Farmica

Executive summary

Farmica is a product I built for direct-to-consumer farms that outgrew phone calls, messages, and spreadsheets but do not want a generic webshop. It starts as a spletna naročilnica — a link or QR code customers already understand — and becomes the operational layer behind weekly offers, order cutoffs, stock by delivery day, and fulfillment lists.

I shipped it in private beta with a small number of paying farms while building the public story at farmica.si. The build proves I can design for a non-technical buyer, ship a real multi-tenant workflow, and sell with language the market actually uses.

The problem

A farm selling direct does not have a single order inbox. Orders arrive as Facebook comments, WhatsApp replies, SMS, email, market conversations, and phone calls. Before each delivery or pickup day, someone must merge all of that into one list — often in a notebook, Excel, or memory.

Three failure modes repeat every week. Collection chaos: orders get missed or duplicated across channels. Offer fatigue: the farm answers "what do you have this week?" on every channel separately. Fulfillment friction: routes, pickup points, and quantities are assembled manually the morning of the run.

Shopify and marketplaces solve a different problem. Retail assumes permanent catalog and standard shipping. Farms think in weekly availability, delivery days, cutoffs, and routes. Fighting retail tooling for that workflow is why many farms stay manual — and why "we don't want a webshop" is a rational objection, not stubbornness.

The thesis

The bet is that farms will adopt software if the entry point matches what they already say — spletna naročilnica — not "platform," "system," or "ecommerce." Under that familiar surface, the product organizes the work they already do: publish this week's offer once, collect orders from link, QR, and manual entry for phone/message orders, close at cutoff, allocate stock per delivery day, and export fulfillment lists by route or pickup point.

The business experiment is whether subscription revenue at farm scale (base package around €70/month, no setup fee on the narocilnica tier) can grow through operational dependence: narocilnica first, then upgrade to a fuller storefront when the farm is ready. The competitive set is not Shopify — it is the status quo of messages, calls, and paper lists.

Validation

What exists today is specific and limited — by design.

• farmica.si is live: trilingual marketing site (Slovenian, English, German), pricing and onboarding story, contact funnel, blog, and a product demo video on the landing page.
• 2–3 paying farms use the platform in private beta. Onboarding is deliberate, not open self-serve.
• Kmetija Mehak (kmetijamehak.si) is the reference deployment. They entered through the naročilnica concept; the relationship validated that farmers understand the hook immediately and that upsell to a broader storefront is a real path, not just positioning copy.
• Positioning check: the first client conversation confirmed that leading with "spletna naročilnica" beats abstract product language — documented in the positioning changelog, March 2026.
• GTM tooling in the repo (farm prospect scraping, outreach CSV generation) supports outbound sales; that is pipeline work, not product-market proof on its own.

What does not exist yet: a public self-serve launch, open-source code, or broad marketplace-style distribution.

What I built

Order platform (private beta)

• Weekly offer / naročilnica: farmer defines products and stock for the week, publishes a shareable link and QR entry point.
• Customer ordering flow: pick products and delivery or pickup day without app install or registration (positioning for the customer side).
• Farmer dashboard: centralized orders, order cutoffs, stock allocation by delivery day.
• Quick order entry: convert phone, WhatsApp, or Facebook message orders into structured orders in seconds (V1 channel strategy before Messenger/WhatsApp integrations).
• Fulfillment outputs: lists grouped by delivery route or pickup point; export to Excel or Google Sheets for the pre-delivery workflow.
• Notifications: customer and farmer alerts on new orders and weekly offers (one-click broadcast story across email, Facebook, WhatsApp text — as designed in product docs).
• Upgrade path: full storefront and catalog layer for farms that outgrow the narocilnica-only package (Mehak as case study).

Marketing site (farmica.si)

• Next.js 16 App Router site with React 19 and TypeScript.
• next-intl routing for sl, en, de with localePrefix: always.
• Core pages: home, farmer landing page, how it works, pricing, offer detail, audience fit, contact, blog, and marketplace comparison page.
• Contact form with Zod validation → Brevo transactional email (internal notification + farmer confirmation).
• Cookie consent gating Google Analytics 4 and Vercel Analytics.
• Markdown blog with remark pipeline, sitemap, robots, and JSON-LD helpers.
• Locale-specific YouTube demo embed on the landing hero.
• Deployed on Vercel.

Business and sales foundation

• Positioning, pricing, ICP, competitive responses, and page-level copy in separate business-foundation docs in the Farmica product repo.
• Outreach scripts and farm prospect datasets maintained alongside that repo.

Architecture

The system splits into two deliverables that share a brand but serve different jobs.

Marketing site. A single Next.js application on Vercel. All user-facing copy lives in locale JSON files; routes are locale-prefixed. Server actions handle the contact form; Brevo sends templated mail. Analytics load only after consent. The blog is file-based Markdown — no CMS dependency for posts — which keeps the marketing layer fast to change without touching the product codebase.

Order platform. Multi-tenant farm accounts each get their own naročilnica and optional storefront. The weekly workflow is the spine: create offer → set stock and fulfillment options → distribute link/QR → collect orders (self-serve + manual entry) → cutoff → fulfillment lists. Data model concepts are farm-native: delivery days, routes, pickup points, and stock tied to fulfillment dates rather than a permanent catalog SKU model.

Key technical choices on the marketing side: next-intl for real trilingual SEO and farmer trust in Slovenia with DACH expansion in mind; Brevo for reliable deliverability on a small-volume sales funnel; consent-first analytics because farm audiences are privacy-sensitive and EU-facing.

Current status

Farmica is shipped-private / private-beta. The marketing site is public. The ordering platform runs for a small set of paying farms with hands-on onboarding. I am actively selling into the Slovenian direct-to-consumer farm segment while refining packaging (Spletna naročilnica at €70/month as the entry tier, full platform as upgrade).

Not paused, but not "launched" in the sense of open signup or scaled distribution. Next steps to move forward: more reference farms like Mehak, clearer in-product demo for cold traffic, and eventually channel integrations (Messenger/WhatsApp) per the product roadmap — V1 manual quick entry already reduces notebook workflow.

What I learned

Lead with market language, not architecture. Farmers searched for a better ordering form, not software. Renaming the offer to spletna naročilnica and pushing "platform" below the fold came from a real client reaction, not theory.

The objection "we don't want a webshop" is data. It forced a product story where channels stay the same and structure appears behind the scenes. That framing applies beyond agriculture — any buyer with an entrenched manual workflow needs the familiar surface first.

Done-for-you onboarding is part of the product. Farms will not configure delivery routes and weekly stock alone on day one. Setup fee vs. no setup fee, seasonal pause pricing, and "we build from your materials" are commercial decisions tied directly to adoption.

Separate marketing velocity from platform velocity. A Next.js marketing repo can iterate copy, locales, and funnel experiments daily while the tenant platform moves slower. Keeping them decoupled avoided blocking sales on refactors.

Reference customers beat feature lists. Mehak on the site does more than any bullet about fulfillment lists. One honest deployment with a named farm beats generic SaaS proof.