Finding, understanding, engaging and eventually knowing your customers used to be so simple. A business had mailing and phone lists, tracked appointments, visits, invoices and purchases and taught employees to remember loyal visitors. It was all about delivering the personal touch. How quaint.
A Brave New World
Today's world of mobile devices, social networks, messaging apps, online retailing, recommendation engines and geo-aware shopping guides means that customers can come from anywhere, start and resume interactions on different venues, leave public feedback on others and expect the entire experience to be seamless. Retailers have labeled it the omnichannel: communications and transactions across multiple touch points spanning the physical and online worlds.
What started as a retail phenomenon -- providing the ability to shop online and pick up in store or, conversely, to order unusual or odd-sized items online that are not stocked by the brick and mortar location -- has rapidly spread to every other form of customer interaction. Airline travelers expect to be notified of flight delays and gate changes via smartphone apps or text message. Befuddled buyers calling a technical support desk want to shift the conversation over to a chat window while sharing their desktop screen. Customers complaining about a faulty product on a company's Facebook page expect an email or phone call offering a replacement.
Consider the Customer
Everyone wants ads, information and promotions specifically tailored to their wants and needs. All of these require stitching transactional, application and messaging data from multiple sources into a unified customer experience. It's what some in the industry now call the "customer journey," which encompasses the entire lifecycle of a business relationship: browsing, evaluating, buying, using, supporting, repairing and upgrading.
It's a powerful vision, but one that's deceptively complex as it involves gathering, organizing, correlating and analyzing data from many sources -- the quintessential big data integration and analysis problem. Generally in the form of log files or transaction records, the right data can provide insight into and context for what customers actually did or intended versus what they or the company thinks might have happened. But making sense of past data is only part of the omnichannel challenge.
Truly delivering on the seamless customer experience involves the contextual part: anticipating what the customer is likely to do or need and predicting it in real time. Proposing to switch a user's cell phone plan to one providing global roaming does them no good if the offer comes in the next bill, after they've returned from a two week trip and racked up hundreds of dollars in roaming charges. Delighting the customer means contacting them the first time you notice their phone accessing a foreign network, thus preventing problems, angry customers and potentially lost business.
Putting Data to Work
It all starts with data, something which most businesses have plenty of. According to 451 Group storage analyst Marco Coulter, data shows that enterprise storage capacity is growing faster than Moore's law with many firms already managing over 10 petabytes. Much of the growth is due to machine-generated data, just the kind needed to correlate customer interactions across channels and derive meaning and insight from a jumble of transaction records and log files. And make no mistake: log files can be a jumble.
Generated up and down the technology stack -- custom applications, middleware systems, databases, operating systems, hypervisors and network hardware -- the data is in a variety of formats, using different user and session identifiers and capturing various types of events. Often, the only common denominator is a time stamp. But with this thread, a company can stitch together an information trail that helps them understand their customers' interactions across their every network, system and application.
Log and machine data management software has already proven its value to IT administrators and security pros, enabling them to sift through gigabytes of previously inscrutable data from dozens of systems to troubleshoot problems, predict and prevent performance bottlenecks and identify security anomalies. The true benefits occur when pattern recognition and self-learning is combined to provide proactive insights into everything from datacenter operations to business trends.
Theory in Practice
The same software used to correlate a sophisticated cyber attack that traverses multiple networks and systems can just as easily track, analyze and (most importantly) correlate web clicks, social media comments, email conversations, help desk requests and mobile app usage. This sort of big data analysis can provide marketing, sales and customer support teams insight by product, geography, customer demographic, interaction channel and client device. When paired with external data sources, log analysis can even correlate customer activities with the competitive pricing environment, public or online opinion, even the weather.
Examples and applications with real business benefits abound:
- Discarded shopping carts are often the result of slow website response. A machine data platform can correlate abandonment rates with network latency and application performance and allow administrators to take preventive measures when certain performance thresholds are crossed.
- A manufacturer or online service might monitor online chatter on popular social networks to detect a rising level of negative comments and product complaints. These can be the proverbial canaries in a coal mine that when paired with internal log information might uncover product defects, process flaws or infrastructure bottlenecks before problems get completely out of control.
- Retailers can use analytics to detect and address changing product demand in real time. A wireless carrier might find that accessory sales and support calls are directly correlated to new phone activations. Predictive analytics would allow them to increase inventory and help desk staff in anticipation of added demand when activations spike.
Although the trend started in retail, business execs and marketing pros must internalize the new reality that omnichannel customer interactions are affecting every business. One key to turning the situation from a source of customer frustration into one of delight is an analytics platform that can digest, correlate and summarize machine data from across channels, and then analyze and predict customer behavior and intentions in real time. It's a tall order, but there are plenty of products that can help turn this dream into a reality. Now it's time for IT, marketing and business execs to do some homework.
Editor's Note: Want more on the challenges of real time marketing? Read: Marketers Say Real Time, Customers Say Deliver