We’ve talked a bit about the Internet of Things (IoT) in terms of what technologies and trends are driving the market: Huge volumes of information are being collected, whether in structured or unstructured form, whether from customer input or geographical positioning information, transactions or telemetry; how mobility not only provides transactional information, but the ability to provide locational information that allows you to provide the right offer to the right customer at the right time in the right place; how analytics turn that sea of data into actionable information. But we’ve only scratched the surface in terms of how that can affect the customer experience.
Delivering an IoT customer experience is a formula, but the customer is often overlooked as an element in that formula. The customer is often—though not always—the starting point in the IoT experience, and should almost always be the finishing point. (In some applications, like predictive maintenance, a piece of equipment may be an end-point in the equation, but we can trace that to a better customer experience, as well; HVAC units that don’t fail provide a more clement environment for the customer, and while that doesn’t necessarily directly convert to dollars spent, it does keep the customer in your environment longer, improving the likelihood of more spend.)
But let’s focus on the human version of “customer,” the person you’re trying to serve, the person who pays your bills, the person whose satisfaction contributes to your bottom line. As we’ve discussed, the IoT customer experience is a formula, and the customer provides much of the data that’s fed into that formula.
Transaction history: This is the bedrock, and always has been, of customer relationship management. A lot of transaction information is, fortunately, recorded in a structured format: A database gets information in specific fields that detail when, where, what, and how much. This breeds easy analytics. But touchpoints have become more scattered than the point-of-sale or order-entry terminal, transactions aren’t simply purchases anymore, and it’s not just debit and credit cards providing the information.
Customers can contribute search history by using a branded smartphone app. They can provide valuable traffic and dwell time information through smart applications of video surveillance technology (or Internet searches in non-retail or wholesale environments). Aggregated information can be collected through sensors in parking lots and readers at airline check-in counters.
Social data: Sentiment analysis can be used in an aggregate fashion, but the nature of social media also makes the customer individually identifiable. Often brand-specific posts are made from mobile devices, with geographically-specific identification as well. Users post 20 million tweets an hour, and 874 million users access Facebook from a mobile device every month, according to 2013 statistics. How do you harness this unstructured data?
We’ll get to the how shortly. There’s another source of data to be brought into the fold that’s not customer-generated. It’s the flipside of the IoT data generation coin. While customers themselves generate most of the data you need to customize their experience on the fly, there’s a whole environment of data collection devices that don’t rely on your customer to be collected. This is often where the discussion of IoT is focused, though I’d argue ignoring customer-generated data is a colossal mistake. Let’s call it ambient data, for lack of a better term.
Your car generates terabytes of data, now that it runs on computer chips as well as carburetors. Much of that data—if it’s relevant—can be connected wirelessly by your dealership to autonomously schedule tune-ups and oil changes. It can be used by your insurance company to determine that you’re a low-risk driver, and thus calculate a lower rate. There are hundreds of sensors in your car, all with the potential of connection to the Internet.
But that’s just a start. Shop-floor machinery can be connected to the Internet to detect potential maintenance issues. Parking lots can tell a mall’s web site where it’s crowded and where it isn’t. If there is an object, there is a way to connect it to the Internet. In 2012, Cisco Systems Inc.’s Rob Soderbery estimated that there were 8.7 billion devices—things, if you will—connected to the Internet. Ninety-nine per cent of things that could potentially be connected weren’t at that time. By 2020, that number should be about 40 billion.
I like to call this “ambient data” because much of it is noise in terms of the customer experience, but much of it creates an environment or context for data use. Does the volume of traffic on the QEW east of Ford Drive have any impact on the experience of a customer at a Shoppers Drug Mart in Etobicoke? We can collect the data from open sources, but does it mean anything in the context of the customer?
In fact, much of this data is collected in a “fog” computing infrastructure—it’s like cloud computing, but closer to the ground. The data is collected and exchanged at the device level—for example, to tell you through a roadside display how long it will take to Bay Street by taking the QEW versus Lakeshore Drive. It doesn’t go north-south to a data centre for analysis; it stays close to the ground, moving east-west, to provide up-to-date information, unless it’s necessary for more in-depth analysis.
And this is why the key to leveraging the Internet of Things is in analytics. Moreso, it lies in making analytics tools accessible to business users. If I were a programmer, I could spend several weeks creating a query that compares weather patterns in Outer Mongolia with delivery times to a customer in Rankin Inlet; it would be a perfectly valid analysis, it would simply be completely irrelevant. Analytics tools that bring together the capability to detect trends and demonstrate opportunities that can be harnessed by those with business acumen, not just programming talent, are key to wresting value out of the Internet of Things.
