Mapping data between steps
Mapping means telling a step which piece of earlier data goes into which of its own fields. It's how a customer name from a form ends up in an email greeting.
What you need
- Sample data from the earlier step, so you can see the real field names.
- A clear idea of which field in the next step should receive each value.
- A sense of the kind of value each one holds: text, number, date, yes or no, or a list.
How it works
n8n defines it simply. Referencing data, or data mapping, means accessing information from previous nodes in your workflow.1 The values aren't changed on the way. When you reference data, you're not changing it.1
In the editor you usually click or drag. n8n lets you create expressions by typing them in a parameter's field or dragging and dropping fields from the Input panel.1
Make explains it with a box image. You can map any information that a module gets in a bundle.2 A bundle is like a box of information. To know what you can map from a module, you run only that module and look at its bundles.2
Get sample data first
Don't map from memory. Run the earlier step once and read what it returned. You want the field label the tool will actually fill in.
Be careful with test records, too. Zapier warns that test records may not match the data your Zap receives during live runs.3 n8n has a similar trap with pinned data, which is frozen output used while you build. Production workflows ignore pinned data and request new data on each execution.4
So test with a record that looks like your messiest real one, not your tidiest.
Lists and single values
Some fields hold one value. Others hold many. Make separates two kinds of container. An array has information of only one kind, such as a list of dates. Collections can have different kinds of information and are like a moving box.2
If you map a list into a field that expects one value, check what your tool does with it before you trust the result.
What goes wrong
- Wrong field picked. Two fields can have similar names, such as a sent date and a created date. Check the sample value, not just the label.
- Empty field. An optional value is missing on some records, so the next step gets nothing. In Make, the Exists operator checks whether a specific bundle item is filled in, and a filter can use it.5 Filters get their own page, linked in the branching notes.
- Type mismatch. Zapier notes that rules only work for the type of data specified in parentheses, so a text rule won't work on a number field.6
- Shape changed upstream. If an app renames a field, old mappings can point at nothing. Re-run the step and re-pick the field.
- Text from an AI step. The reply is text, even if it looks like a number or date. If a later step needs a real number, test that it arrives in a form the step accepts. The AI step page explains why replies can vary.
Keep customer data in mind
Mapping moves real data. Before you map a customer's details into an AI step or a message, read your business's written rules. Passwords, API keys, customer records and card numbers shouldn't reach an AI prompt unless those rules allow it.
For legal rules about personal data in your own country, go to your regulator's official page.
Cited material
- Reference data, n8n Docs
- Mapping, Make Help Center
- How Zap triggers work, Zapier Help Center
- Key concept glossary, n8n Docs
- Filtering, Make Help Center
- Add conditions to Zap workflows with filters, Zapier Help Center