Notes from Building
1. The API is the secret sauce, we just weren’t treating it that way.
We didn’t need a new CMS or a second database of project blurbs. The website already structured portfolio data for case studies, filters, and internal tooling. The deck app is a different presentation of the same records. That’s the “marketing site as app” idea in practice: the site isn't pages, it’s an interface to your own data.
2. Deck URLs are a tiny programming language.
Two-character tokens, read left to right: project slot, category sweep, newsletter issue. Parsing is finicky — M35 is not "category M + project 35" unless the remainder is a valid project code. We learned to treat the URL segment as a token stream, not a fixed pattern. Worth it: a link like /01MD02 is simple, shareable, and encodes real intent.
3. Integration over replacement.
We didn’t rebuild the portfolio. We attached to what existed. The deck builder, the public slideshow, and the admin picker are thin layers on top of endpoints and content we are already maintaining for the marketing site. Less to go stale. Fewer questions about “which version of the Figma deck is correct?”
4. The bigger picture
A good marketing site isn’t just something you look at. It’s something your business runs on: proposals, decks, case studies, newsletters, all fed from the same well.
That’s what we mean when we say a “marketing site is an app.” Not because we just add more features. Because the same platform and content can power a website, a sales deck, a client portal, and whatever comes next.