All Articles Marketing Marketing Strategy Redefining real time when it comes to data

Redefining real time when it comes to data

For data to provide real value, it needs to be on-demand and fresh, and enable accurate decision-making.

5 min read

Marketing Strategy



In this digital era, just how real is real time? We know the drill: No matter what we do or what we need to do, it must be in real time. But for marketers in many disciplines, always seeking to enhance the connection between company, product, message, medium and target, shouldn’t operability be as important? Let’s step back and, yes, take a little time to reassess priorities.

First, a lot of so-called real-time data truthfully isn’t.  Second, that’s not the real problem; there are larger issues than, for example, the delay of a few milliseconds in the transfer of information. Third, a more comprehensive approach to the definition of real time — one that encompasses other critical variables — can indeed make a major difference to everything from operating models to target markets and the bottom line.

To get started, we should acknowledge the reality of many technologies currently in place. It’s routine for data platforms to exist in complex architectures that support both structured and unstructured data streams, feature high-velocity transactions and so on. Moving forward, the number of disparate data sources will only increase, and the data will become even more complex. Meanwhile, when adtech and martech companies need to synchronize with each other’s databases, they often use manual APIs or batch processes. These are costly, clunky and not really in real time.

But let’s be honest: The universe of applications that can only function with data that’s absolutely “real” down to the millisecond is quite small. The rest of us can be more flexible — we don’t want it to be a week or a month old, but a few seconds won’t make a massive difference. Of course, more agile technologies that can scale faster and process data more effectively would be very welcome, and they are emerging. 

However, it’s time we set our sights higher than that, spanning the entire spectrum from aggregation to usage. It’s just as important to capture and serve up data in real time as it is to provide access. Specifically, we need to approach the concept of real time not in isolation but through the prism of other crucial variables: specifically, audience, context and message.

“Real-time data” truly adds value only if it arrives to meet specific needs — as in, on-demand. Data providers of every stripe increasingly have the need to slice the market more narrowly than ever, dealing with a smaller universe than before. With every data lake overflowing, what’s the best way to draw just a sip and extract only the subset they need? This ability also enables the pursuit of new markets and other business opportunities.

Next, we should redefine freshness, particularly with a focus on retargeting. Is there common ground between real time and obsolete? Consider the downside of barraging customers with ads for products they recently acquired, or when they’re out of market for the business entity making the pitch. It happens often, and it’s never good. Real time in this context is vital, and it can’t happen with updates being pushed out on a preset schedule established with content partners and data platforms — it only works continuously and on demand.

Finally, and perhaps most importantly, we have the issue of real-time decisioning, which must build on accurate data, robust performance and even real-time scoring: What’s the response to the creative involved, and how does it match up to predictions? We’re all familiar with audience dimensions and dynamic creative optimization (which typically relies on multivariate testing), but it’s now possible to go further. For example, does a particular segment perform better on a mobile news site than a YouTube video accessed via a desktop PC, even with similar content? The algorithms must be not only multivariate but also multi-dimensional — recognizing interdependencies between predictive features and the axis of audience, context and message.

In sum, successful customer interactions come about only through customized and contextually relevant experiences, and that mandates real-time information about users and their context. That can be a challenge: A data-centric marketing framework that can respond to millions of requests in almost real time requires major investments, such as for geographic coverage, redundancy and support. On the plus side, there are technologies available that offer real-time access to entity (think user, household or device) and entity relationship (identity graph) data, even as they manage complex identity relationships to build a unified 360° view of consumers.

Sure, the idea of “real time” makes for a great sound bite and satisfies the need for instant gratification. But, true data analytics to drive business decisioning requires time to be measured in a broader context, one that includes message and audience. In the past, without the right technologies, this was an elusive goal. Now, those capabilities are very real — and it’s time they were adopted. 

Marc Sabatini is the CEO of aqfer. He has over 20 years of experience consulting on technology and data solutions in direct marketing, CRM and omnichannel digital marketing.

If you enjoyed this article, sign up for SmartBrief’s free e-mail from the Mobile Marketing Association, among SmartBrief’s more than 200 industry-focused newsletters.