The time has come to make peace with data. What was once a beautiful partnership has become contentious. Like the repair of any relationship, turning the page requires taking stock of how the problem started, but more important, looking ahead to a brighter future. Yes, marketers have expected too much of their data. Yes, there have been great challenges and even greater inconsistencies in the ways data has been collected, shared, and integrated. But there are answers, and they’re worth discovering.
Not surprisingly, some of the answers do involve integrating more data.
Don’t turn away yet. Bringing data sources into harmonious agreement with business strategy doesn’t necessarily mean building a giant data warehouse, or taking on a complex new set of analytical tools. Brand objectives, not technological milestones, are at the heart of the new peaceful understanding of data. Along the way, enterprises can tune in to the voice of the customer, which is far more valuable when it speaks the language of every touchpoint, not simply the sales transaction. Understanding customer behavior and preferences over time can, for example, help sales reduce discounting while maintaining or growing volume, because deeper customer insights can help brands decrease the magnitude of incentives and better discover which customers have inelastic demand.
In short, a healthier relationship with data does more than bring peace of mind—it produces results. “Our clients achieve a 5 to 7% improvement in pricing and a 10 to 15% increase in customer lifetime value after integrating data across multiple channels,” says Randy Watson, VP of global client services at Acxiom, citing an internal study the company conducted.
Strategy gone awry
So, what went wrong? In a word: everything. Digital channels and the explosion of data storage and processing capabilities over the past 15 years made it seem as though brands could collect data about anything, anywhere, at any time. And so they did. “Marketers have struggled because they got so focused on capturing all the data that they lost sight of what the data means in terms of the customers,” says Jay Henderson, strategy director for IBM Smarter Commerce.
Customer relationship management (CRM) was touted as a solution to the bloat and the growing number of data silos by bringing all relevant customer data under a single umbrella. However, in most cases that meant asking CRM systems to do far more than they were originally engineered for. “CRM systems are still designed for sales VPs,” says Mike Volpe, CMO at HubSpot. “In most CRM systems you can’t really integrate social media interaction data; you can’t integrate website visit data or mobile app usage data, at least not without custom code.”
When data piles up but there are no consistent and repeatable processes to integrate, analyze, and act on it, necessity becomes the mother of invention. It takes laudable initiative to roll up one’s sleeves, open a spreadsheet, and create one-off reports or custom data sets to solve a particular business problem. But without standard procedures to publish and refine those queries and reports, the time-consuming process will be repeated whenever someone has a complex data problem.
Brands that think they struggle with data integration might be surprised at just how resourceful their employees have become—and just how much insight is hidden away on the average corporate laptop. “More people have more data in spreadsheets on their [hard] drive than you would believe,” says Michael Kean, president at IT services firm Altico Advisors. “You’ve got to fight to get that information out to the rest of the enterprise.”
In fact, it’s people and processes that are ultimately to blame for most data shortcomings in the enterprise. Even the best unified customer records can’t produce positive outcomes if brands are focused on such narrow objectives that they fail to cooperate internally to win and retain customers. When a call center agent’s average handle time is allowed to dictate customer retention, the brand suffers. “I don’t think this is purely a technical problem, and I don’t think we can apply a purely technical solution,” says Lyn Robison, VP of research at Gartner. Without concerted efforts to place customer-minded objectives first, “people will generally decide to do the work in a way that satisfies their own interests, as opposed to the larger enterprise’s interests.”
Stripping to the bare essentials
Data integration is not a goal, but a means to an end. Start treating it that way, and the way forward will be much less painful. “You’re better off working backward from some outcomes you care about and finding the integrations you need in bite-size chunks. Don’t try to solve world peace and have a unified customer record,” says Peter Chase, executive VP and founder of CRM integration vendor Scribe Software. “From the marketing team’s perspective, most companies don’t have the political will or the money to make that happen.”
