An honest comparison
Custom automation
vs Airtable
Airtable earns its reputation. It is the fastest way to give a team a shared, structured place for data without involving a developer. The trouble starts when the base stops being a place where data lives and becomes the place where your business runs. That transition happens quietly, and there are four specific signs it has already happened.
What Airtable is genuinely great at
Structured lists many people edit. Content calendars, applicant pipelines, inventory that changes weekly, project trackers. It gives non-technical teams relational thinking without the word database, and the interface layer means people can work without seeing the raw grid. For that job it beats a spreadsheet decisively and it beats custom software on time-to-value, which is not a small thing.
The four signs the base became the system
One: there are automations inside the base that other automations depend on, and the order matters. Two: you are hitting record limits or per-seat pricing that grows with headcount rather than with value. Three: a customer or supplier needs to see part of it, and you are handling that with a shared view and a promise. Four: someone has written scripts inside the base, and those scripts are now the actual business logic. At that point you are running software without any of the safety that software normally has.
What replacing it usually looks like
Rarely a big-bang rebuild. The common path is that the data model moves to a real database, the parts humans touch daily get a proper interface, and the base stays for the things it is good at. A useful benchmark for what the destination can hold: a production schema with 183 tables, 229 API routes and 25 background jobs, designed and operated by one person. The number that matters there is not 183, it is that a structure that size stays maintainable, which a base of comparable complexity does not.
The comparison, straight
| Airtable | Custom system | |
|---|---|---|
| Best for | Structured lists many people edit | Operations your business depends on |
| Time to first result | Same day | 2 to 4 weeks for a first scope |
| Cost model | Per seat, growing with headcount | Built once, then nothing to rent |
| Record volume | Fine into the tens of thousands | Hundreds of thousands is unremarkable |
| Business logic | Automations and scripts inside the base | Versioned code with a test suite |
| External access | Shared views and trust | Real permissions, per account |
| Interfaces | Interface layer, within its limits | Whatever the job needs |
| Ownership | Their platform, export is not the same as owning | Repo, database and cloud on your accounts |
The specific limits, in order of when they bite
Per-seat pricing bites first, usually when the team grows past the people who actually need to edit. Record limits bite next, and the workaround of splitting data across bases is where reliability starts to erode. External access bites third, because there is no honest way to give a customer their slice without giving them a view of the shape of everything. And scripts bite last and hardest: logic that lives inside a base cannot be reviewed, tested, or handed to a new developer in any normal way.
The rule of thumb
Keep Airtable while it is a shared, structured list your team edits and while a mistake is annoying rather than expensive. Move when outside parties need scoped access, when the automations have dependencies on each other, when scripts hold the logic, or when the per-seat bill has quietly become a five-figure annual line. And keep it even then for the things it does well; replacing everything at once is almost always the wrong call.
Common questions, answered straight
Can we keep Airtable as the interface and change what is underneath?
Sometimes, and it is a reasonable middle step: the database of record moves, Airtable stays as a view for the team that likes it. Whether that works depends on how much of your logic currently lives in the base rather than in the data.
What happens to our existing records?
They move, and that is usually the bigger half of the work. As a sense of scale, one migration involved 863,484 contacts imported, deduplicated to zero duplicates and normalized, including roughly 1.2 million empty fields that had been counted as data.
How long does the first scope take?
Typically two to four weeks to production. The teardown that tells you which scope to pick takes 72 hours and costs nothing.
What if Airtable is genuinely the right answer for us?
Then the teardown says so. I do not sell Airtable builds, which is exactly why you can trust that recommendation.
Related reading
Has your base quietly become your system?
Free systems teardown: your three biggest time sinks, priced, and an honest call on whether the answer is a tool or a build. In writing, within 72 hours, no obligation.
Get your free systems teardown