How analytics helped retail grocery giant Haggen

Haggen Inc. is a rapidly growing US food and grocery retailer based in Bellingham, WA and operates 32 supermarkets throughout Washington and Oregon states. Haggen employs over 3, 700 people and operates two store formats, Haggen food & pharmacy and Top food & drug.

Challenges at Haggen
Haggen was facing a growing need for data-driven reporting and analysis to keep up with the decision-making needs of its business managers. Haggen's manual, unstructured and develop-as-you-go reporting culture was proving to be a major constraint in their growth plans.

The information needed for decision-making was scattered over various source systems like JD Edwards, ISS POS, and a host of homegrown systems covering planning, finance, loyalty and HR functions. To generate cross functional analysis and reporting, the company's business was spending an increasing amount of time and energy generating data and reports using slow, time-consuming and cumbersome methods. Average weekly report packs took more than 24 hours to prepare and deliver to decision makers. It was hard to maintain and manage data quality at Haggen's manual reporting environment that needed complex integration and transformation.

Complex merchandise and category performance analysis was impossible to report because of multiple levels of granularity (department, category, sub-category and SKU levels), multi-dimensional nature (product, brand, store and time), and the number of source systems that needed to be connected. Special reports like volume reports could not be drilled or filtered. Haggen runs a number of promotions at frequent intervals and it was a challenging task to compare pre-promotion, during promotion and post promotion sales. Haggen had significant overheads on IT resources for extraction and transformation of data for generating analytics and presenting such reports.

Search for a silver bullet
Haggen discovered that a lack of single, integrated data warehouse, the right BI toolsets and retail analytics were missing elements in their BI strategy. After a few rounds of evaluation, Haggen realized that the available generic BI options were prohibitive in cost and time when it came to customizing those platforms to meet their requirements. The development lifecycle of a generic BI implementation is usually a prolonged, intensive effort and takes typically anywhere between 12-18 months. In Haggen's competitive usiness environment, the availability-usability timeframe, and scalability for BI was a critical factor.

Manthan Systems and Teradata jointly implemented ARC - Manthan's integrated retail BI suite on the Teradata platform. The solution encompassed an enterprise retail data model, and cutting-edge BI functionality that spanned usage across store operations, category management, inventory and promotions. This unique deployment gave Haggen the high performance, high availability capability of Teradata coupled with the retail centricity of the ARC solution. The solution is currently being extended to HR, Payroll and Loss Prevention Analytics.

The benefit

An enterprise-class BI-DW in just 16 weeks Being pre-architected for retail, ARC dramatically cut down the time to deploy BI in Haggen to about 16 weeks. The data sources to the warehouse include JD Edwards, ISS POS and homegrown planning systems. The data warehouse holds 5 years of historical data at week granularity with rolling 13 months of data at day granularity. The solution has been deployed on HP servers operating in a Windows environment. There are currently over 100 business users using ARC.

The food and grocery variant of the ARC retail data model, specifically designed to recognize the nuances of food and grocery retail business helped Haggen's business managers be productive from day one. ARC enabled a much needed item-level reporting on merchandise performance. Low margins in food and grocery segment required merchandisers and category managers to constantly monitor and analyze sales and margin performance of food categories at multiple levels of granularity across multiple dimensions. Complex performance indicators like ‘all-in-gross' helps Haggen uncover profit making opportunities, previously considered impossible through analysis.

ARC gives the senior executives of perishables team, sales team and category managers visibility into movement, sales and gross margin across merchandise hierarchy (SubDepartment, Major Category, Sub Category, Minor Category or UPC), which they did not have. Quick generation of reports like item and price comparison reports enables the store department managers to take better replenishment decisions for their store's unique demographic needs. Day-part reporting allows the ability to monitor stock/sales throughout the day. The store managers and merchandisers can now easily compare private label with national brands. Financial analysts and controllers are able to analyze costs by separate buying and selling merchandise hierarchies. The time frame to generate these critical reports has reduced dramatically to less than 5 seconds across all reports. Incremental daily data loads take only 30 minutes to support timely day-part decision-making. The enterprise BI licensing model of ARC has enabled Haggen to scale up number of users from 15 to over 100 at no incremental cost. The current architecture is maintained by just 1 person for administrative and monitoring activities.