Structuring Airtable bases
Last updated: November 11, 2026
Plan availability | All plan types |
Platform(s) | Web/Browser, Mac app, and Windows app |
Related reading | |
Airtable terminology | Base - a collection of data in Airtable, designed to contain all of the information a user needs for a project or workflow. Bases can have multiple tables, each containing data; within each table, there are records, with data for each record stored in fields. |
This article details best practices for effectively structuring Airtable bases to ensure the best performance possible.
How can I improve base performance in Airtable?
Create individual views that use filters and hidden fields that only display individual collaborators and associated information.
Delete any unused records.
Remove non-essential formulas, linked records, or rollup fields .
Reduce API usage or insert pauses between calls, when possible.
Refer to to this section of the base performance article where you will learn more about how automations and syncs can affect base performance.
What base design behaviors degrade Airtable performance?
Base performance, including things like speed, collaboration, and usability, is often related to redundant tables—all living in the same base. The following behaviors should be avoided whenever possible:
Creating multiple tables in the same base with similar fields that don’t use linked records.
Creating multiple tables in the same base with similar structures and uses.
Creating tables that summarize information in other tables without using linked records in the same base.
Creating multiple tables in the same base, tracking different stages of projects.
Creating individual tables for different time-frequencies (weeks, months, quarters).
Creating individual tables for categories (products, projects, types of events).
Creating individual tables for each collaborator.
What base design behaviors improve usability?
Creating collaborator, text, or select fields to assign tasks to individual collaborators and group records to keep them separate instead of creating tables for each collaborator.
Using views to section out related data instead of having data unnecessarily spread across multiple tables.
For the decisions behind these behaviors - field versus linked table, stages as records, and where to aggregate - see Choosing how to model your data in Airtable.
Example: one table instead of many
The guidance above is easier to apply against a concrete case. Here's the most common restructure.
Before - a base with a separate table per team member. Each table repeats the same fields (Task name, Due date, Status), so a single view of all work in progress is impossible and every new team member means a new table.
"Tasks - Alex" table
"Tasks - Priya" table
"Tasks - Sam" table
After - one table, one field, many views. Adding a team member now means adding a value, not a table. The same pattern applies to the other cases above; replace a table-per-month with a date field, and a table-per-category with a select field.
Create a single "Tasks" table containing every task.
Add a collaborator field named "Assignee".
Set each record's "Assignee" to the person responsible.
Create a grid view per person and filter it to "Assignee is [person]".
Group the unfiltered view by "Assignee" to see everyone's workload at once.
Example: A parent record with repeating occurrences
The example above collapses many tables into one. The opposite question comes up just as often: when one thing happens repeatedly—a webinar program and its individual sessions, a course and its cohorts, or a project and its milestones—do you keep everything in one table and separate it with views, or do you create a second table and link the two?
Use one table with views when each occurrence stands on its own and the parent is only a label you filter or group by. Use two linked tables when the parent holds details that are the same for every occurrence while each occurrence has details of its own. The quickest test: if changing one piece of shared information would mean editing more than one record, the parent belongs in its own table.
For a webinar program, that means two tables:
Create an “Events” table with one record per program, and add the fields that stay the same across sessions, such as the audience, the program type, and the description.
Create a “Sessions” table with one record per session, and add the fields that change from session to session, such as the date, the registration link, and the recording link.
In the “Sessions” table, add a Link to another record field pointing at “Events” so each session belongs to one program.
In the “Events” table, add a rollup or count field over the linked sessions to summarize them—for example, the number of sessions scheduled or the date of the next one.
In the “Sessions” table, group a grid view by the linked “Events” field to see each program’s sessions together, and filter separate views by date or status for production tracking.
Supporting tables—speakers, content assets, and production tasks—link to “Sessions” the same way. A speaker who appears in several sessions, and a session that has several speakers, form a many-to-many relationship that a linked record field handles without duplicating anyone; see Understanding linked record relationships in Airtable. For the decisions that sit behind this structure—field versus linked table, stages as records, and where to aggregate—see Choosing how to model your data in Airtable.
How can I reduce the amount of information I see or share in bases?
Check out our Limiting linked record selection to a view article to learn how limit what information is viewable.
How can I customize my base’s appearance?
Check out our Creating and managing Airtable bases to learn how customize base appearance.