skip to Main Content

Working With Big Data and eCommerce

The term “big data”, or using complex datasets to further understand operations and customers, has now been a buzzword in the eCommerce industry for several years. But is this the year big data stops being a buzzword and becomes a reality? Here are some tips to help understand the expanding relationship between big data and eCommerce…


Customer acquisition is only one element of big data – Key data metrics can also be used to up-sell existing customers and retain long-term customers based on behavior across multiple touch points. Business models and strategies can now leverage new data to create informed and effective promotions to attract not only new customers but also to target specific segments of existing customers. For example, smaller promotions can be sent to existing customers who are likely to purchase anyways and larger promotions can be sent to customers who need more of an incentive to make a purchase.


Big data can be used to improve existing products – Big data is being used more often to predict and understand consumer behavior, both online and offline. For example, some fitness brands have started linking products with smartphone apps to manage fitness endeavors and use the data collected by the app (running distance, interaction times, etc.) for future product design and marketing.


Big data can grow eCommerce site customization exponentially – A small number of online retailers are now able to customize content at a device-specific level in real-time. This can be accomplished using consumer interaction data with mobile devices to push top mobile purchases to the index page. More customization options include one focused on specific past purchases – if a customer likes a sweater in a certain style (i.e. scoop neck), those will be given priority when the customer logs in to their account on the site. According to Monetate, some eCommerce sites have up to eighty on-site experiences (geography, gender, etc.) which can be tailored for each unique visitor.


Passive “micro-categories” are now available through big data – Innovated by the “micro-genre” of the online video service Netflix, passive micro-categories can now be assigned by some retailers to products with great depth. Want a pair of black shoes without laces that has gold trim? That’s now a micro-category, and the possibilities for some retailers are essentially endless.


Computer-based recognition networks are becoming a reality – It may sound like science fiction, but deep learning and artificial neural recognition networks do exist. These networks are highly complex and can be trained to recognize images, language, and specific speech patterns, allowing for easier product discovery and interaction. Speech pattern recognition in particular can be used to automatically summarize lengthy product reviews into accurate tag-friendly product keywords. Keyword tagging for images also exists as well – if a product picture of a yellow hat is uploaded, it is automatically tagged with the keywords “yellow” and “hat”.


Predictive marketing can help boost future conversion rates – With devices such as Google Glass, the predictive abilities of big data can become very powerful in the future. An individual could try on several green colored jackets while wearing Google Glass and if they choose to share their data with their favorite stores, could receive notifications when new green jackets are in stock. Anticipatory marketing could also become a possibility in the future. If an online retailer knows a customer has a trend in buying winter gear in October, they could send the customer specific offerings based on past purchases – if the customer just bought a winter jacket in their size on clearance in March, winter jackets can likely be excluded from offers in October.


Use big data from other professions in the retail industry – Some retailers are using big data in atypical ways. Retailers can use factors out of their control, such as predicting product volumes for future seasons by using weather-related big data – if a long winter full of snow is predicted by weather experts for next year, it may be time to stock fulfillment centers with items that sell well in winter conditions.


In short, the main goal of big data is using sophisticated information to drive engagement and sales. However, given all of the data available to online retailers and the potential costs associated with obtaining and managing this data, research will likely be needed by each online retailer to conclude which kind of data is most beneficial to their efforts. In the eCommerce industry, it’s usually more important to gather data which is actionable and can help a retailer’s bottom line than to simply gather as much data as possible.