Zapier fundamentals: Zaps, paths, formatter and code

Transcript of the narration, chapter by chapter.

0:00 Zapier fundamentals

If you need an automation live by this afternoon and the people maintaining it aren't technical, Zapier is often the answer. It connects more than eight thousand apps, and its simple, linear model is easy for anyone on the team to read. In this lesson, you'll learn Zap anatomy, the built-in tools like filters, paths, formatter, looping and code, how mapping works, reliability features, and you'll build a lead-to-CRM Zap with deduplication and routing.

0:32 Lesson roadmap

Here's the roadmap for this lesson. First, why Zapier and when it's the right choice. Then an analogy that makes Zaps easy to reason about. Then the anatomy of a Zap and the built-in tools, with a closer look at code steps and reliability. Finally, two worked examples: a lead-to-CRM Zap and an onboarding flow for a coaching business in the UK.

0:59 Why Zapier

Why does this matter? Automation only helps if someone can maintain it. A brilliant workflow that only one engineer understands becomes a risk the day they go on holiday. Zapier's strength is readability: a marketer, a coach or an office manager can open a Zap and follow it from top to bottom. The trade-off is less fine-grained control than n8n or Make for complex data work, and costs that scale with tasks.

1:30 Analogy: a recipe card

Here's the analogy. A Zap is a recipe card. The trigger is: when this happens. Each action is a numbered step: then do this, then do that. Filters are notes like only continue if the eggs are fresh. Paths are: if you're baking for kids, do these steps, otherwise do those. Formatter is your measuring and chopping. And code is your special technique for when the recipe needs something unusual. Anyone who can follow a recipe can follow a Zap.

2:05 Zap anatomy

Now the anatomy. A trigger is an app event, or a Catch Hook from Webhooks by Zapier, or a schedule. Actions create, update and find records, and search actions with create if not found give you upserts without code. Filter by Zapier stops runs that don't match. Paths by Zapier branches. Formatter handles text, numbers, dates, line items and lookup tables. Delay waits. Looping runs actions for each item in a list. Code by Zapier runs JavaScript or Python. Sub-Zaps package reusable pieces. And Storage, Tables and Interfaces give you data and simple forms.

2:46 Code by Zapier

Code by Zapier is your escape hatch. In JavaScript, the values you map come in through an input data object, and you set output to return results. The example in the lesson text trims and lowercases an email and converts a phone number to international format using a small country code table for the UAE, Pakistan, Saudi Arabia, the UK and the US. Keep code steps short and commented. They run with time and memory limits, and the next person to maintain the Zap may not be a developer.

3:25 Reliability

Reliability features matter as soon as you go live. Zap history shows every run with the data in and out of each step, filterable by status. Autoreplay, depending on your plan, retries failed steps automatically. Error notifications tell you when things break, and higher plans add versioning and change history. Check your plan's current features. Make it a habit to look at Zap history after every change.

3:54 Example 1: lead-to-CRM

Example one, simple. A lead-to-CRM Zap. A form submission or Catch Hook triggers it. A code step normalizes the email and phone. A CRM find contact step searches by email and creates the contact if not found. A filter continues only for new contacts. Paths assign an owner: Gulf leads to one rep, Pakistan leads to another, everything else to a third. And a Slack message tells the owner with a link to the CRM record. Seven steps, clear to anyone.

4:29 Example 2: UK coaching onboarding

Example two, realistic. A coaching business in the UK sells one-to-one coaching and a group program through Stripe. When a payment arrives, the Zap finds or creates the customer in HubSpot. Paths split by product. One-to-one clients get a booking link email and a Trello card for the coach. Group program clients get tagged in Mailchimp and invited to the community platform. Formatter shows the start date in UK format, Slack notifies the coach, and a delay of three days sends a check-in email. No code, built in an afternoon.

5:08 Common mistakes

Three mistakes are common. Building twenty-step Zaps without sub-Zaps, so every change is painful. Split reusable pieces into sub-Zaps. Assuming polling triggers are instant. Many triggers check every few minutes, so check each trigger's type. And letting task usage explode by looping over large lists. Filter early so you don't pay for work you'll discard, and batch where apps allow it.

5:35 Watch me do it: intake Zap

Watch me do it. I build the same intake as a Zap. Trigger: Webhooks by Zapier, catch hook. I copy the URL and send the sample payload, and Zapier shows the fields. Step two: Code by Zapier, JavaScript. I map email, phone and country into input data and paste the normalization snippet from the lesson text. The test output shows the clean email and the international phone number. Step three: HubSpot, find contact by email, with create if not found switched on, mapping the clean values. Step four: a filter that continues only if the contact was just created, using the created flag that the find step returns. Step five: Paths. Path A, Pakistan, sets the owner to the Lahore team. Path B, UAE or Saudi Arabia, sets the Dubai team. Path C catches everything else. In each path, a Slack message to the owner with the name, company, source and a link to the record. I test each step with real sample data. One problem: my Slack message shows the raw phone from the trigger, not the cleaned one, because I mapped the wrong step. I remap it from the code step. Then I publish, submit three real test leads, and check Zap history: two new contacts, one duplicate stopped at the filter, and the task count for each run.

