Skip to main content
Every business names its data a little differently. Your cost might be called averagecost in one system and unit_cost in another, and ERPs are full of custom fields. Field mapping is where you tell Elastly which of your fields mean what. You rarely start from scratch. When you connect a source, Elastly reads its schema and a sample of real rows, then proposes a mapping. You review what it flags, fix anything specific to your account, and save. Every sync after that follows the mapping.

​
What you actually have to map

Three fields. Everything else is optional and makes the price sharper when you map it. Quotes and orders also need an id, a customer reference, and a sale date, but your connector fills those in itself. You never map them by hand.

​
How Elastly proposes a mapping

For each canonical field, Elastly ranks every candidate in your source on three signals:
  • Name. How closely the source field’s name resembles the canonical one. This carries the most weight.
  • Type. Whether the source type fits (a money field should not be mapped from a date).
  • How often it’s populated. A column that is empty in most of your sample rows is a poor match, even with a perfect name.
Each field then lands in one of three states, which you see as pills in the mapping board:
Money is always your call. A field that holds money, or that needs converting to cents, is never auto-verified, even when the match is obvious. Cost is the basis for every margin Elastly stacks, so a wrong guess there is wrong everywhere. It always lands in Review.
If two candidates score close together, Elastly treats the field as ambiguous and asks you rather than picking a winner.

​
Product fields

sku and costCents are required. The rest sharpen the price or help match products when you map them.
floorCents and ceilingCents let a product carry its own limits in from your source, which is useful when a specific item has a contractual price band. They sit alongside the workspace guardrails in Guardrails & rules, and the tighter of the two always wins.

​
Customer fields

Only externalId is required. The others are what let a price vary by who is buying.

​
Quotes and orders

Elastly maps these itself. A quote carries an id, the customer it belongs to, its status, and its date. An order carries an id, the customer, and when it sold. Orders are the signal the pricing learns from, and quotes tell it what was won and lost. Only NetSuite syncs quotes today. See Connectors.

​
When your data isn’t a plain column

Real systems hide things. A mapping doesn’t have to be one column on one table:
A field on the record itself, like product.name. The common case.
Follow a code to its human label. Useful when your source stores a category as an id and you want the name behind it.
Always use the same value. Useful when a source has no field for something and it never varies, like a single-region account.
A mapping can also transform the value on the way in: convert money to cents, coerce to a number, treat “greater than zero” as in stock, parse a day-first date, trim, uppercase, or take the absolute value. And it can fall back: try one source, and if it’s empty, try the next.

​
What each source starts with

Each connector ships defaults for its standard fields, so most accounts barely touch the mapping. The gaps are deliberate: where a value lives is specific to your account, Elastly asks rather than guesses.
Category and brand have no default. NetSuite accounts keep them in different places, often a custom field, so Elastly asks instead of guessing wrong.
When your account keeps something somewhere else, like a cost in a metafield, point the mapping at that field instead.

​
Where you do it

Open the connector from Integrations and go to Field mapping. The board shows one column per record type: Product, Customer, Quote, and Order. Fields needing attention are flagged, and you can open any field to see the sample values behind Elastly’s guess before you accept it. You can change a mapping at any time. If your source schema changes, re-run discovery to pick up new fields without losing what you already verified.
Worth a second look. Category and customer segment are the two mappings to confirm on a real account. They decide how prices and recommendations get grouped, and the field they come from is specific to how you set up your system.