Big Data is everywhere in the Retail Industry. It would be be hard to find any part of the management of retail operations that is not deeply touched by Big Data. In fact, it is already clear that to survive in the Amazon era, all retailers will have to rely heavily on Big data to help them store the right merchandise, at the right times, in the right quantity, and at the right price. Those that ignore it will die.<\/p>\n
This does not mean that only the large companies that have the resources to exploit Big Data will thrive while small companies will die. Small companies will be able to harness the power of Big Data \u2013 but they are likely to do so through niche consulting firms that have developed the professional and technical skills, hired a stable of experts, built the computing platforms, and acquired access to the massive data required to operate effectively in this field. Smaller retail outfits that buy their services will find them expensive \u2013 but the benefits should far outweigh those costs.<\/p>\n
The biggest costs will likely not appear on the company\u2019s financial ledger. The biggest costs will be the time, focus, and energy that senior and middle management will need to invest to come to grips with how to leverage Big Data and how to build investment arguments that make sense. This last point proved to be a major stumbling block when general data processing began to make inroads into large and then medium sized companies some 40 years ago. It is quite liable to prove to be a stumbling block in the application of Big Data as well.<\/p>\n
With regards to department stores, the most promising use of Big Data is through recommendation engines.<\/p>\n
Recommendation engines<\/strong> — Recommendation engines use the historical purchasing decisions of customers to predict future purchases and recommend other products to customers that they may be interested in. Big Data using these engines have the potential to generate accurate product recommendations to customers before they even leave the webpage. Amazon<\/em>, for example, sees a 30%-60% revenue uplift due to these recommendations alone. These recommendation engines are a widely-used way of incorporating Big Data into department stores because they are easy to implement and have an immediate positive impact on revenue: recommendation engines are shown to have the potential to boost revenue by 24% on average.<\/p>\n Trend Forecasting<\/strong> — The second biggest use of Big Data in department stores is predicting trends and forecasting demand. Trend forecasting algorithms comb social media and web browsing habits to find what products and services are causing buzz. These algorithms also analyze ads to see what products marketing departments are pushing. The algorithms then compare the data gathered from social media with the data gathered from current ads to accurately predict what the top selling products for a given quarter will be, how to better market products, and how to develop more cost-effective marketing strategies.<\/p>\n These predictive algorithms assist retailers in making better informed decisions about stocking and product ordering. This capability is particularly helpful during the holiday season when shopping rates increase \u2013 machine learning can use past historical shopping data to forecast future purchasing and revenue outcomes. It is anticipated that this kind of predictive analysis in department stores will grow from a $2.7 billion global market in 2015 to a $9.2 billion by 2020, a CAGR of around 27%. In the US alone, predictive analysis from big data is expected to reach a $3.6 billion market by 2020. As of 2015, less than 25% of department stores had adopted predictive analytics. Between 2018 and 2020 this is anticipated to grow to 70%.<\/p>\n After identifying trends, Big Data (particularly in regards to customer economic and geographic information) can be used to understand where and when this demand will come from. This helps business to generate effective marketing and advertising campaigns. For example, Ozon.ru (Russia\u2019s first online retailer) analyzed that demand for books rises when it gets colder during the winter months, and thus increases the number of book ads their customers see. This ability to accurately forecast demand, that comes with using Big Data, is helpful in lowering a business\u2019s costs, as it is expensive to keep excess inventory on shelves and having too little stock drives down revenue and decreases customer engagement and loyalty.<\/p>\n Price Optimization<\/strong> — The third main use for big data in department store retail is optimizing pricing. In retail, Big Data can be used to help assist in determining when prices should be dropped (marked down optimization) or when they can be raised without customer dissatisfaction (reflected by a lack or reluctance to purchase). Previously, before the advent of Big Data, markdowns occurred at the end of the buying season, with stores hugely discounting their remaining merchandise. The problem with this approach is that demand is already gone by the time that markdowns occur. Big Data analytics demonstrate that what is actually most effective in increasing revenues is to gradually lower prices once demand initially begins to decrease. When the US retailer Stage Stores<\/em> employed this technique, it could increase its traditional end of season sales revenue over 90% of the time.