Rethinking enterprise data management in our new normal
Ever since Gartner’s Doug Laney coined the “3 Vs of Big Data” back in 2001, there have been endless iterations and attempts to define the terms of engagement. The core three were always “volume, velocity, and variety.” In subsequent years, there has been a consensus around at least two more: veracity and value.
In general, it’s easy to agree that we need the ability to manage ever-increasing amount of data (volume), collect and activate it in real-time at the speed of the modern consumer (velocity), manage increasingly diverse types of data (variety), understand the truth inside of data to make it relevant (veracity), and make it actionable in the different endpoints that impact business (value).
A lot has remained the same since 2001, and this structure continues to pay dividends as a framework for the broader discussion of data management. But how should modern CMOs, CIOs, and “Chief Data Officers” be thinking about this in 2022?
The excitement and growth in the CDP category have provoked some interesting questions about how to think through customer data management, so let’s apply the lens of current context to think through the “five.”
When it comes to leveraging data to drive customer experience (arguably the primary and most important use case for today’s CDPs), we are basically in the very first innings of a long and interesting game. There are so many possibilities for unleashing data on modern CX challenges. Loyalty data, commerce data, and marketing data are just the tip of the iceberg. Going a step further, imagine in-store purchase data, call center data, and social data combining to provision a truly rich “360 degree” view of customers. With such a profile, a smart marketer could do practically anything.
The ugly truth is that most marketers use relatively tiny amounts of CRM and e-mail data for personalization. While some more advanced marketers have managed to combine commerce, loyalty, and marketing interaction data at scale to drive better CX, many are stuck in the closed-loop of using marketing data to drive better marketing use cases and activating that data in channels such as display media and e-mail.
Another truth is that ERP data is hardly being used at all. That’s a huge opportunity. Enterprise resource planning systems are the foundational systems that store the real data tied to business outcomes that directly impact customers. Is the product in stock? How much does it really cost to manufacture and ship it? How soon can I get it there? ERP systems – and “supply chain management” in general – seem to offer the missing variety to power impactful experiences. This argues against CDP being an evolution of CRM, especially if we look at CRM tools as endpoints that only manage engagement between the enterprise and its customers.
I think there is an interesting argument to be made that the foundational data layer of customer experience starts a layer below with ERP — and connecting ERP and CRM through data is where the magic of CX might happen.
How do we connect people to the enterprise, not just the marketing department?
Managing data at scale has always been a challenge, but I don’t think even the most forward thinkers in 2001 could envision the truly massive growth in useful customer data over the past several years as mobile usage, and the “Internet of things” continues to create and make available massive amounts of highly granular data that can be pointed at customer experience use cases.
Marketers have indicated that they are using X data types today and estimate that they will use X in 2022. Those data are increasingly coming from contextual and behavioral signals – many of which need to be consumed and activated in real-time to be useful. Another element of volume hard to envision even a few years ago was the maturity in machine learning as a method of extracting value from these massive data sets.
Why did Google make TensorFlow – arguably one of the world’s best library of ML models – free to any developer in Silicon Valley? Basically. Because it’s not the algorithms that create value (they are largely undifferentiated), but rather the volume, scale, and fidelity of the data they need to run against. In short, more data means better results. And not just more – but more valuable. Looking at these three elements that drive success in AI and machine learning, it seems like “volume” is only as important as the quality and fidelity of the data. Yes, they need to be unified at the customer profile level, but the attributes also need to be connected to the business, beyond marketing and advertising inputs.
For example, I may be able to attach petabytes of data related to online advertising interactions at my CDP – but is the processing and storage of such ephemeral data worth the expense? Are there more valuable attributes that can contribute to enterprise data volume that is tied to the enterprise? We need to consider that better data is greater than more data. Better data is data that is connected to the business, and that can create better outcomes in endpoints like Sales, Service, Commerce, and Marketing.
How do we create engagement at enterprise scale?
