Transorg Analytics

Overview
The client’s earlier approach of centralized inventory planning missed out geographical insights of demand thus incurring sale loss & customer churn.

The objective was to come-up with a framework that allows planning for assortment and customer offers at a localized level, hence improving customer retention and thereby revenue.

Solution
The process involves systematic analysis of data at various levels like customer, product and geography level information of sales patterns.
Store/City level insights helped in better planning of assortment and combo offerings suited to local demand.
Created reporting framework to understand weekly and monthly purchases at customer and product level.
Geography level analysis helped to give a view of location wise top selling products in different months
Product category heatmap helped to analyze MoM performance across different geographies & timeframes.
Combining product and geography data with customers, we categorize the customers in different tiers and identified loyal base.

Output
Key Impacts

Our localized and seasonal recommendations resulted into ~20% increase in combo order count and improved inventory planning.
Improved the personalization experience of loyal customers
Increased customer retention and new customers count.

Overview
Client is a leading low-cost carrier airline embarking on a journey of excellence in customer service. Consistently ranked as one of the most punctual airlines globally.

Solution
CX-LLM is an advanced GenAI chatbot tailored for seamless ticket booking and social media integration, revolutionizing user interactions in the realm of customer service.

Client required a CX-LLM chatbot that addresses user inquiries and booking activities with a focus on general questions and ticket booking tasks, providing a comprehensive solution for user engagement
The chatbot utilizes various sources of data, including user inputs, social media interactions, and backend systems such as Navitaire (airline management system) API, to generate accurate responses and facilitate bookings.
The chatbot’s model undergoes periodic updates to incorporate improvements and adapt to evolving user needs and preferences
Data preprocessing and transformation techniques are applied to input data to ensure compatibility with the chatbot’s natural language understanding and response generation capabilities.

Output
Key Impacts

CX-LLM delivers prompt responses to user queries, facilitates ticket bookings seamlessly, and integrates with social media platforms for enhanced accessibility and convenience
TransorgGPT was the preferred solution among models like OpenAI’s LLM, to understand user queries, generate responses, and facilitate ticket bookings with efficiency and accuracy.

Overview
Our client, world’s first performance branding company collapses the silos between performance and brand to unify marketing objectives, targets, & strategy

Solution
Transorg Analytics adeveloped a market mix model and a KPI forecaster (both embedded in one end to end framework) on top of it by using advanced analytics and machine learning techniques covering at least 8 brands across US, UK and Canada regions to provide granular level insights at the ‘Country X Brand X Sub-brand X level and at the ‘Channel (platform plus audience) X’ level.

Models are refreshed weekly to remain calibrated with updated spend budgets and most recent actuals.

For each brand the input data used for building the models included:

Channel wise spend on a weekly basis.
Weekly total business revenue
Promotional Periods
Monthly spend budgets.
Monthly revenue goals (just for comparison)
Modelling (MMM)

Analyse historical spend distributions and their contribution towards total business revenue.
Manual weights can be assigned towards specific channels and months based on previous year data.
Output the spend recommendations based on MMM model on a monthly/daily level.
Modelling (Forecasting)

Take the spend recommendations of MMM model as an input.
Generate weekly revenue forecasts with in-built cross validation.
Interpolate forecasts into monthly and daily level.
Output
Daily/Monthly Spend Recommendations
KPI (like revenue) forecasts daily/monthly on spend recommendations

Overview
Our client, Global 500 FMCG company wanted to improve reporting structure using Power BI as the base visualization tool with inclusion of Power BI Report Builder tool and other Power Apps working in conjecture for various reporting needs.

Transorg Developed methodologies for report creation, scheduling, and bursting using Power Automate, Report Builder, Online Services and Subscriptions.

Refined the logic of multiple metrics to bolster performance efficiency.
Utilized Tabular Editor for dynamic measure creation.
Reporting Migration to Power BI platform from legacy SAP BO for Sales Department in APAC
Change Management with training of users and supporting in adapting the new platform.
Created 2 data models in Power BI utilizing Sales, Stock, Scheme and Hierarchal data as the base for the reporting structure.
Access Management done utilizing Azure Active Directory.

Solution
Input Data

Primary Sales (SAP BW ~ 2 tables)
Secondary Sales (Azure SQL Server ~ 1 table)
Stock (Azure SQL Server ~ 2 tables)
Scheme and Scheme Budget (Azure SQL Server ~ 5 tables)
Masters – Distributors, Retailers, Products, Calendar (Azure SQL Server ~ 7 tables)
Manual user data integration for applying RLS or including target files through SharePoint Excel files and SharePoint Lists

Transform

Designed 32 SQL views by applying transformations and implementing necessary filters to maintain data integrity in Azure SQL Server.
Resolved Many to Many Cardinality with respect to relationships among SQL Views.
Replicated ~75 BO and Time Dimension measures in Power BI Desktop.
Fixed all number mismatch between SAP BO and Power BI.
Enhanced data pipelines and optimized views to align with specific requirements

Output
Used Power Automate to Schedule / Burst (RLS) reports.
Conducted training to 100+ users to facilitate scheduling of 860+ reports which were scheduled through SAP BO

Overview
Our client, one of the world’s largest food and beverage companies, wanted to build a retailer recommendation engine which looks at the historical order placement data and suggests products for the next month.

The objective is to provide retailer-level recommendations to the company by unlocking key patterns in the historical purchase order at channel x retailer x region/district level.

Solution
Data Pre-Processing and Feature Engineering

Optimization for unique hierarchy ID assignment in combined databases.
Data cleaning and validation.
Focusing on the top 90% retailers and products based on revenue.
User-item matrixas implicit feedback derived from transactional data.
Features based clustering.

Retailer segmentation with comparable purchasing behaviors based on their sales channels.
Algorithms used such as RFMand KNN.
Segmentation based on revenue such as:
Premium
Medium
Occasiona
Recommendation Engine

Implemented product-based recommendations for Retailers through Collaborative Filtering techniques via ML Models like Neural Collaborative Filtering.
Applied Association Rules to find popular items and co-purchases via Apriori Algorithm.
User Interface

Updated SQL table with tailored recommendations post model training for consumption over web and mobile app for Distributors and Retailers.
Output
Projected improvement in revenue by ~3%
brand consumption improvement from 4 to 6
Strategically upselling new product variations by increasing product awareness through recommendations.
Improve Precision: & accuracy by 80%