Client
Pottsalat GmbH
Unit
prodct [CX] · /skills · [Labs]
Duration
12 months · 2021–22
Shape
Headless platform · ground-up build
Headline
Storefront, apps, checkout, subscriptions, order engine, delivery + route planning — one event-driven platform Pottsalat owns end-to-end.
// prodct [CX] + /skills + [Labs] case · Pottsalat · 2021–22 · food delivery · headless e-commerce

WooCommerce hit the ceiling.
So we built the platform.

Pottsalat had outgrown WooCommerce — business logic it couldn't carry, checkout friction, performance that capped growth, no real path to scale. We made the ambitious bet: a ground-up, headless, event-driven platform on a Python backend with React + React Native frontends — storefront, apps, checkout, subscriptions, order engine, inventory, delivery management, route planning. All self-built. All owned by Pottsalat.

1 platform
// headless, event-driven, Python + React / React Native
// self-owned
AOV
// higher basket value post-launch, sustained
// customer-visible
churn
// subscription retention up, time-to-order down
// direct effect
// 01 / the situation

WooCommerce couldn't carry the business.

Pottsalat had outgrown the stack. WooCommerce couldn't model the business logic — delivery windows, subscriptions, per-zone availability, dynamic menus. Checkout had too much friction. Performance capped the growth curve before the growth curve even started.

Scaling on the existing platform meant stacking plugins on plugins. We stepped back and made a different call: own the platform. Build for Pottsalat's actual model, not a shop template.

  • 01

    Business logic the platform couldn't model.

    Windows, subscriptions, regional availability — forced into plugin workarounds.

  • 02

    Checkout friction that cost orders.

    Load time, flow breaks, form-state drift — measurable drop-offs at the last step.

  • 03

    Performance didn't fit the ambition.

    Every traffic spike hit the ceiling. No room to grow the catalogue or the markets.

// 02 / approach

Headless. Event-driven. Self-built.

Python backend. React on web, React Native on mobile. Every business action — order placed, window closing, route assigned, subscription renewed — is an event on the bus. Services listen, compose, react. No monolith, no plugin stacking.

The model is deliberate: own the logic that defines the business, use best-of-breed for the rest. We built the commerce spine. We didn't rebuild what the market already solves.

  • A

    Headless architecture.

    Frontends and backends decoupled. Storefront, app, admin — all independently deployable surfaces.

  • B

    Event-driven core.

    One bus. All state transitions are events. New capabilities subscribe without reshaping the core.

  • C

    Python + React / React Native.

    Pragmatic stack, strong typing at the edges, hireable tech — Pottsalat can keep building without us.

// 03 / what we built

Every surface of the business, shipped.

[Storefront]
Web storefront + mobile apps
React web, React Native iOS/Android. Shared design language, independent release cadence, performant on real networks.
[Checkout]
Custom checkout flow
Window picker, address validation, payment, subscription sign-up — one flow, built for Pottsalat's model, not bolted on.
[Engine]
Subscription engine
Plans, pauses, skips, renewals, partial swaps. Customers run their own rhythm — ops see it aggregated for planning.
[Engine]
Order engine
Order lifecycle, state machine, SLA tracking. Handles single orders and subscription-generated orders on the same rails.
[Admin]
Internal admin + inventory
Menu, pricing, stock, product data, window capacity. Ops configures the business without a ticket.
[Ops]
Delivery management + route planning
Delivery windows, capacity model, sequencing, dispatch. Fewer trucks chasing the same postcodes.
// 04 / outcomes

The bet paid off.

performance
// page speed, time-to-order, checkout completion
// measurable
AOV
// higher basket value, sustained post-launch
// directional
churn
// subscription retention up, time-to-order down
// direct effect
orders / window
// delivery capacity model lifted throughput per window
// ops-visible
planning KPIs
// forecast quality, window fill, route density all up
// event data pays off
Owned end-to-end
// Pottsalat builds on top, on their own cadence
// no vendor lock-in
// similar engagement?

Outgrown your commerce platform? We've built the replacement.