or you want to self host, n8n's execution pricing and self hosting option are worth a look.
• If you need many ready made connectors to business apps, check the connector lists of each tool for the apps you use, not the totals.
• If the job is a conversation, a flow is the wrong tool, and no price comparison will change that.

A checklist

• [ ] I wrote the flow as a list of steps and counted them.
• [ ] I compared usage on my own step count, not on a headline price.
• [ ] Every create step checks for a duplicate.
• [ ] The trigger is a webhook where I need speed, and I know the polling delay where I do not.
• [ ] I know which part of the job needs a conversation.
• [ ] I have checked the pricing pages again this month.