BlueGlassFarm.com

When we started Blue Glass Farm we had a real field-of-dreams mentality. If we built it, they would come. Then the summer of 2022 arrived, our crops started growing, and we found ourselves with an abundance of vegetables and no customers. We needed a way for people who weren't already friends and family to find us and buy from us, so we decided to solve the problem with technology, and blueglassfarm.com was born. We didn't quite know what to do with it at first. I knew the technical side of building a website cold. What Blue Glass Farm taught us was the whys and hows of building a good web presence.

The product started humbly. I wanted customers to be able to find out about the farm, and I knew the site would be image heavy, because people eat with their eyes and we were selling food, so it had to be performant too. I'd been playing around with Gatsby, which had great image optimization baked in at the time. I also liked its markdown-based approach to content. The whole site lived as files that were easy to modify, version controlled like code, with no CMS to stand up behind it. A two-person farm does not need a monthly software bill to publish a paragraph about tomatoes.

The Gatsby site worked great for a couple of years. Then I sat down to do a full rewrite, new styling and a bunch of new content, and realized Gatsby had gotten slow. Builds crawled and the developer experience had fallen behind. I'd been keeping an eye on Astro, a really cool new project that let you pull together components from different frontend frameworks and serve them in a genuinely performant way. Next was the big player in the room and gaining adoption fast, but it was entirely React based, and I liked Astro's framework-agnostic approach. At the time it wasn't clear React would stay the winner as long as it has. The thing that really sank the decision, though, was content collections. Define your types of content, markdown or JSON, and pull them in without pageantry, fully type safe, which eliminated an entire class of bugs and got rid of the need for a CMS once and for all. Three years down the road I still think it was a great call.

The deeper reason for the rebuild is that the farm didn't know what it wanted to be when it grew up. We had sold every way we could find, and we knew the business would keep changing shape, so I wasn't just picking the best framework for the site we had. I was picking the one that could turn into whatever site we needed next, fast. Astro gave me everything Gatsby had gotten right, fast static-first pages, a real image pipeline, markdown as the source of truth, and it got rid of the creeping elephant in the room, the GraphQL content layer that sat under every Gatsby build and made it so slow. I'm telling you, that era put GraphQL into everything. Somewhere out there is a coffee pot with a GraphQL schema in it.

And the bet paid off, because the farm is a startup, and like any good startup we've pivoted in a meaningful way every single year. We pulled away from our online store, moved toward a CSA, where members buy the season up front and pick up a share every week, doubled down on it, and spun up u-pick on top. Every pivot meant new questions to answer, a redone navigation so things stayed easy to find, and whole encyclopedias of crop information for the foods we grow and what to do with them. The site absorbed all of it without a fight. A member handbook that walks the season, pages for how pickup and weekly ordering work, waitlists with embedded Klaviyo email-capture forms. And Astro handles all of it in a way that's incredibly easy to maintain.

We even looked at pulling our Shopify store into the site to have one cohesive experience, rather than a Shopify presence and a website presence, which we could do through Hydrogen components thanks to Astro's agnosticism. We haven't quite decided it's worth the friction yet, but there may come a day. Shopify still does checkout pretty darn well, and I'm hesitant to take all of that onto our own shoulders.

The thing I still can't quite believe is what happened on search. I wasn't even trying to optimize the crop pages for SEO. But the site was so performant, and making well-built content was so easy, that one day I realized we were capturing two percent of all Google searches for high-value keywords like the word beet. Our tiny little farm was stealing thunder from websites with millions of users, because it had been built so well. Once I saw what was happening I leaned into it, and our little site now pulls over 4,000 organic visits a month on what is largely text-only content. There's still a bunch of meat left on those bones.

Talk to CSA farmers and you hear the same friction stories, confused members, missed pickups, an inbox full of the same five questions. We don't have those stories to trade, and it isn't luck. The site answered the questions before anyone had to email them. I've said, and I'll defend it, that this is a better website than most $5M companies run, SEO optimized, best practices across the board, for a tiny farm in Woodinville. But the claim I care about more is quieter. We held off on running a CSA for years, until we were sure we could deliver a great experience in the field and on the screen. This site is half of what made that possible.