Think Big - Large Scale

Analytics Business Suite i3v7 is geared to handle large data sets in record processing time. Historically, forecasting, like any other data mining applications must handle large sets of data. A typical forecast will align SKUs, Stores, and Hierarchies on a daily basis that will explode into managing billions of data sets. Example of a typical retailer having a forecast project for:

• 150,000 SKU,
• 5000 Stores,
• 100 Segments,
• 4 time frame projections (next day, next week, next month, and next year).

Providing related forecast information will result into 3 Billion records. Analytics Business Suite i3v7 will produce forecast data for such a project set in less than 5 hours, making the information immediately available to the business and shortening their reaction time.

Models can often be more complex, and aggregation can be done at Product Category, and Subcategory levels, by using Store Clusters performing in optimised fraction of time.

Open for Business

In the new release, ibs AMS is no longer working exclusively with ibs Application Suite: It can be accessed as a service and used as an open system that can be easily integrated with non Soft Solutions applications by using Smart Data Connectors or staging tables that pulls data from flat transaction files, POS Logs, relational or non-relational data warehouses.

Ibs AMS was built exclusively and entirely by Soft Solutions Team and interfaces with all applications of the IBS Suite.

Ibs AMS within ibs presents many advantages such a seamless integration with other modules e.g Pricing, Allocation, Promotions, Category Managers.

Forecast results are provided directly into these applications in the appropriate context.

As cascade results and thanks to embedded optimisation engines, Analytics Business Suite i3v7 has proven tangible results in various business domains:

• DC margin optimisation from 1.5% to 3% on everyday goods categories
• Store margin optimisation from 0.5% to 2% on everyday goods categories
• " Allocation & Replenishment planning optimisation with a drop of 2% of overstocks

ibs Analytics - Forecast engine

The Engine main forecast method is based on the "Advanced Data Decomposition Model" (ADDM) internally developed by Soft Solutions Science Team. The model uses a number of algorithms using outlier detection and normalization coupled with unique projections and extrapolation of cyclical trends, seasonal effects, and promotional lifts.

The forecast engine is not limited to the ADDM method and will automatically select alternative methods suitable for products with different properties, velocities, behaviours and life cycles. This advanced feature newly introduced in Analytics Business Suite i3v7 provides a full meta-engine, defining automatically the patterns of sales of any given item, and forecasting each with the most appropriate method.

The i3v7 Engine automatically selects one of the alternative algorithms such as Exponential Smoothing, Auto Regressive Integrated Moving Average (ARIMA), Elasticity models and Product Life Cycle Analogy to produce the optimum accuracy for specific products.

This new approach allows a fast implementation and adjustment of the engine to any situation, adding or removing new methods to it without changing the architecture or introducing any risk in accuracy decrease.

Based on this new architecture, Analytics Business Suite i3v7 achieves high rates of accuracy:

• Average accuracy rate of regular and promotional sales over 95% for non seasonal segment level forecast
• Average accuracy rate of regular and promotional sales over 90% for seasonal segment level forecast
• Average accuracy rate of regular and promotional sales over 75% for regularly carried items
• Average accuracy rate of regular and promotional sales over 70% for slow moving items

ibs AMS - Forecast visualization & Optimization

Supporting the Analytics engine, ibs AMS - Forecast visualization & Optimisation service allows to monitor and enhance results of the forecast engine:

• Accuracy Monitoring:
  • Constant forecast comparisons with the respective observed values
  • Generation of accuracy (distance) and bias (directions) ratios at different aggregation level (category, Sku...) with standards of industry (MAPE, WMAPE, MAE, MAD...)
  • Alerts and notifications sent when accuracy levels are below expected ranges
  • Proposals of corrective measures to improve the accuracy levels for specific data subsets.
  
  
• Forecast Visualization Portal:
  • AMS Data is accessed using different applications within the suite, or via the Forecast Visualization Portal. It is used to view forecast data directly and perform simulations and complex "What If" scenarios.
  • Changing a combination of parameters such as price, elasticity, trends, seasonal peaks, method of calculation, does simulation.
  • Simulations can be stored and monitored over time and eventually used to execute corrective measures or to study the behaviour of a group of products.
  • ibs AMS has an impressive number of reports and charts helping to visualize the forecast data sets.
  
  
• Optimisation & Simulation
  • Regular Price optimisation.
    • Based on forecast decomposed models (retail price elasticity in particular) and joined to ibs Pricing service, ibs AMS provides optimized Pricing policies, allowing a better competitor position without sacrificing any margin point
  • Markdown optimisation.
    • ibs AMS is also able to determine the best rate off by markdown wave to optimize costs or sell the full remaining stock at a given date.
  • Assortment optimisation.
    • The combination of forecast model and cross-elasticity (halo & cannibalisation effects) and ibs assortment service ables AMS to determine the best set of items to be included in the assortments based on given objectives.
  
  
• Allocation & Ordering optimisation
  • Based on store level forecast and considering supply chain information provided by ibs Referential and ibs Central, ibs AMS can also optimise the allocated and ordered quantities, taking into consideration, required service rate, security stock, TIL, BOH...

ibs KPI - Universal Dashboard

Analytics Business Suite i3v7 is equipped with an intuitive Dashboard that enables the user to monitor different indicators of the application such as the accuracy levels, the behaviour of certain products or group of products. The Dashboards can also group information in logical subset that can be triggered by action or events that users are monitoring.

The Dashboard can be used to drill down into different levels from Categories to single SKU and ultimately view any cross section of the data.

The Analytics Business Suite i3v7 Dashboard can also generate complex view of the accuracy and keep track of the performance of the tool while monitoring the corrective measures. Soft Solutions has deployed its own indexing and storage techniques to perform real time reporting and consolidations on given KPI in a configurable and evolving scheme:

• Dedicated information management - Our approach combines high performances of indexes of databases servers with our own technology of the information to strengthen the volume of data
• KPI Information Model
  • Our approach includes a descriptive layer of the Key Performance Indicators, providing the ability to embed definition, units and calculation formulas within the data storage and ensure 100% information integrity. It also provides the ability to configure additional KPIs based on evolving needs of our clients