Big Data is the current shiny object when it comes to data-driven marketing. “But not all that data matters,” says Joe Stanhope, a principal analyst at Forrester Research. So marketers need a strategy to figure out which data does matter. That starts with having a specific goal for using the data, Stanhope says—whether it’s to, say, understand how a campaign is performing, improve segmentation, predict lifetime value, or target specific customers.
Next, examine the data to decide which elements from each source are necessary for what will be measured, Stanhope advises. Marketers who take this approach may in the process find that data they thought was worth saving really wasn’t. “That’s what a lot of companies that manage so-called Big Data very well get to be very good at,” Stanhope says. “Curating that inbound flow of information.”
Big Data isn’t the answer to everything
“I’ve been hearing from a lot of executives, from CEOs and CMOs, ‘Hey, get me that Big Data,’ without them understanding what exactly it is,” says Wilson Raj, global customer intelligence director at SAS. “The reality is that, obviously, no business will improve just because you have data or Big Data. You have to use it. That’s where the magic comes in.”
“Pick a very cogent, pressing business issue,” Raj says. “Maybe it’s acquiring new customers. Whatever that challenge is, focus the data questions and analytic questions on that particular business issue and solve that first.”
Not all companies are even ready to use Big Data. “We’re a few years off from the democratization of Big Data,” notes Stewart Pratt, director of data and analytics at SapientNitro. “Big Data is great. It’s created incredible marketing efficiencies for large players like Best Buy [and] Staples….” But, he says, many companies haven’t yet exploited all the information available from unstructured data. Nor do they have the people, processes, or technology in place do so.
And that’s a good reason to get back to some of the basics of data management. Those include data quality and optimizing list buys. Other issues to pay attention to: boosting personal data analysis skills and finding ways to deal with the shortage of marketing-analyst talent.
1. Keep on top of data quality
In Raj’s view, two data-quality issues are important now: data integration and effective usage, while the latter encompasses security issues and compliance. The two areas in fact go hand-in-hand. “There’s a lot of compliance and security risk that comes with integration,” Raj says. “That’s something else that marketers should think about.”
Since important customer information can come from non-marketing areas of the business, marketers should stay in close contact with IT. “It’s important that marketing and IT leaders…have a shared understanding of what customer issues they’re going to solve,” he says. It’s a good idea for the CMO to connect with other members of the C-suite, too. For data quality:
Don’t think you have to put all your data into one big database to access it or keep it consistent. Companies that spend so much time focused on integration neglect to dive in and use the data to its fullest potential, or at all. “You get to a point where you’re never pulling the trigger to act,” Raj says. A federated approach puts a layer on top of the various databases that lets each department see only what it needs to, including data from other departments or functional areas that may be relevant. “You get a faster time to market, smaller IT footprint, and start getting benefit immediately,” Raj explains.
Do weight the data based on how trusted the source is. “That’s a big factor in how we manage quality,” says Jim Swift, CEO of Cortera.
Don’t neglect to use the data as part of the maintenance process. If you don’t, “you can just introduce more junk into the system,” Swift says.
Do keep multiple iterations of data. This can help ensure that the current data is correct, as well as provide the ability to reference past data that may be pertinent. “I can keep different versions of the address in date ranges, so I can zero in on which is most likely to be the more current address,” Swift says.
2. Use lists, carefully
Catalog and print-magazine subscription lists were the go-to lists for years. Today there’s an ever-increasing number of lists created specifically to be sold. But there’s a problem with many of these newer lists. “You’re not getting buyers,” says Jay Schwedelson, president and CEO of Worldata. “You’re getting people who respond to opportunities. And so the data quality has really been diminished.”
Some companies that create these lists to be sold are aggressive in their data-collection efforts, which can result in privacy issues. List members may have opted in, but it’s not always clear what they’ve opted in to, he says.
