Super Enterprises
Super Enterprises is a traditional wholesale business that had grown successfully over nearly three decades using a legacy ERP and largely manual workflows. Over eighteen months, I led its digital transformation—not by replacing existing systems, but by understanding how information moved through the business and designing software that fit those operational realities.
The result was an integrated operational platform spanning sales, inventory, receivables, analytics, and product data, giving management far greater visibility into the business while remaining usable by a largely non-technical workforce.
Selected technical highlights
- ERP ↔ Django web platform integration
- Entity resolution & data standardization pipelines
- Human-in-the-loop image annotation workflow
- Receivables & dispatch workflow automation
- Inventory ageing & operational analytics
Tech Stack
Python, Django, SQL, JavaScript, AWS, Pandas
Disclosure: This business is owned and run by my family
About Super Enterprises
Super Enterprises is a wholesale distributor of infant garments and accessories based in Ghaffar Market, Karol Bagh, New Delhi. Founded in 1997, it supplies over 1,000 retailers across North India, offering a wide range of products sourced from multiple domestic brands.
The Problem
I joined to implement a point-of-sale system. The ERP already included one, so I expected the project to take only a few days.
Instead, it revealed a deeper problem. Information critical to running the business was fragmented across the ERP, WhatsApp conversations, paper records, and employees’ tacit knowledge. Inventory, product images, receivables, and customer interactions all existed, but not in a form that supported efficient operations or informed decision-making.
What began as a software implementation therefore evolved into a broader effort to redesign how information moved through the organisation—building systems that integrated with existing workflows while making them more visible, measurable, and easier to act upon.
Operational Platform Architecture
Legacy ERP
MARG
Central Web Platform
Django · PostgreSQL · AWS
Internal Operations
- Orders
- Inventory
- Receivables
Client Portal
- Invoice Access
- Dispatch Tracking
- Communication
Operational Intelligence
Credit · Cash Flow · Sales · Geography · Dead Stock
Product Catalogue Pipeline
Key Contributions
1. Web App
Although the ERP stored inventory, orders, and customer records, interacting with it required technical familiarity and was largely limited to a handful of operators. As a result, routine tasks—checking stock availability, creating sales orders, or reviewing customer history—often depended on relaying information through intermediaries.
Rather than replacing the ERP, I designed a web platform that extended it. Built around existing workflows, the platform exposed inventory, order processing, receivables, and dispatch information through interfaces simple enough to be used on a smartphone by sales and dispatch staff.
This shifted operational information from being concentrated among a few ERP operators to being directly accessible across the organisation, reducing bottlenecks while preserving the ERP as the underlying system of record.
As the platform matured, it gradually became the operational hub of the business. Beyond internal workflows, it introduced customer-facing capabilities such as invoice access, dispatch tracking, and automated order communication, while also supporting receivables management through collection reminders and interaction histories.
By the end of the project, the application was used daily by sales and dispatch teams as the primary interface to operational information, with the legacy ERP continuing to operate in the background.
2. Digital Catalogue
With sales accelerating online after the COVID-19 lockdowns, managing digital sales of over 8,000 SKUs posed a unique set of challenges for this wholesale business.
Initially, online sales ran through WhatsApp - via broadcast lists and customer groups. Factories sent product images that were forwarded to clients periodically or shared in response to inquiries. The result was a web of inefficiencies: when new images arrived, older ones became impossible to track; salespeople couldn’t tell which items were still in stock, or how many units remained.
To address this, I designed a workflow that combined software with human judgement. The objective was not simply to build a catalogue, but to create a sustainable process for continuously linking incoming product images to the underlying inventory database as the catalogue evolved.
Annotation Pipeline
- Factory images uploaded in bulk through ZIP archives.
- Automatic duplicate detection, image compression, and cloud storage.
- A web-based annotation interface that presented unlinked images and allowed staff to associate them with one or more inventory items.
- Continuous integration of newly annotated products into the searchable catalogue.
The resulting annotation pipeline enabled any staff member to contribute to catalogue maintenance with minimal training, gradually building a structured, image-linked product database without disrupting existing operations.
This transformed how the sales team worked. Instead of searching through message histories or requesting photographs from colleagues, they could browse a searchable catalogue connected to live inventory, respond to customer enquiries immediately, and enable clients to place orders on their own.
3. Finance and Strategy
Years of financial records already existed, but they were difficult to analyse. Clients frequently appeared under different ledger names across accounting periods, addresses were incomplete, and historical transactions could not be reliably linked to the same business. As a result, straightforward management questions often required manual investigation.
I began by constructing a cleaner data foundation. Using entity resolution techniques, I reconciled fragmented ledger records into a consistent client database, while enriching them with standardized geographic information. This transformed transactional records into a longitudinal dataset that could support analysis across customers, regions, and time.
The resulting datasets supported several operational initiatives, including structured credit limits, receivables management, geographic expansion analysis, and inventory ageing reports that identified slow-moving stock.
These insights were subsequently embedded into day-to-day operations through structured credit limits, redesigned receivables workflows, quarterly sales incentives, and geographic planning. Rather than producing one-off analyses, the objective was to integrate better information directly into routine business decisions.
4. Understanding the Business
The technical work was only one part of the project. Before any system could be redesigned, I first had to understand how the business actually operated.
Much of this knowledge was never formally documented. Critical decisions—extending credit, prioritising dispatches, handling payment delays, or introducing new products—relied on routines and judgement that employees had developed over years of experience. Understanding these processes meant spending time on the shop floor, observing day-to-day operations, asking detailed questions, and gradually reconstructing how information moved between people, teams, and the ERP.
This process of discovery shaped every system that followed. Rather than imposing new workflows, I designed software around existing operational realities, introducing structure only where it reduced friction or improved visibility. The documentation produced during this process became a shared reference for onboarding, process improvement, and future development.
As new systems were introduced, I also worked closely with sales, dispatch, and accounts teams to refine workflows, incorporate feedback, and encourage adoption. Many of the most important improvements came not from writing software, but from understanding why existing processes had evolved the way they had.