Data Nexus

Merchy

A Multi-Vertical Transactional SaaS Platform for Telegram WebApps

Data Nexus case study: architecting a multi-vertical transactional SaaS platform for Telegram WebApps with a unified, state-driven transaction core.

The Merchy platform on a desktop display: the sign-in screen beside phone screens for its clothing, car, food and manicure verticals.
Record
Client
Merchy
Year
2023
Duration
6 months
Scope
SaaS, E-commerce, CRM, WebApp, UX/UI

Invariant

Every vertical is a specialisation of one transaction sequence. A new business line changes the data and the rules, never the structure.

01/Context

Merchy is a transactional SaaS platform built specifically for Telegram WebApps, enabling businesses to sell products, services, and bookings directly inside Telegram bots — without standalone websites or native applications.

The platform was conceived as a universal transaction engine, capable of supporting fundamentally different business models (retail, food ordering, services, bookings) on top of a single architectural core.

Rather than being a collection of domain-specific interfaces, Merchy is a system-driven product where user flows, pricing logic, scheduling, fulfillment, and post-order interactions are expressed as variations of the same transaction model.

The Merchy seller dashboard on a laptop — weekly revenue, cost of goods, item count and user counts — beside phone screens of a watch catalogue and a product page. The figures on the dashboard are demonstration data, not a client’s results.
02/Diagnosis

Telegram WebApps impose strong structural constraints:

  • mobile-first environment with limited navigation patterns

  • high sensitivity to cognitive overload

  • fragmented user attention

  • absence of conventional web UX affordances


At the same time, Merchy needed to support:

  • complex product configuration (variants, add-ons, discounts)

  • time-based services and bookings

  • fulfillment logic (delivery, pickup, scheduling)

  • analytics and operational control for merchants

  • scalability across multiple business verticals


The core challenge was not interface complexity, but systemic complexity.

Without a strong architectural foundation, each new business scenario risked:

  • creating new screens

  • duplicating logic

  • increasing development and UX debt

  • breaking product coherence


Eight phone screens of Merchy running inside Telegram, one vertical each: watches priced in dirhams, cars and motorcycles, a food menu, manicure booking, a dry-cleaning checkout, address selection on a Dubai map, courier tracking in chat, and a car-rental date picker.
03/Intervention

Data Nexus approached Merchy as a product architecture problem, not a UI project.



1. Canonical Transaction Model

Instead of designing domain-specific flows, Data Nexus defined a universal transaction sequence that acts as the invariant core of the platform:

Browse → Select → Configure → Schedule → Fulfillment → Payment → Post-Order


All supported verticals — retail, food, services, bookings — are specializations of this same transaction graph, differing only in domain rules and data, not in structural logic.


2. State-Driven UX Architecture

Merchy was designed around states, not screens.

Key principles:

  • variations are modeled as state changes, not navigation branches

  • configuration flows are nested states, not new pages

  • filters and options are overlays, not separate interfaces

  • post-order interactions are continuations of the transaction

This approach:

  • reduced UX and development duplication

  • prevented interface explosion

  • preserved cognitive clarity for end users


3. Multi-Vertical Scalability on a Single Core

Using the same transaction engine, Merchy supports:

  • Retail & E-commerce

    catalogs, filtering, variants, discounts

  • Food Ordering

    menus, ingredients, add-ons, cart, delivery tracking

  • Services & Bookings

    scheduling, time slots, duration, capacity management

All verticals reuse:

  • the same interaction patterns

  • the same state logic

  • the same transaction semantics

Only domain-specific constraints and content change.


4. Enterprise-Grade Operational Layer

In parallel with the Telegram WebApp, Data Nexus designed a full SaaS admin platform functioning as a merchant operating system.

Capabilities include:

  • product and service management

  • pricing and discount configuration

  • order and fulfillment control

  • subscription and tariff management

  • advanced analytics (revenue, cost, conversion, behavior)

This ensured that Merchy was not only usable, but operationally scalable.

5. Architecture Built for Growth

By abstracting logic into:

  • transaction states

  • reusable interaction modules

  • domain-agnostic components

Merchy can expand to new business models without redesigning its core system.

The platform scales through structure, not through interface multiplication.

Outcome
  • A unified transactional SaaS platform operating entirely inside Telegram

  • Support for multiple business verticals on a single architecture

  • Reduced UX and technical debt through state-driven design

  • Enterprise-ready admin and analytics infrastructure

  • Strong foundation for white-label, API, and enterprise deployment

This case demonstrates Data Nexus’s ability to design systems that remain coherent under complexity, and to transform fragmented requirements into scalable SaaS architectures.

Eight phone screens tracing one Merchy Food order: a dish carrying a “Contains lactose” badge, priced ingredient add-ons, the menu, payment confirmed in chat with the delivery address, courier tracking on a map, checkout, and the order summary.
Eight phone screens tracing one Merchy Clean order end to end: catalogue, categories, the item, cleaning speed, completion date, ready time, delivery method on a map, and checkout.
Eight phone screens tracing one Merchy Beauty booking: manicure styles with duration and price, a choice between the salon and a home visit that adds to the total, address on a map, an appointment slot picker, payment method, checkout and the order summary.
Eight phone screens of the Merchy watch vertical: a catalogue priced in dirhams, collections by model, a product page with a case-size selector, featured brands, an empty shopping bag with its own explanation, and the bag with one watch in it.
04/Outcome

Merchy was delivered as a coherent, scalable transactional SaaS platform, capable of supporting multiple business models inside Telegram WebApps without fragmenting UX or system logic.


Architectural Results
  • A single, canonical transaction core reused across retail, food ordering, services, and bookings

  • State-driven UX architecture, reducing dependency on screen-level branching

  • Clear separation between domain logic and interaction logic, enabling controlled product growth

  • Platform extensibility without redesigning core flows or interfaces


Product & Operational Impact
  • Reduced UX and system complexity by consolidating multiple verticals into a unified transaction model

  • Faster time-to-market for new business scenarios and verticals

  • Lower long-term design and development overhead due to reusable states and interaction patterns

  • Enterprise-ready admin system enabling merchants to operate, analyze, and scale their businesses from a single control layer


Strategic Outcome


Merchy evolved from an initial e-commerce concept into a multi-vertical transactional platform, positioned for:


  • white-label deployment

  • enterprise clients

  • API-driven integrations

  • further expansion beyond Telegram


The result is not a collection of interfaces, but a transactional system that remains structurally stable under growth and complexity.

–40%
UX & flow complexity via state-driven architecture
–60%
Time-to-launch for new business verticals
4+
Business verticals on a single transaction core

The first two are engineering estimates against the implementation that preceded the rebuild, not instrumented measurements, and they are stated here as estimates. The count of verticals on one transaction core is a count, and it is the one figure in this record a reader can check.

The work behind it

A method page can argue anything. This is the measurement it argues from, which is why the two are linked in both directions rather than one.

Platform and transactional architecture
The domain model is the product. Everything above it is an interface to it.

Custom software development and AI engineering for Dubai
A probabilistic component is only safe inside a deterministic boundary.

UX, UI and user flows
Model states, don’t multiply screens.

From idea to a product that exists
Find the assumption that kills it, and test that one first.
Built with

Each names what it is good for and the point at which it stops working. This record is the evidence those pages cite, which is why the link runs in both directions.

Telegram Bot API
Telegram's programmatic interface: bots that hold a conversation, and web apps that render a full interface inside the chat.
Next

What made this record possible was a diagnosis before a proposal. That is the same first step whatever the system is, and it is the one thing worth asking for now.