<\/p>\n Weather Optimization<\/strong> — Big Data is particularly helpful in optimizing prices in accordance with weather conditions. The Weather Company<\/em> (part of IBM) has found that \u201cweather is one of the largest swing factors for economic and business performance\u201d – 60% of shoppers change their behaviors when it is either raining or it is hotter than average out. A 1o<\/sup> F drop in temperatures below 60o<\/sup> produces a 2%-3% drop in apparel sales. Approximately 60%-70% of a retailer\u2019s excess expenses are due to weather-impacted supply chain costs (e.g., trucks held up due to poor weather conditions). In the UK, if temperatures reach over 65o<\/sup> F there is a 22% rise in fizzy drink sales, a 20% rise in juice sales, and a 90% rise in lawn furniture sales. In the US, temperatures below 64o<\/sup> F increase sales in soup, porridge, and lip care. Food, drink, pharmaceutical, and apparel sales are the categories most impacted by weather.<\/p>\n Targeting Individual Consumers<\/strong> — The final use of Big Data in general retail is identifying individual customers and how to most effectively market to and target them specifically \u2013 whether through email, text, or location-based alerts. Retailers, for example, can install sensors in their stores to identify customers\u2019 locations through their smartphones. If a customer\u2019s smartphone\u2019s WiFi is turned on, it will attempt to connect with the store\u2019s internet and this is how a customer\u2019s location can be sensed and tracked. Retailers can then track what specific stores she visited, what departments she visited, and what products she purchased at what time and on what date. This information can be used to better understand each customer\u2019s movements and patterns when it comes to shopping. Retailers can then use this information to reorganize their stores to optimize customers\u2019 shopping experience and even to offer special deals and coupons to bring further business to their stores.<\/p>\n In addition to the Big Data applications listed above, there are four other applications that apply specifically to online retailers.<\/p>\n Dynamic Pricing<\/strong> — Dynamic pricing is Big Data at its finest. Dynamic pricing is highly responsive to external factors such as consumer demand and competitors\u2019 prices. Dynamic pricing collects trend data about which products are being bought to automatically adjust prices. Its analytic capabilities slowly increase prices on items that are popular and discounts prices on items that are less popular. Dynamic pricing is key to increasing online retailers\u2019 overall revenue.<\/p>\n Individual Customer Experience<\/strong> — Big Data analysis gives sellers insights about customer behavior and demographics and provides customers a personalized experience. For instance, customer data can be used to create\u00a0buyer-specific e-mails for promotional campaigns. For example, \u00a0Amazon<\/em>\u2019s<\/em> \u201cCustomers who bought this item also bought\u2026\u201d recommendation feature increased sales\u00a0nearly 30% when it was first implemented. This is a simple and remarkably effective way to keep customers on a retail site and keep them buying. Consumers might have reservations about their\u00a0favorite retailers\u00a0knowing intimate details about their lives, but they\u2019re going to love the results in practice. Sharing all those personal tidbits is helping companies like CNA<\/em> identify\u00a0fraud and prevent customers from having their identities compromised. Retailers can use information from live transactions and other sources (such as social feeds and geo data from apps) to\u00a0prevent credit card fraud\u00a0in real time.<\/p>\n Better Quality of Products<\/strong> — Amazon<\/em> is the\u00a0e-commerce standard when it comes to smart, effective pricing. It can easily access its\u00a0competitors\u2019 pricing data\u00a0and respond quickly with its own deals \u2014 changing some items\u2019 prices up to 10 times a day. The industry-wide shift to dynamic pricing means that companies will no longer be competing on price alone. They will now need to establish a reputation for offering their customers the best\u00a0value<\/em>\u00a0and the best\u00a0experience<\/em>.<\/p>\n Reduce incidents of shopping cart abandonment<\/strong> — Companies can also use cross-device tracking to reduce shopping cart abandonment rates. EBay research found that the average consumer uses as many as\u00a0three or five devices or platforms\u00a0during the course of her buying journey. Mapping this journey with data allows retailers to help their customers\u2019 transition from one device to the next and complete their purchases.<\/p>\nGeneral Online Retailers<\/h2>\n
\n