One useful pattern for a small product site is a static website form that writes to Google Sheets. You get a simple data store and operational view without deploying a dedicated backend.
I used this pattern for the Slap AI pre-launch flow.
The architecture
The flow is deliberately small:
Static website -> fetch() -> Google Apps Script web app -> Google Sheet -> email workflow
The website stays static. Google Apps Script acts as the HTTP endpoint and writes the request into a Sheet.
This works well for a launch list, contact form, lightweight preorder interest, or internal lead capture where a full application backend would be unnecessary.
What I capture
A form becomes much more useful when it records attribution from the beginning.
For Slap AI, the payload is designed to support fields such as:
- email;
- name;
- type of signup;
- quantity;
- source;
- campaign;
- consent;
- current page;
utm_source;utm_medium;utm_campaign;utm_content;utm_term.
That gives the Sheet enough context to answer questions later without trying to reconstruct where a signup came from.
Keep the endpoint configurable
I store the endpoint in configuration rather than hardcoding it throughout the frontend.
Conceptually:
export const SIGNUP_ENDPOINT = "https://script.google.com/.../exec";The form can then post URL-encoded or JSON data depending on the Apps Script handler.
If the endpoint is missing, the frontend should show a configuration error rather than pretending that a submission succeeded.
Validate before writing
The Apps Script should not append every request blindly.
At minimum I validate required values such as the email field. For more important forms, I also recommend validating length, allowed types, expected campaign fields, and any quantity fields before the row is written.
A useful pattern is:
- acquire a script lock;
- parse and validate the request;
- find the target sheet;
- append the row;
- trigger downstream email logic;
- record success or error status;
- release the lock.
The lock matters because multiple visitors can submit at roughly the same time.
Track email state separately
A row should distinguish between “form received” and “confirmation email sent.”
For example, a status column can move through values such as:
- Pending;
- Sent;
- Error.
This makes failures visible. Without that field, an appended row can look successful even when the email step failed.
Static-site form challenges
The difficult parts are usually integration details rather than the Sheet itself.
CORS and response handling
Apps Script web apps have their own behavior around redirects and browser requests. Depending on how the endpoint is implemented, a normal fetch() may need careful response handling.
Duplicate submissions
A user may click twice, refresh, or retry after a slow response. For important flows, check for duplicates using a stable identifier or recent matching email.
Formula injection
If arbitrary user input is written into a spreadsheet, values beginning with spreadsheet formula characters should be sanitized before writing.
Secrets
A static frontend cannot safely contain secrets. The endpoint can be public, but credentials and privileged operations belong inside Apps Script or another backend.
Add analytics only after consent
The form itself can work without analytics.
If analytics consent exists, I track events such as:
- form view;
- submit attempt;
- success;
- error;
- preorder CTA click.
UTM parameters can still be submitted with the form even when analytics scripts are disabled, depending on the site’s privacy design.
When Google Sheets is the right choice
Google Sheets is useful when:
- traffic is modest;
- humans need to inspect or edit rows;
- the workflow is operational rather than transactional;
- you want a low-cost bridge before building a full backend.
It is not a replacement for a transactional database when concurrency, strict consistency, authentication, or complex relational queries become important.
For Slap AI, the lightweight approach matched the stage of the product and kept the pre-launch architecture simple.
See Building Slap AI for the wider site architecture.