Home
/
Blog
/
Guides
Guides

SFMC Data Extension Naming Convention: A Template That Scales

A copy-ready SFMC data extension naming convention: seven type prefixes, SQL-safe rules, a matching folder tree, and a rollout plan for accounts full of copies.

Open Contact Builder in an account that has been live for three years and you will find Newsletter_final, Newsletter_final_v2, test_copy_john and forty data extensions nobody dares to delete. Every new audience starts with the same question: which table is the right one?

An SFMC data extension naming convention fixes this with one rule: every name states what the table is, where it comes from and who it is for, in the same order, every time. The template below is the one we recommend: a type prefix, a business area, a descriptor and a scope, separated by underscores, with no spaces.

This guide gives you the template, worked examples, the SQL rules that make names query-friendly, a folder structure that mirrors the prefixes, and a short rollout checklist you can apply to an existing account without renaming everything at once.

‍

Why an SFMC data extension naming convention matters

‍

What breaks without one

Names are the only index SFMC gives you. There is no tag system for data extensions, search works on names, and a SQL query references a data extension by its name, not by its external key. When names are inconsistent, every audience build begins with a hunt through folders and a few exploratory queries just to confirm which table is current.

The cost compounds. Teams duplicate tables because they cannot find the original, retention is never set on the copies, and the account slowly fills with near-identical data extensions. Salesforce's own account optimization guidance notes that while there is no set limit on the number of data extensions, too many of them can slow down both the interface and the API.

‍

What a good name has to do

A good name answers four questions before anyone opens the table: what kind of data is this, which business area or brand owns it, what does it contain, and how wide is its scope. It should sort sensibly in a folder list, survive a copy-paste into a SQL editor, and still make sense to a colleague who joins next year. It should also make sense to an AI assistant for Salesforce Marketing Cloud that reads your table names to pick the right source for a query.

It should not try to encode everything. Row counts, dates of last refresh and who built it belong in the description field or in your documentation, not in the name.

‍

The naming template, with worked examples

‍

The pattern: TYPE_Area_Descriptor_Scope

The template has four parts joined by underscores. The first part is an uppercase type prefix that tells you the role of the table in your data flow, and the second is the business area, brand or program.

The third part describes the content. The fourth states the scope: a region, a language, a frequency or a channel.

A loyalty audience for the French newsletter becomes SND_Loyalty_GoldMembers_FR. The daily order feed that lands from your e-commerce platform becomes IMP_Orders_Ecommerce_Daily. The customer master that every business unit shares becomes MASTER_Customers_Global.

‍

The type prefixes we recommend

Keep the list short enough to memorize. Seven prefixes cover almost every table an SFMC account holds, and each one maps to a typical data action in Automation Studio.

PrefixWhat it holdsExampleTypical data action
MASTERSource-of-truth tables, one row per customer or productMASTER_Customers_GlobalUpdate (Add and Update on import)
IMPRaw imports landing from files or integrationsIMP_Orders_Ecommerce_DailyOverwrite or Append
STGStaging tables produced mid-automation and read by the next querySTG_Loyalty_Scores_MonthlyOverwrite
SEGNon-sendable segments used as building blocksSEG_Loyalty_Gold_ActiveOverwrite
SNDSendable audiences used by sends and Journey Builder entrySND_Loyalty_GoldMembers_FROverwrite
SUPSuppression and exclusion listsSUP_HardBounce_GlobalAppend or Overwrite
LOGSend logs and audit tables that grow over timeLOG_SendLog_GlobalAppend

Temporary and test tables get a lifecycle suffix instead of their own prefix: SEG_Loyalty_Gold_Active_TST while you build, and _DELETEME when it can go. Salesforce suggests exactly this kind of suffix in its account cleanup guidance, because it makes an audit a simple search.

‍

SQL rules that make names query-friendly

‍

No spaces, no special characters

SFMC allows spaces in data extension and field names, but SQL does not love them. Any name with a space must be wrapped in square brackets inside a query, and the same applies to field names. A convention that bans spaces and punctuation other than the underscore removes a whole class of query errors.

The same thinking applies to columns. Name a field OrderDate, not Order Date, and EmailOptIn, not Email opt-in?. If you ever need to join two data extensions in SFMC SQL, bracket-free names keep the join readable.

Here is what a query looks like when every table follows the convention. It runs in a child business unit and reads the shared customer master through the ENT. prefix.

SELECT
    c.SubscriberKey,
    c.EmailAddress,
    c.Country
FROM ENT.MASTER_Customers_Global AS c
INNER JOIN SEG_Loyalty_Gold_Active AS s
    ON s.SubscriberKey = c.SubscriberKey
WHERE c.EmailOptIn = 1
  AND c.Country = 'FR'
  AND NOT EXISTS (
      SELECT 1
      FROM ENT.SUP_HardBounce_Global AS b
      WHERE b.SubscriberKey = c.SubscriberKey
  )
  • The SELECT list pulls only the fields the sendable audience needs: the key, the address and the country used for routing.
  • ENT.MASTER_Customers_Global is the shared customer master in the parent account; the ENT prefix is how a child business unit reaches a shared data extension in SQL.
  • The INNER JOIN on SEG_Loyalty_Gold_Active keeps only customers present in the Gold segment built by an earlier query in the same automation.
  • The WHERE clause applies consent and country, then NOT EXISTS removes anyone found in the shared hard-bounce suppression table.
  • Target data extension: SND_Loyalty_GoldMembers_FR, sendable, primary key SubscriberKey, data action Overwrite so the audience is rebuilt from scratch on every run.

