A surf forecast for Montañita. An encebollados stand in Cuenca run by a Canadian who moved here in 2017 and fell for the dish. A barbershop. A bubble-tea brand. A Mexican cantina, a menswear atelier, a car dealership, a shoe label. Over the past year I've designed and built around a dozen sites for small businesses in Ecuador, and almost every instinct I carried over from building products and apps turned out to be wrong for them. Designing for a venture-backed product and designing for a marisquería are not the same craft wearing different clothes. They are different crafts. Here is what the small businesses taught me.

The website isn't the storefront — WhatsApp is

The single biggest mental adjustment: for most of these businesses, the website is not where the transaction happens. It is a landing strip that exists to hand the customer to WhatsApp, to Google Maps, or to a phone call, as fast as possible. Nobody in Cuenca is going to fill out a multi-step contact form to book a haircut. They tap, a wa.me deep link opens with the message half-written, and they're talking to the barber. So the most important element on the page is usually a single button, and the highest-leverage engineering is making sure that button is thumb-reachable, obvious, and one tap from anywhere on the page. I've shipped sites where the entire conversion funnel is see the thing → open WhatsApp, and that was correct. Build the beautiful site, absolutely. But know that it is a bridge to a chat, not a destination.

Mobile-first is not a slogan here, it's the whole brief

In product work "mobile-first" is a responsive-design discipline. Here it's a description of reality: the traffic is overwhelmingly Android phones, frequently on mobile data, sometimes on a connection that charges by the megabyte. A heavy hero video or a 4 MB uncompressed photo isn't a Lighthouse ding — it's a customer who closes the tab before your menu loads. That constraint quietly dictates the stack. Static sites on a free Vercel tier. Aggressively optimised images. No framework weight that doesn't earn itself. The reward is a site that opens instantly on a mid-range phone on 3G outside a restaurant — which is the exact moment the customer decides whether to walk in.

Trust is shown, never stated

A small business lives or dies on trust, and in Ecuador that trust is deeply local and visual. Nobody is reassured by the words "quality service." They're reassured by the Google rating, the review count, a real photo of the actual food, the owner's face, and whether you're open right now. Encebollados del Gringo leads with its story — Josh, from Montreal, 4.8 stars, a thousand reviews, thirty-odd thousand followers — because that story is the trust. The barbershop shows the fade, the price, and the booking slot. The design job isn't to decorate; it's to surface the three or four proofs a local customer actually checks, and to mark them up properly: LocalBusiness schema, opening hours, a map, so the business shows up well the moment someone searches it by name.

The brief is a voice note

None of these projects arrived as a spec. They arrived as a WhatsApp voice note, a logo someone's cousin made, and a feeling. The actual first task is reverse-engineering the brand from how the owner talks about their own work — the way the encebollado guy talks about his broth, the way the tailor talks about a lapel. That's the part no product ticket ever taught me. You're not implementing a requirements doc; you're translating someone's pride in their craft into a visual language, and then showing it back to them so they recognise themselves in it. When a client says "yes, that's us," you did the research right — even though the research was a conversation, not a user-testing session.

One page that does its job beats ten that nobody updates

The temptation is to give every client the full treatment: about, services, blog, gallery, team, contact. It's the wrong gift. A small business owner is running a kitchen or a chair, not a CMS. A ten-page site they can't maintain rots in a month; the hours go stale, the menu is last season's, and stale is worse than small. So I design for the smallest honest surface: the thing, the proof, the hours, the button. If there's going to be ongoing content — a photographer's portfolio, a restaurant's changing menu — I reach for whatever the owner will actually keep up, even when that means WordPress instead of the stack I'd personally prefer. The best site is not the most impressive one; it's the one that's still accurate in six months.

It has to feel like there, not like a template

The fastest way to lose a local business is to hand it a site that could belong to anyone. Monarca needed the papel-picado swagger of a Mexican cantina. The encebollados stand needed to feel coastal and a little irreverent. Buono needed the restraint of a tailor who says the art of dressing well and means it. The barbershop needed to open with "something has changed" and back it up. Designing in the local visual language — the colours, the humour, the references a Cuencano reads instantly — is what makes a template-built web feel like a specific place with a specific person behind it. That specificity is the whole product.

What it taught me

The web most tutorials teach you to build is for venture-backed products with analytics teams and quarterly roadmaps. The web most of the actual world uses is this one: a phone, a data plan, a WhatsApp button, and a decision made in ten seconds on the sidewalk outside. Building for small businesses strips away everything you can hide behind. There's no budget for excess, no spec to blame, no second screen, and no patience for a slow load. What's left is the oldest design question there is — can this stranger trust you, quickly? — and answering it well, over and over, for a surf forecast and an encebollados stand and a barbershop, has made me a sharper designer than any product ever did.