7:11 Recap + try this now

Recap. Zapier trades some control for readability and breadth. Use find or create for upserts, filters and paths for logic, formatter and small code steps for data, sub-Zaps for reuse, and Zap history for debugging. Try this now: build the lead-to-CRM Zap from the lesson text, test each step with real sample data, publish it, and check Zap history after the first three live leads.

7:39 Try this now

Try this now. Create a Zap with your form app or a Catch Hook as the trigger. Add Code by Zapier to normalize email and phone, using the snippet in the lesson text. Add a CRM find contact step with create if not found. Add a filter for new contacts only, and paths for Gulf, Pakistan and other. End each path with a Slack message to the right owner. Test every step with real sample data, publish, and review Zap history after the first three live leads.

Why Zapier

Zapier is the most widely adopted no-code automation platform, with a catalog of more than 8,000 app integrations. Its linear Zap model is the easiest for non-technical teams to understand, and its breadth means the app you need is probably already connected. It is cloud-only and billed primarily by tasks (successful action steps), with plan limits and features that change over time; check current pricing.

Zap anatomy

Element	Purpose
Trigger	Event in an app ("New form entry", "New deal") or Webhooks by Zapier "Catch Hook", Schedule by Zapier
Actions	Create/update/find records, send messages
Search actions with "create if not found"	Upsert patterns without code
Filter by Zapier	Continue only if conditions match
Paths by Zapier	Branch into different sequences based on rules
Formatter by Zapier	Text, numbers, dates, line items, lookup tables, utilities
Delay by Zapier	Wait for a time or until a date
Looping by Zapier	Run actions for each item in a list
Code by Zapier	JavaScript or Python for custom logic
Webhooks by Zapier	Receive webhooks and send custom HTTP requests
Sub-Zaps	Reusable Zap components called from other Zaps
Storage / Tables / Interfaces	Simple key-value storage, database tables, forms and portals

Mapping fields

Each step's output fields appear in a picker for later steps. Zapier often provides both a "pretty" and a raw version of fields. When fields are line items (arrays), some apps accept them directly; otherwise use Formatter "Line-item" utilities or Looping.

Code by Zapier

JavaScript example (inputs mapped into inputData):

// Code by Zapier (JavaScript). Map email, phone, country into Input Data.
const email = (inputData.email || "").trim().toLowerCase();
let phone = (inputData.phone || "").replace(/[^\d+]/g, "");
const cc = { AE: "971", PK: "92", SA: "966", GB: "44", US: "1" }[inputData.country || "AE"];
if (phone.startsWith("00")) phone = "+" + phone.slice(2);
if (!phone.startsWith("+")) phone = "+" + cc + phone.replace(/^0+/, "");
output = [{ email, phone }];

Python version uses input_data and sets output. Keep code steps small and well commented; they run with time and memory limits.

Reliability features

Zap history: every run with data in and out; filter by status.
Autoreplay (plan-dependent): automatically retry failed steps.
Error notifications and custom error handling paths (check current features).
Versioning and change history on higher plans.

Worked example: a UK coaching business onboarding flow

Trigger: Stripe "New payment". Steps: Find or create the customer in HubSpot; Paths by Zapier: Path A (1:1 coaching product) -> create a Google Calendar booking link email and a Trello card; Path B (group program) -> add to a Mailchimp audience tag and a Circle community invite; Formatter to format the program start date for the UK; Slack message to the coach; Delay 3 days -> send a check-in email. No code, built in an afternoon, readable by the whole team.

Hands-on: build a lead-to-CRM Zap with dedupe and routing

Trigger: your form app ("New submission") or Webhooks by Zapier > Catch Hook.
Code by Zapier (normalize email and phone, as above).
CRM > Find Contact by email, with "Create if not found" enabled.
Filter: only continue if the contact was newly created (or if lifecycle stage is empty).
Paths: Path "Gulf" (country is AE or SA) -> assign owner A; Path "Pakistan" (PK) -> owner B; Path "Other" -> owner C.
Slack > Send Channel Message with name, company, source and a CRM link.
Test each step with real sample data, then publish and watch Zap history for the first live runs.

Pitfalls

Building 20-step Zaps without sub-Zaps, making maintenance painful.
Assuming polling triggers are instant; check each trigger's type and frequency.
Letting task usage explode with Looping over large lists.

Measuring success

Time from lead submission to owner notification (target: under five minutes).
Duplicate contacts created per month (target: zero, thanks to find-or-create).
Task usage per lead (watch for loops and late filters).
Failed runs per week and time to fix, from Zap history.

When to move beyond Zapier

Consider n8n or Make when you need complex data transformations across many line items, very high volumes where task costs grow quickly, self-hosting for data residency, or deep custom code. Many teams keep Zapier for simple, broad integrations and use another platform for the heavy lifting.

Zapier fundamentals: Zaps, paths, formatter and code