Star Alliance Airline Soars with Alteryx Server
At a Glance
1000+
Workflows Governed
10
Departments Migrated
25
Alteryx Champions Trained
Overview
Service
Data Engineering & Infrastructure
Industry
Airlines
Stack
Machine Learning & Gen AI
Pouring Clarity into Sales: How ML Powers Category Management
Author(s)
Ethan Pirso
Analytics Consultant
Emery Dittmer
Analytics Manager
Technology Stack
The Challenge
A global beverage leader struggled to manage its vast and complex sales data across multiple markets and product lines. Without a centralized reporting system, identifying critical sales anomalies was cumbersome, leading to delayed insights and reactive decision-making. These inefficiencies resulted in stock imbalances, missed opportunities, and a slower response to market shifts.
The Solution
To tackle these challenges, our team developed a machine learning-powered Sales Anomaly Detection solution that identifies significant shifts in sales trends. By consolidating insights into a centralized report, the system provides a single source of truth, allowing the company to proactively address market changes, optimize strategies, and respond faster to emerging opportunities.
Impact
$430,000
in Sales Variance Identified
25-30%
in Brand Share Deviation Identified
50%+
Shifts in Market Share Identified
Stack
Our Client’s Context
Our client, a global leader in the beverage industry, manages a vast portfolio of products across multiple markets, processing thousands of transactions daily. However, their fragmented and reactive reporting process made it difficult to detect sudden shifts in sales trends, leaving decision-makers without a clear, centralized view of anomalies.
Without a streamlined approach to anomaly detection, the company struggled to identify unexpected changes in brand performance, resulting in missed opportunities, delayed responses to market fluctuations, and challenges in optimizing inventory and mitigating revenue losses.
Seeking a data-driven solution, they partnered with Compass Analytics to gain real-time visibility into sales anomalies. Leveraging Databricks for advanced data processing and Tableau for visualization, we transformed their sales monitoring process—consolidating insights into a single, reliable system and empowering teams to make faster, more strategic decisions.
The High Cost of Low Sales Visibility
Data Overload
With thousands of transactions flowing through diverse distribution channels, gaining a clear view of sudden sales shifts across markets and product lines was a challenge.
Inconsistent Reporting
The absence of centralized reporting and cross-region insights caused delays in detecting and addressing sales anomalies.
Missed Opportunities
Undetected demand spikes or sudden drops led to stockouts and revenue losses, as reactive planning replaced proactive decision-making.
Pouring Precision into Sales with Machine Learning
To tackle these challenges, the Compass Analytics team developed a solution with two key components:
- Change Point Detection: Using bin segmentation, this method identifies shifts in sales trends over time, pinpointing when significant changes occurred—whether due to a marketing campaign, distributor change, or other external factors.
- Anomaly Detection with Isolation Forest: This machine learning model detects outliers in sales data, flagging sudden spikes or dips that require further investigation. Its flexibility allows for fine-tuning at both the market and brand level, ensuring precise anomaly detection across different sales dimensions.
By utilizing Databricks for data processing and Tableau for visualization, we developed a user-friendly dashboard that provided stakeholders with daily visibility into sales anomalies, enabling swift and informed decision-making across the company.
Brewing Success: Turning Data into Actionable Insights
The implementation of the Sales Anomaly Detection solution delivered high-impact results, equipping the client with daily insights for smarter, faster decision-making. By detecting significant market shifts and financial outliers, the solution enabled strategic resource allocation, optimized promotional and distribution strategies, and minimized revenue losses. With enhanced visibility, proactive decision-making, and centralized insights, the company became more agile and responsive in an increasingly competitive market.
Market Shift Detection
The solution identified brand share shifts of up to 57%, providing critical insights into evolving consumer preferences and competitive pressures. Leveraging Change Point Detection with bin segmentation, it pinpointed when shifts occurred—whether due to marketing campaigns or distributor changes—allowing decision-makers to quickly adjust pricing, marketing, and distribution strategies.
Financial Outlier Detection
The model uncovered a $430K variance in sales and a 25.6% deviation in brand share, helping the company respond swiftly to financially significant anomalies. The Isolation Forest model detected outliers in sales data, flagging sudden spikes or dips that required immediate attention. With these insights, the company optimized promotions, adjusted distribution, and improved inventory management to minimize revenue losses and maximize growth opportunities.
Anomaly Tracking & Centralized Insights
By consolidating sales anomaly data into an interactive Tableau dashboard, the company gained real-time visibilityacross brands and regions. This centralized, user-friendly platform provided a single source of truth, improving communication, alignment, and efficiency across teams—empowering them to make data-driven sales decisions with confidence.