Integrated data objectives should include improving the quality and relevance of information presented to business users. For example, let partner intelligence data decisions be driven by objectives, such as growing wallet share with existing customers. Make it easier for sales to understand the relationship between existing customers, and subsidiary, sister, and parent organizations that might also make excellent prospects.
Give customer service the data they need not simply to resolve issues in front of them, but to anticipate the likely reason for a call, as well as identify any potential engagement or sales opportunities.
Additionally, never let the quantity of data dictate terms. The flood of social and mobile data is driving integration for integration’s sake. Fight that urge, and instead develop social and mobile strategies supported by data. The necessary data links, and stakeholder involvement, will be much clearer than any uniform data warehouse ever will be. “Some pharma companies are tapping social platforms to learn about off-label uses of their drugs. Finding a potential new use for a drug is valuable and opens up a realm of possibilities,” says Wilson Raj, SAS global customer intelligence director. That business objective—expanding into new markets—is readily supported by ensuring that social discussion of prescription use makes it to the relevant brand managers and product development teams, rather than letting it linger in a dedicated social media database.
Social data can be as overwhelming as it can be invaluable. Brands built on highly personalized experiences and a strong grasp of individual characteristics should resist the urge to buy the ever-expanding aggregated data streams from social and display networks unless they understand how that data can complement current campaigns. Allowing the aggregate data to overshadow valuable personal insights will hurt in the long run. And be aware that not all that glitters is gold and not all data is worth owning, let alone integrating. “There’s always going to be an 80/20 rule. Out of 100 billion data points, some small percentage is going to be the most actionable when it comes to targeted, personalized marketing,” says Steve Krause, senior VP of product management at Responsys. “Only a small subset of the facts will end up being the most important. The interesting exercise is finding out what those facts are.”
Despite its source, inconsistent data leads to inconsistent treatment of customers—across touchpoints and channels, across interactions, and even across simultaneous campaigns. Brand leaders, including the CMO, can limit the drift of both the message and the data by building a strong sense of enterprise collaboration and a focus on results that are best for the enterprise as a whole, not for a particular department. “When messaging is more line-of-business oriented, there is more incentive to keep data in silos. When it’s more enterprise-oriented, there is more incentive to leverage everything you have,” Acxiom’s Watson says.
Pioneers and silos
Not all data silos are inherently bad. In fact, many of today’s silos were yesterday’s pioneering concept, born in a forward-thinking department or division to follow a path the rest of the brand could not walk. Although poor planning can certainly create a siloed environment, they are also the product of innovation—an attempt to do something new for the brand, something untried, with tools, techniques, and campaigns that were simply unavailable to the broader enterprise when the project began. “In the late 1990s, when it was time to get email and Web operations up and running, [brands] set up silos with good reason: it made the teams more agile, they weren’t burdened with legacy infrastructure and processes,” IBM’s Henderson says. “Silos allowed us to get online quickly and get operations up and running.”
The lesson, then, is not that every new customer intelligence initiative must be part of a monolithic database. Rather, any new endeavor should be seeded with a data structure that is compatible with the master records of the parent database, and designed with an exit strategy—a clear concept of how the data, actions, and insights produced in the new silo can be repatriated to the broader enterprise if it succeeds. “If the experimental silos start with good master data, they have a better chance of succeeding or failing on their own merits, in an environment that can later be replicated in the rest of the enterprise,” says Robison of Gartner.
Just as pioneers can create silos, so too are they often responsible for the most productive and effective integration and reconciliation efforts. “A division or department will lead the way into more holistic data analysis and a higher degree of customer experience 90% of the time,” SAS’s Raj says. “Rarely does the C-suite mandate that we look more carefully at our data.”
Fixing the relationship between marketers and data requires compromise. Don’t fear it. Let go of the passion for a “single view of the customer” and the need to assign statistical significance to billions of data points. Create visionary campaigns that engage customers and challenge the brand’s ability to deliver in new ways. The odds are that the necessary data is already waiting, and those aggressive business goals will provide the framework necessary to corral it.
