Competitive Advantage in Retail Through Analytics: Developing Insights, Creating Value

Register now

The Current Retail Context

As differentiation between retail chains decrease and the need for convenience and value-added services increases, customers have become more discerning and demanding but less loyal than before. One retailer's customer today is a potential customer for all other chains, formats and channels tomorrow.

To create sustainable advantage over competition, retailers are trying to enhance their product offerings, service levels and pricing models. To prevent value erosion and to protect margins, retailers are trying to reduce their cost-to-serve per customer and thereby ensure that the total cost of ownership of a customer over time is reduced. Managing promotional spends is another critical area for retailers to focus and target customers more effectively and efficiently.

Paradigm Shift for Retail Focus

Over the last decade or so, retailers world over have focused their resources to create efficiencies in their sourcing, supply chains and store operations with the purpose of reducing the operating costs and maintaining a hold on their thin margins.

Figure 1: Customer Satisfaction

However, today these process improvement efforts have become par for the course. There is an urgent need for retailers to revisit the core aspect of their business, which has been out of the center of their attention: customers. The task ahead is to reorient the organization to shift the paradigm from customer relationship management (CRM) to customer relationship building.

Customer satisfaction is no more a logical conclusion of management of the selling process; it requires in-depth understanding of customers and products (see Figure 1). Development of strategies with thorough knowledge of customer needs, behaviors and motivations along with product trends and associations clubbed with execution is the only approach that will lead to service levels significantly superior to competition. Technology acts as an enabler to provide insights into customer behavior, product interrelationships, and successful promotion types for different channels.

The wealth of data available to retailers presents the opportunity for competitive advantage, but the gains only accrue to those who exploit it the best. Though CRM, data warehouse (DW) and business intelligence (BI) tools implemented by retailers are useful to a great extent, retailers are still seeking decision-enabling insights. The objective of data analysis has evolved from knowing a customer to predicting his behavior.

The objective of this article is to look at the applicability and benefits of analytics in retail, with special emphasis on predictive analytics. Some of the key challenges in implementing analytics program are also examined, and suggestions to overcome these trials have been outlined.

Introducing Analytics

Without utilizing the powers of analytics, the ROIon the large amounts of data collected will always be suboptimal. Analytics goes beyond the preformatted logic in BI tools by analyzing the problem in a given context and then using as large an element of domain knowledge as the statistical techniques.

Analytic techniques, such as design of experiments, cross tabulations, statistical analysis and data mining, help in uncovering patterns and trends within large databases. When used for creating forward-looking suggestions, they provide the edge to decision-making.

While descriptive analyses help to identify issues and examine causes, predictive analytics enhances the accuracy and effectiveness of decisions that directly impact your bottom line.

Descriptive Analytics: Developing Insights from Data

As the term suggests, descriptive or causal analytics enable the uncovering the reasons behind any event or trend and brings critical marketing issues to the forefront. The techniques used answer queries of what is happening, where and amongst whom. Some analyses applicable for retail are:

  • Out-of-stock analysis,
  • Category and brand dynamics,
  • Promotion-effectiveness analysis, and
  • Benchmarking analysis.

Predictive Analytics: Going Beyond "What and Why"

Data warehousing tools feed data into descriptive analytic techniques to answer the questions of "who/what/where/when/how." However, all these answers help in interpretation of past events. The critical question of how a customer will respond or how an event will unfold in the future remains unanswered by descriptive techniques or BI reports. Predictive analytics play a decisive role in a similar way as diagnostic checks do in a medical treatment scenario as opposed to symptom-based treatments. It allows the retail organization to enhance its decision-making powers by looking at the future with analytical rigor.

Predictive analytics enables:

  • Targeting customers more effectively for campaigns,
  • Improving response time to market changes,
  • Increasing employee productivity, and
  • Improving customer service at stores.

Power of Analytics in the Retail Domain

A customer-focused industry with several channels of engagement with the customer, the retail industry has great potential for application of analytics, which can help it to serve its customers better.

As data from the point-of-sales, terminals and credit cards/store cards are synchronized and mapped along with data flowing in from other sources, analysis of this integrated data has the power to impact the retailers' bottom line positively (see Figure 2).

Predictive analytics hold the key to taking advantage of these opportunities such that retailers can increase their ability to forecast their customers' behaviors and plan accordingly.

Figure 2: Power of Analytics

The Analytics Complexity Model

An organization trying to implement analytics should plan for the natural evolution of the analytics process. For an entrant into analytics space, the need is to establish credibility for data analysis using simple techniques rather than advanced models (see Figure 3). Similarly organizations with a mature DW/BI foundation and a culture of data analysis should progress to the advanced stages and create analytical models, which enable improved decision-making. Thus, in the beginning, the level of analytics is dependent on the current familiarity and sophistication of the data analysis process in the organization.