Read that query cold and you can tell what each table is, where it lives and why it is there. That is the whole point of the convention.

‍

Keep the external key short and stable

The external key (customer key) is what the APIs, the import activity and many integrations use to find a data extension. Salesforce's developer documentation recommends limiting external keys to 36 characters and warns that longer keys can negatively affect other processes. Set it explicitly to the same value as the name, truncated if needed, rather than accepting the generated GUID.

‍

Never rename a table a query depends on

Because SQL references data extensions by name, renaming a table silently breaks every query activity that reads from it. The activity fails at its next run, and validation flags the missing table the next time someone opens it. When you adopt a convention in an existing account, create the correctly named table, repoint the queries, run them in parallel for a cycle, then retire the old one with a _DELETEME suffix. Skip that step and the first sign of trouble is a failed automation on a send day.

‍

Folder structure and lifecycle governance

‍

A folder tree that mirrors the prefixes

Folders and prefixes should tell the same story, so that a table is findable whether someone browses or searches. A tree that works well in most accounts has one top-level folder per type prefix, then one sub-folder per business area.

So SND_Loyalty_GoldMembers_FR lives in Data Extensions, then Sendable, then Loyalty. The shared customer master lives in Shared Data Extensions, then Master. Imports land in a folder called Imports, where retention is set aggressively because the raw file is not the source of truth.

This is also where AI tooling earns its keep: when a tool reads your data extension metadata to turn Gold loyalty members in France who can receive email into a query, clean names and a clean tree are what let it pick the right source table instead of a stale copy.

‍

Lifecycle suffixes and retention

A convention is only as good as the cleanup it enables. Three habits keep the account tidy over time.

  1. Set a retention policy at creation on every IMP, STG and TMP table, deleting rows or the whole table after a fixed period, so that temporary data expires without a human remembering it.
  2. Mark anything you are about to retire with _DELETEME and run a quarterly search for that suffix, deleting what is older than the agreed grace period.
  3. Give every MASTER and SND table an owner in the description field, and review the list of SND tables every quarter against the sends and journeys actually running.

If your account is already past the point where this feels manageable, the post on why data extensions become unmanageable in large SFMC orgs explains how the sprawl happens and where to start. For teams running several brands, scaling audience management across business units covers the shared-data-extension side of the same problem.

‍

Rollout checklist for an existing account

‍

You do not need a big-bang migration. Apply the convention forward first, then clean backward where it pays off.

  1. Write the convention on one page: the four parts, the prefix list, the lifecycle suffixes and three worked examples. Store it where every SFMC user can see it.
  2. Apply it to every new data extension from today, including the external key and the folder location.
  3. Inventory the existing tables and tag each one as keep, migrate or delete. Start with the sendable audiences and suppression lists, because they carry the most risk.
  4. For each table to migrate, create the new name, repoint the queries and imports, run both for one cycle, then retire the old one.
  5. Add retention to every import and staging table that lacks it.
  6. Put a 15-minute naming review into the monthly ops meeting until the convention holds on its own.

‍

FAQ

‍

Can you rename a data extension in Marketing Cloud?

Yes, the name is editable in Contact Builder and Email Studio. But SQL query activities reference the data extension by name, so every query that reads from it has to be updated, and any Journey Builder entry source or send that uses it should be re-checked. Treat a rename as a small migration rather than a cosmetic change.

‍

Should data extension names include dates?

Only when the date is part of the table's identity, such as a one-off campaign audience you will never refresh. For recurring tables, a date in the name goes stale on the first rerun and encourages copies. Put refresh cadence in the scope part instead, as in IMP_Orders_Ecommerce_Daily.

‍

How do I reference a shared data extension in SFMC SQL?

From a child business unit, prefix the shared data extension's name with ENT., for example FROM ENT.MASTER_Customers_Global. Salesforce's Trailhead SQL module documents this prefix, along with the rule that names containing spaces must be wrapped in square brackets.

‍

Is there a limit on how many data extensions an account can have?

Salesforce states there is no set limit, but it also warns that a very large number of data extensions can degrade performance in the interface and the API. That is the practical reason to combine a naming convention with retention policies and a regular cleanup.

‍

Keep the names clean, let QAiry do the SQL

‍

A clear naming convention makes every audience faster to build, whether a person or an AI writes the query. QAiry reads the data extensions in your SFMC account and turns a plain-language request into production-ready SQL and a target data extension, and well-named tables are what make that translation precise.

You can install QAiry for free in your own SFMC account, or watch the product demos to see a request become an audience end to end.

Share this article
QAiry for SFMC

Skip the SQL. Build segments by chatting.

QAiry turns plain English requests into Salesforce Marketing Cloud audience segments and data extensions — no SQL, no IT ticket, no waiting.

Built for SFMC · ISV Partner · GDPR-ready