In today’s digital world, no other industry has the potential to witness an exponential rate of growth of data than retail. More and more retailers are warming up to draw blueprints on ways to identify, capture and harness mountains of raw data into manageable business cases and actionable insights with tangible business benefits
Big questions, big data, big answers, big returns – this would have been a short and sweet story for big data in retail and that too with a happy ending. For a generation of retailers who have grown up on MIS reports evolving into dashboards and trend analytics recently, it will be a steady and gradual journey towards comprehending and realizing the big-data business case. The current reports and dashboards would have been created out of transactional data, with the data sourced from within the boundaries of the retail organization. These sources would range from point-of-sale and customer database to supply chain transactions
As retail channels evolve, the customer interactions transcend into areas outside the boundaries of a retailer’s control. Shoppers are influenced even before interacting with the retail channels, and purchase opinions are shared way beyond the point-of-purchase. In such a case, it becomes imperative for retailers to capture customer feedback from data sources outside their realm of physical influence.
So, the first question that many retailers ask would be, how big is big data? The answer is simple: any data volume and data type that is beyond the current capabilities of a retailer to manage, process and analyze can be classified as big data. Each item on a retail invoice, customer service call, email, social media activity and each tweet present an opportunity for data to be generated in retail.
Consider these opportunities in the light of an increasing consumer population with increasingly complex demographics and a growing affinity towards digital channels and devices. The proliferation in product lines, stores, social-media platforms and digital devices represent a potential permutation to generate data of the highest magnitude. In reality, this is just the beginning of things to come. We have only accounted for scenarios where humans generate data. How about a future where devices interact with other devices and generate volumes of data? This will create a data heap that will make finding a needle in a haystack look easy.
The case for big data can very often get overwhelming and intimidating for many retailers. Often, the business value of harnessing big data is wrapped along with additional layers of technology such as cloud, mobility and analytics. Since a lot of these aspects are in an evaluation stage in many markets, there are chances of a big data idea being put on the back burner. This often tangles the business value and the ROI model becomes complicated. Therefore, it is important to understand how retailers can form a pragmatic strategy and approach towards big data within the retail organization.
Defining and implementing a big-data strategy involves decisions across multiple technology layers to decide on the execution model. The journey for a retail organization should commence with first evaluating if the organization is a big-data candidate. The evaluation process consists of forming a business case and associated business scenarios. In most cases the scenarios are in the area of offering highly customized promotional offers, analysis of promotional effectiveness in real time or analysis of store space effectiveness by analysing the shopper path. Breaking up the business case into scenarios helps retailers identify current technology and skill limitations, investments and the associated returns.
Identification of various candidate data sources that will serve as inputs to the business scenario is essential to get a firm grip on the data characteristics. These characteristics are popularly classified into areas of variety, volume and velocity. For each candidate business scenario, the variety of data can be as structured as point-of-sale data and as unstructured as store-video feeds. The volume of data amassed by retailers can easily be a few terabytes a day depending on the variety. Velocity is the speed at which data is captured from its source and turned into meaningful insights. It is evident that technology decisions depend highly on the variety, volume and velocity of data.
Fortunately for today’s retailer, inexpensive hardware and innovations in processing capabilities coupled with the scalability of cloud computing can be ingredients to an attractive big-data value proposition. This way, the retailer can concentrate more on the operational and business-oriented aspects of big data. This involves integrating to various sources of big data and translating the data into a manageable and analysable form. These are used to define analysis models where substantial inputs are required from the business users to define the various actionable items that can be created out of the data. Now we are entering the realm of big data analytics which is ultimately the wow factor in big-data management.
Retailers can benefit from big data to such an extent that each customer can be classified as a separate segment. Shopper behavior can be tracked across all channels and the information can be correlated with the current assortment characteristics. Retailers can also tailor the store and shelf space through insights gathered from store-video feeds. These can provide information around how long do shoppers spend in a particular aisle or their interactions with products.
The implications of big-data insights range from customer analytics and merchandising decisions to store performance, assortment management and loss prevention. These are areas where it is necessary for retailers to rapidly take decisions based on a quick understanding of the customer profile, brand interactions, sentiments and basket size. The decisions can be in the form of accelerating product shipments from various locations, mark-down decisions or initiating vendor replenishments in advance
In an industry with razor-thin margins, defining and executing your big-data strategy has the potential to bolster the retail margins by a significant extent. Granular customer insights leading to highly personalized messages to the shoppers will lead to an increase in the basket size and as well as in the number of shopper visits to retailer stores.
Another area where retailers stand to gain from big data is customer service where service requests can be fulfilled in the shortest possible time and in the most precise manner. Big data implementation eventually pays for itself when the ROI is measured against reduced cost of operations, efficient assortments and reduced stock-outs.
So, hopefully within a few years, if a customer tweets about how a rain spoiled her day, she would immediately receive a discount offer for an umbrella. To her surprise, the color of the offered umbrella would match with the dress she bought two weeks earlier from the store!