Figure 3: Complexity Model

The Role of Analytics in Retail Data Flow

 Figure 4: Data Flow in Analytics

Analytics has a pivotal role in the scheme of data flow within a retail organization. The data warehouse structure feeds into the analytics layer drawing from multiple sources (see Figure 4). The analytics process takes care of the data preparation and modeling activities and generates reports. These serve as planning and monitoring guides across levels of the organization, and enable effective decision-making. These reports also feed back into the data source, thereby becoming a predefined model ready for future reference and use. Analytics models are natural extensions of the DW/BI structure rather than a deviation.

Applications of Analytics in Retail

Given the multifaceted nature of operations of a retail business, analytical models actually cater to all functions by providing standalone findings as well as integrated analytics, which enable enterprise-wide planning and execution. In this section, we look at some of the critical applications of analytics in the retail space.

Customer-Centric Data Analysis. Using data from the customer's shopping basket and customer-specific data (credit cards/store card), analysis can be done to segment customers and predict their response to marketing campaigns and promotions. Purchase patterns are used to cross-sell and up sell products to identified customers. Analytic models examine a customer's frequency and recency of visits along with purchase behavior, and provide a customer's churn or attrition probabilities on which retailers can take corrective action to reinforce loyalty. Some of the key analyses are:

  • Profiling and targeted marketing,
  • Customer profitability analysis,
  • Basket analysis for opportunities for up selling or cross-selling, and
  • Churn prediction and satisfaction analysis.

Store-Level Data Analysis. At a store level, analytics help predict sales (or profits) and enable benchmarking to improve store performance management and employee productivity. The critical task of identification and selection of location for new stores can be made easier through analytical models that can predict expected performance across options and suggest an optimal solution, such as:

  • Product placement based on product associations at a store level,
  • Store revenue or profit forecasting, or
  • Store performance assessment.

Category Management Areas. Category management is an aspect affected by various extraneous factors. Analytical models can study historic data along with loads of seemingly unrelated variables and develop more accurate forecasting models. These models not only enable promotion planning and stocking at stores but also can be used for making adjustments for the current context.

Challenges in Implementing Analytic Programs

Though very attractive in its potential, analytics implementation holds some challenges. Some key challenges are:Historical Data Availability. Organizations with low quantities of historical data or with data that has been restructured over time without ensuring continuity will have issues with implementation of analytical models as these models rely upon patterns from the past and utilize them to predict future events. While there are several ways one can use substitute data and recode and reclassify data points, lack of past data is a huge constraint for several types of analysis.

Multiple Data Sources/Integration Issues. While the multiplicity of data sources provides a diverse data set for analytics and helps create integrated models, it also becomes an issue in the absence of proper data warehousing and data marts. Data silos created by the DW structure can restrict the applicability of analytics. Data identification and cleansing become difficult when several data sources with different formats exist.

Implementation of a proper data warehousing system, which allows for extraction, transformation and interaction of data across modules, is one of the approaches recommended for overcoming this challenge. An ad hoc approach specifically for analytics is to move the relevant data from different systems into a common database to reformat and cleanse the data.

Clarity of Analytics Objective. Most organizations are unable to utilize the power of analytics because of a lack of clear objectives for the analytics program. More often than not, the results from analytical models become good-to-know information rather than the actionable insights they were meant to be.

For an organization to be able to truly utilize the potential of analytics, clear objective setting and usage planning for the results is as critical as the model. A responsible team should invest commensurate time in this preparation. This approach would result in capturing a significant ROI.

Integrated Team Profile and High Cost of Resources. An analytics program is a collaborative effort between various resource profiles in an organization. Integrating a team of data mining and statistics experts, domain and subject matter experts, tool-specific operators and data management experts can pose a significant challenge.

Retailers who want to develop an analytics practice in-house will want to make the investment and see the benefits accrue over the long term. An alternate and equally viable solution is to utilize the services of analytics solutions providers. These service providers bring to the table their expertise and experience in this area and provide benefits in delivering actionable insights that can help drive marketing strategies.

IT or Business. Whether the analytics program should be aligned with business or with IT is another dilemma that creates problems in the effective implementation of analytics. On aligning with IT, centralized management and ease of access across business users become smoother. However, resource management, investments and delivery responsibility remain with IT and increase its burden. Conversely, when analytics is a part of the business side, advantages are in terms of coordination and speed of delivery, but the risks of interfunctional rivalry for analytics resources and misuse of analytics to advance one's own purpose are high.

The service-provider concept discussed earlier can resolve this issue if there is an appropriate program governance structure interface with both IT and business.

Large and varied data sources provide new opportunities for improving customer relationships through analytics. Retailers need to keep rediscovering and understanding the dynamic nature of customer behavior to create a sustainable competitive advantage. Analytics provide actionable and powerful decision insights, increasing decision yield by delivering the intelligence required for developing and maintaining profitable customer relationships.

For reprint and licensing requests for this article, click here.