WibiData's Amit Nithianandan suggests retailers need to do a better job understanding the value of customer data.
Kerosene production was big business in the 19th century. The production process that distilled petroleum into kerosene also created a byproduct called gasoline that was subsequently discarded. It wasn't until the automobile was invented at the end of the century that gasoline began to be deemed useful.
The way in which kerosene manufacturers disposed of gasoline is similar to how many retailers fail to capture or utilize data about their customer interactions today. Despite e-commerce's continued growth, many large retail brands capture a small subset of the data generated by customers interacting with their digital properties. In fact, it's possible some retailers are wasting data that could fuel their company's growth as better data processing tools and technologies emerge.
Amazon.com Inc. stands out as one of the few retailers that leverages much of the data about its consumer interactions to deliver relevant, personalized experiences to its customers. As consumers shop online, Amazon collects a variety of data points, including browsing behavior data and purchase history data, to create robust, up-to-date customer profiles. This allows Amazon to offer consumers the products they want, seemingly before consumers know they want them.
While many retailers aspire to replicate what Amazon has accomplished, the engineering and data science resources historically required to build such a system from scratch were vast and resource-intensive. However, new open-source data technologies, such as Apache Hadoop, are making it economically viable to store, process and analyze massive amounts of customer data.
To realize the maximum benefit from customer data, it's important to start capturing every interaction. Product purchases, or online conversions, are a clear metric to gauge success and profitability, but it is also imperative to record information about how your customers arrived at the purchase decision:
Effectively capturing and processing this information allows retailers to create an optimal shopping experience wherein products and content can be precisely placed, tailored and marketed to maximize purchases. By looking at data and analyzing it accordingly, organizations can make real-time decisions that will satisfy each customer.
Netflix Inc.'s $100 million investment to produce two seasons of the TV drama "House of Cards" is a good example of successfully utilizing customer interaction data. Data played a pivotal role throughout the development and production of the series. For example, Netflix relied on its warehouse of data to identify that customers who liked the BBC series the show is based on also watched and liked movies starring Kevin Spacey and films directed by David Fincher. Netflix cast Spacey as the star of the series and hired Fincher to direct. Netflix further tailored the cast, storyline and directing styles to cater to a large segment of its subscribers based on their viewing preferences.
However, Netflix doesn't stop at producing appealing content that many people are likely to watch. It captures information such as users' playback devices, scrolling behaviors, search terms, and clicks on the play button and then feeds that data into algorithms it has developed to present even more relevant recommendations to each customer.
These interactions are important as they are the primary ways customers interact with Netflix. For example, if you like "Road House," you will probably also like "Red Dawn." Or, if you're a Tina Fey fan, you'll probably like to watch episodes of "30 Rock." Netflix uses data to create a personal concierge by presenting consumers with the movie recommendations that they are most likely to enjoy.
Some examples of data online retailers can capture to move toward similarly tailored recommendations include:
This kind of data can help optimize product placement based on context. For example, it might make sense for an apparel retailer to promote sweaters to consumers shopping via a mobile phone and dresses to desktop viewers. This data can also help the retailer better identify customer and product segments. Understanding what the customer sees but doesn't act on can be just as important as what the customer sees and does act on. With powerful data processes, organizations can rank products based on context and personal preferences instead of aggregate or predetermined tastes.
Today, it is possible for an e-commerce technology team to build a data application using free frameworks such as The Kiji Project, Apache Hadoop and Apache HBase. These frameworks allow engineers and data scientists to quickly and efficiently process and store customer interaction data to create a comprehensive view of their customers.
Designing and implementing such a data capture system is certainly a complicated project. However, viewed through the lens of retailers like Amazon.com that have implemented such systems, the potential revenue growth—in the form of customer acquisition and retention—far outweighs the costs.
Taken as a whole, these interactions represent a comprehensive blueprint of the customer journey. Not capturing data or capturing it in a way that prevents fast analysis means that you could be discarding the gasoline that may fuel your company's growth for the coming decades.
Amit Nithianandan is a "big data" engineer and member of the technical team at WibiData, a company that builds and sells data applications.