Marketers recognize the value of using data to make smarter decisions, a trend borne out by recent research surveys. Questions to 700 marketers showed 68 percent said they plan to increase their spending on data-related marketing expenditures this year, according to the survey sponsored by Yesmail Interactive and Infogroup Targeting Solutions. A 2012 study by the New York American Marketing Association (NYAMA) and Columbia University researchers shows a consensus, too: 91 percent of the 253 marketers interviewed said they believe successful brands use customer data to drive marketing decisions.
But while it’s becoming clear marketers feel the need to embrace data-based decision making, they face a number of potential pitfalls in implementing analytics to support their goals. Marketers must work with new technology and build new alliances inside and outside their organizations at a time when the velocity and complexity of marketing programs is accelerating, says Mark Price, managing partner at M Squared Group, a data-driven marketing consulting firm in Eden Prairie, Minn.
At the same time, Price says, the “accepted behaviors” of marketers are also changing. “Marketers have traditionally only had to evaluate their programs on the broadest of measures—share of market over time, change in customer awareness, positioning, etcetera,” Price says. “Now they must learn a new skill set and understand that the new skill set is what is valued the most at the senior levels of the organization.”
Price and other experts identified eight common mistakes enterprises should avoid when instituting marketing analytics in their businesses. Watch out for these issues that can limit your company’s return on investment:
1. Failing to integrate and analyze all available customer data.
The marketing group can learn a lot about customers through data it collects. But marketers will get a limited picture if they don’t incorporate information gathered through other departments like those that handle research, e-commerce or the call center. Consolidating a company’s data assets to get a holistic view of customers can make marketing analytics so much more powerful, says Mike McGuirk, partner at the marketing analytics consultancy iKnowtion, which is a business unit of TeleTech Holdings.
“The more information we gather from all these different groups inside of a business, the better understanding we’ll have of that customer so that we’ll be that much more targeted with our next communication that we send to them,” he says.
Consider this example from McGuirk: A cell phone operator may think a customer is perfectly happy with the service based on reports that show consistent use of monthly minutes over the past six months. But what if that person recently called the contact center to complain about multiple dropped calls? If the usage data isn’t connected with the call center data, the firm’s marketers won’t recognize that customer is at risk of leaving. If it is integrated, marketers could react quickly, perhaps giving the customer extra minutes for the next month or sending a communication explaining how the company is making improvements to avoid future dropped calls.
2. Not asking actionable questions.
It’s understandable that some decision makers would seek wide-ranging insights. But it’s a mistake to frame questions that are too broad. “Analytics are a way to answer business problems but many times people start by saying, ‘Tell me everything about this,’” says Price. Instead, he says, it’s important to “focus on what is the business problem you need to solve and what information do you need to have to be able to answer it in a way that will let you take action.”
Two well-defined and actionable questions are: “Which of my customers have stopped buying my products in the last six months?” and “What did they used to buy?” Armed with that knowledge, Price says, “I can use offers related to those products in the communications I’m going to send to [lapsed customers], so by doing that I’m going to get a higher level of response and I’m going to make more money.”
3. Only exploring quick hits.
Don’t jam the analytical pipeline with too many short-term problems that may not have a strategic impact on the business or on marketing operations—requests that consultant Peter Vandre refers to as “fire drills.”
Say, for instance, a senior executive wants to know how the profile of customers who bought widget A differs from the profile of those who bought widget B. That kind of question can be answered with the data but may be hard to act on in a way that creates major value for the business, Vandre says.
In contrast, Vandre says a request that can have a longer-term impact is an analysis to identify, implement, and automate 10 high-value email trigger programs based on key customer behaviors.
“When we work with clients what we try to do is say, ‘Thirty percent of our time is going to be spent on fire drills, but the other 70 percent needs to be on things we prioritize that are going to have a longer-term impact on the business,’ and usually those things just take longer to analyze,” says Vandre, who is vice president, digital analytics at the CRM agency Merkle.
4. Using analytics to justify marketing decisions rather than drive change.
Sometimes, when analytics sits within marketing groups, its function turns into making a group shine instead of always providing objective insight, says Vandre.
“You mine the data to find the one thing that looks good with the program and ignore the other 20 things that don’t look so good with the program,” he says. “What ends up happening is analytics actually is detrimental in that case. Not only are you focusing potentially on some of the wrong things, but you’re also reinforcing things the company really shouldn’t be doing.”
Marketers may celebrate the 10,000 new email registrations they attained through a contest, Vandre explains, while ignoring that half of the registrations were from invalid emails and only 2 percent of registrants ever opened an email sent to them.
It’s important to design marketing efforts in ways that facilitate objective analysis, he says, and to have the discipline to not only define upfront what success looks like, but also to adhere to that throughout the process.
5. Letting silos stifle communication between marketers and analysts.
Vandre describes a scenario you don’t want to experience: “The marketing organization says, ‘Here’s our problem,’ and the analyst doesn’t have a lot of exposure to what’s happening day-to-day within the programs being executed there so they make recommendations naively, not really understanding what’s possible or not possible.”
Analysts may tell marketers they should spend more on branded paid search, for example, only to discover the team has already bought all the keywords that exist, Vandre says. Integrating analysts into marketing departments can reduce those kinds of glitches by making them aware of existing research projects and data sources.
6. Not applying marketing skills internally.
Marketers must build credibility and buy-in with the people who are going to help them execute on the marketing plan, like those in the C-suite and in departments like sales, operations, finance and IT. “You have to build support with those people to succeed or you will fail,” Price says.
For example, marketers should work with the finance team to determine how to evaluate the success of a marketing program before it launches. “If a common framework for evaluation can be established, the CFO can become the voice of validation to the CEO about marketing results,” he says.
But marketers also need to run what Price calls an “internally focused marketing campaign.” He says it starts with determining which constituents are important to influence in the organization, their existing beliefs about marketing, and the metrics that will matter to them. Next, Price says, marketers should take a customized approach – deciding how often they’re going to send individual decision makers updates about the status of marketing programs and choosing which vehicles will work best for each person, perhaps opting for newsletters, PowerPoint presentations or personal updates.
7. Mismatching human resources.
You need to find the right people who can execute your analytics vision for marketing. Vandre knows of an insurance company that hired some smart statisticians who were expecting to build predictive models, do advanced segmentation or design tests and experiments. But the company had them designing business reports and mining them for insights – tasks they weren’t well suited for and didn’t do very well – and most of the statisticians left within a year after they were hired.
“It’s the classic case of understanding the kinds of problems you’re going to need to answer and building the organization that can answer them,” Vandre says. You may find that for your company, it makes more sense to outsource various functions like predictive modeling or Web analytics.
8. Hiring vendors and consultants who are inexperienced in your problems.
When considering outside help, don’t base your decision on the number of Ph.D.s in a company, says Anil Kaul, CEO of the analytics and research firm AbsolutData in Alameda, Calif. “Yes, you do need a marketing analytics company to have technical chops, but you also need a company to understand your business and industry,” he says. “Do they understand how analytics can be translated into a business impact?”
Also consider whether the service provider has enough capacity and scale to serve you. “Analytics is one those businesses where a lot of people can hang their own shingle,” Kaul says. “The problem you run into is if that company doesn’t have scale and they get busy with another client, your projects get affected.”