Append lists can be a problem, as well. The email addresses might not represent each person’s best address, for example. “[Append is] best utilized where you’re trying to get into a [new] space,” Schwedelson says.
Ways to use lists well:
- Buy only lists that have a clear data trail.
This will help to ensure that the information has been collected properly. “This may impact your volume, but you’ll get a better response rate, and you’ll have…a program that’s not going to run into delivery issues,” Schwedelson says.
- Refine email append data selectively.
Since 20% of your customers likely comprise 80% of your business, it’s worth finding the best email addresses for that 20%. Assigning high-value customers—or any customers—an email address without ensuring that it’s the right one, “you’re really corrupting your own customer file,” he says.
- Collect Twitter handles.
If customers have active Twitter handles, add them to the appropriate database. “The engagement level [of Twitter users] is through the roof,” Schwedelson says. The downside is having more remove requests and complaints from vocal customers.
- Drive action.
Build a sense of urgency or exclusivity into communications. If your objective is to prompt a quick response via email, provide offers that expire in 24 hours.
3. Become a data analyst
In the past the view was that it was difficult or impossible to measure marketing. But the budget cuts of recent years, coupled with advances in technology, have changed that. Not only has marketing become increasingly measurable, but senior management’s expectation that marketers will track metrics and prove ROI has grown, as well. For some marketers, that’s changed the skill-set needed. “The modern marketer either needs to become more technical…or move into the strategic, data visualization, business information space,” says SapientNitro’s Pratt.
Ways to be more data savvy:
- Explore data using existing tools.
If it’s just a spreadsheet, that’s fine, says Swift of Cortera. Often, once marketers see the opportunities presented by digging into the data themselves, they get hooked and look for ways to enhance their analytics skills.
- Research analytics tools designed specifically for business users.
An increasing number of vendors are providing data analysis and data visualization applications specifically created for marketers.
- Use outside partners.
Boutique marketing shops and many larger marketing agencies and customer strategy-focused consulting firms have an analytics practice. “Start to have some of these flex resources look at the data with you and help you understand what it is that you’re trying to target,” he says.
- Take advantage of vendor training.
Most marketing vendors offer training to build engagement and retention among their clients. Aprimo, for instance, provides marketing analytics instruction at its user conference, and Cloudera provides its customers education and certification through its Cloudera University.
- Get educated.
Attend data conferences, join an association for marketing analysts, or enroll in marketing sciences classes at universities like the MIT, Berkeley, or DePaul.
4. Find analytic talent
With all of the customer data available today from a number of sources, it’s often difficult to find the expertise needed to analyze it. But it’s urgent that companies do so. According to ITSMA’s 2011 Online Survey on Data-Driven Marketing, which found that 82% of respondent companies reported improved market share from being data savvy.
“The big challenge is that a lot of folks…in marketing services have a lot of human capital invested in older platforms,” Pratt says. Training those people on the newer, less expensive platforms is a must, he adds, noting that recent college graduates who’ve trained in analytics tend to have also acquired computer skills in newer, related platforms.
According to Pratt, too many organizations hire analysts but fail to provide a career path for them, which is integral to retention. “How do you help them develop their career, so that they don’t just spend their lives cranking out reports and dashboards, which…nobody wants to do.”
If you don’t have that expertise in-house, you may have to look elsewhere. Aprimo’s VP of marketing Paige O’Neill says the following signs may indicate the need for an outside data analyst:
- You don’t know where to start.
In that case, it makes sense to have someone come in and help you set up best practices for data analytics.
- You can’t get perspective.
“Sometimes, as marketers, we get so close to the data that we’re not able to see things that it might be telling us, or we’re trying to lead [the data] in a different direction,” O’Neill says.
- You’re dealing with new types of data, such as social.
“The more experience you have, I think the more danger you have…of [making] assumptions that may not prove true once you really put data up against them,” she says.
- There’s something you want to know more about.
In that case, a third party can help formulate the questions and find answers.