Reconsidering the Product
While building the add-on, a question came up that I should have asked before writing a line of code: Google Docs already has File → Download → PDF. What are we actually adding?
The honest answer: not much. The add-on saves the PDF to Drive instead of downloading it to your computer. That's a real workflow difference, but it's a minor convenience, not a missing capability. It's not something people are paying $5–10/month for.
Where the logic broke down
The market research identified real complaints about third-party PDF converters: silent truncation at 80 pages, dropped links, misleading free claims. Those are genuine problems. The mistake was concluding that the solution was another PDF converter.
The people experiencing those problems are using third-party tools — Lumin PDF, generic converters — when they could just use Google's own export. The correct advice to those users is "stop using that tool." Our add-on calls Google's export API internally, which means we're producing the same output as File → Download, wrapped in a slightly more convenient package. We're not solving a problem that exists for our target users; we're solving a problem that stops existing the moment they use the built-in option.
The PDF gap was real in the market research sense — there are underperforming paid tools with real user complaints. But "underperforming compared to what Google already provides for free" is not a gap we can profitably fill. A paid add-on that does what a free built-in does isn't a product.
What we built anyway
The add-on works. Docs, Sheets, and Slides convert cleanly to PDF. PDF → Google Doc conversion via Drive's OCR is functional. The code is solid and the infrastructure is in place. None of that is wasted — the same Google Workspace add-on scaffold, GCP setup, and Drive API integration carries forward to whatever we build next.
But we're not going to market with a PDF converter. The original market research was right about the platform (Google Workspace Marketplace) and right that the e-signature gap is larger and more structurally defensible. It was wrong about PDF conversion being the right entry point.
What comes next
We're back to the question of what to build. The e-signature category has a real structural problem — PandaDoc's 2.9/5.0 rating at 640K installs is driven by a single architectural decision: recipients get pushed into a paid subscription just to sign a document they were sent. That's a design choice, not a technical limitation. A tool that lets anyone sign without an account would directly address the complaint.
The floor is higher. E-signatures carry legal weight — ESIGN Act, audit trails, identity verification. These aren't optional. But "higher floor" doesn't mean "wrong direction." It means more work before shipping, not a different category.
We may build toward that. Or we may find a different angle. Either way, this post exists because documenting the wrong turns is the point of this blog, and walking away from a working but unbuyable product after two days of build time is the right call.