When we think about the velocity of data, it’s tempting to focus on real-time engagement use cases that require a CDP profile (personalizing real-time journeys across text and e-mail, or website/app personalization based on behavior, as examples), and that is not wrong. There is a ton of low-hanging ROI to be gained through personalizing faster. Every CMO wants to connect their customers the business in real-time and, given the incredible amount of channels customers can engage a company on, it is starting to become a necessity.
The missing element when it comes to data velocity, however, is related to overall business agility – not just the ability to match an e-mail offer with web content. What do I mean by that? Think of today’s approach by most brands: they have a roster of new products they are going to put into the market, and they configure their outbound marketing plans to reach customers on the channels they are engaged with. New sneakers coming out are put into broad-scale national campaigns, with each channel slightly optimized for personalization. Makes complete sense. What brands market is directly related to what they try to sell in campaigns.
The hidden layer when it comes to delivering meaningful velocity in data management has more to do with data that can unleash true business agility. Instead of answering the “what do I market” question based on products alone, businesses need to add some additional context. What’s in my actual inventory, and how many of each SKU can I actually sell (supply chain). Also, what is my potential profit if I am successful (finance)? To whom can I sell, based on their permission (identity)? And also, where can customers buy it, and how quickly can orders be fulfilled (commerce)?
How we can give the people not only what they want, but what they actually ordered?
If the rise in CDPs has done anything, it has put a laser-like focus on the importance of unified data. Whether you call it a “customer 360,” a “single source of truth,” or the fabled “golden record,” enterprises have a mandate to connect as much data to a universal ID or profile as possible. Obviously, doing so eliminates the problem of stitching people together with their many different postal, e-mail, and pseudonymous IDs – and creates value in terms of unifying data for the purposes of cleaner analytics. At its core, getting to the “veracity” or truth behind the data is an identity challenge.
That said, if you think about how most non-technical folks think of “identity” when it comes to customer experience, the term has taken on a decidedly “marketing” flavor. Back when 3rd party cookies were in vogue, “identity” meant cross-device identity (stitching cookies and device IDs to the universal identifier), or maybe was thought of as data “onboarding” (matching hashed e-mail addresses to an active cookie ID). Or, when we thought of “identity enrichment” we were mostly thinking of big marketplaces where 3rd party data attributes like “car intender” could be appended to first-party data. Obviously, these tactics still have a place at the table in the identity conversation, but we need to broaden our thinking.
Now that we live in a world in which 3rd party cookies cannot easily be shared, and first party-data, when collected, must be done with consumer-driven permission, it is clear that the top-down approach to identity (I buy identity services by companies that capture data) has to change to a bottoms-up approach (I use tools that give consumers the ability to grant me permission to use my data). In a word, we have entered the CIAM era of identity management, and the new token of value is “zero-party” (permission-based) data.
In this new world, zero is everything.
Finally, as we think about the value delivered through modern data management, trends are becoming readily apparent. CDPs that are good at personalizing online experience – mostly marketing, advertising, or loyalty experiences – are shrinking, and “enterprise” scale CDPs that go beyond marketing to offer personalization in channels like sales, service, and commerce are on the rise. This makes sense, especially considering how much more meaningful “real-life” touchpoints are to consumers. In short, I’d rather you got my name right and knew my order number when I call your 800 number than get just the right ad in Instagram.
This trend is not going away, the big CRM companies will continue to expand their data offerings through tangible connections to CRM-based customer experience delivery. But, for every valuable use case that can be delivered by connecting marketing and call center data, there are dozens of opportunities to go a layer deeper and connect real enterprise data to CX.
As above, who is doing personalized ecommerce using “available to promise” data about products hidden in the supply chain? Or leveraging real-time inventory data from the manufacturing line to influence campaign execution? The answer is almost nobody, which is also the opportunity.
Is “enterprise data” hidden in the chain become the true driver of “enterprise CDP?”