Stock the customer can trust
Availability is checked at the cart, so the store stops selling what the shelf no longer has.
Retail · Online store and order operations
A storefront and an operations dashboard built as one product, so every order carries its own stock, promotion and fulfilment status instead of being reconciled by hand.
The operational challenge
Selling online is rarely limited by the shop window. The costly work begins after checkout: keeping stock honest across variants, applying promotions without breaking prices, and knowing which orders are new, packed, shipped or still unpaid without asking someone.
Leadership and scope: Mihail owned product definition, architecture and delivery across the customer storefront and the internal dashboard, so the two were designed against the same record rather than integrated afterwards.
The delivered system
How the work moves
What it changes day to day
Availability is checked at the cart, so the store stops selling what the shelf no longer has.
Discount rules are validated by the system rather than applied by hand at checkout.
Products, variants, categories, media and content are edited in the dashboard without a release.
Why both halves were one project
When the storefront and the back office are separate products, someone becomes the integration. Stock is retyped, promotions are tracked in a side document and the answer to “where is this order?” lives in a person rather than a system.
Defining the catalogue, the stock rules and the order states once meant the customer view and the operational view could never disagree about the same product or the same order.
Built to be run by the owner
The dashboard was shaped around the tasks that actually recur: adding a product with its variants, correcting stock, launching a promotion, and moving orders through their states.
Media, categories and static content are editable without a developer, which keeps the ongoing cost of the store with the business rather than with the person who built it.
What changed operationally
shared by the storefront, stock and order pipeline
catalogue, promotions and content, without developer involvement
for every order from new to paid, with its own history
Published at the client’s request without naming the business. Product data, order volumes and commercial figures remain private.
What can be verified
Retail assortment, supplier terms and sales figures are commercially sensitive, so this case study describes the system and the operating model rather than the numbers behind them.
The useful evidence for a similar business is structural: how variants are modelled, where stock is authoritative, how a promotion is validated and which order states the team actually needs.
When the shop works but the back office does not