Why You Suck at Web Analytics
Many organizations make at least some of the 10 common web analytics mistakes outlined by analytics expert Adam Greco at a recent online marketing event. Luckily, he also mentioned some fixes for the mistakes.
Many companies spend pots of money on analytics for their websites and social media channels. Yet too often they don’t know exactly what they’re looking for in the data, they don’t fully understand the analytics software tools, or they haven’t “tasted the soup,” to borrow a line from an Eddie Murphy joke.
“When you’re bad at web analytics, you’re throwing money out the window,” said Adam Greco, senior partner of Web Analytics Demystified, a Portland, Ore. consulting firm. Greco spoke on the topic of “Do You Suck at Web Analytics?” at the recent Online Marketing Summit 2012 in San Diego.
In his presentation, Greco outlined common mistakes organizations make with web analytics and what you can do to improve:
1. Incorrect business requirements. An understanding of your organization’s business needs is essential to knowing what to measure and what to look for in the data. “Too often, people lose sight of their business requirements,” Greco said. “You have to remember why you’re doing web analytics in the first place. If you’re not sure, stop what you’re doing and ask yourself, ‘Is this data answering the key questions we have about our business?’”
2. Poor data quality. Greco said he often sees analytics reports with gaps in the data. In addition, people don’t always stop to closely check the data they’ve acquired. Greco said he’s done surveys in which he’s asked web analytics team members such questions as “Do you believe your data is correct?” and “Are you willing to stake your reputation on the data?”
“You get three strikes with web analytics data,” Greco said. “If your data is wrong the third time, no one will trust it again.” In severe cases, Greco said you may need to shut down the web analytics process you have in place and start over until you get it right.
3. The wrong web analytics team. Poor data quality is easy to blame on web analytics software or the software vendor. But in most cases, Greco said, it’s not the software’s fault. “It’s more about having the right web analytics processes and people,” he said.
In general, you should have a web analytics architect who has met with the business side and determined key performance indicators (KPIs) to track. You’ll also need a technical implementer who understands the tools; a data quality person who verifies the data is accurate and relevant; a “report monkey” who cranks out reports from the data; one or more web analysts who offer their insights on the data; and an evangelist who can enthusiastically share the data and its importance with others in the organization.
If you don’t have the team members you need, don’t assume you can’t afford to add more, Greco said. “One of the biggest excuses I hear is that it’s too expensive to hire a web analytics person,” he said. He explained that he had hired a junior-level web analytics person via Twitter. “I hired someone who was pretty green and within six months, our data quality acceptance rate went from 17 percent to 90 percent.”
4. Lack of tagging skills. You can’t do web analytics reports without proper tagging, Greco advised. If you suck at tagging or it takes too long to get tags on your site, use a tag management system such as Tealium, Ensighten, or TagMan, Greco said. “You can outsource to a service that will help you set up your tag management once, and then you can take it from there,” he added.
5. Cubicle syndrome. Too often, web analysts spend a lot of time in their cubes. “They need to be out talking to the people in the business side of things, talking to IT, talking to customers, talking to executives,” Greco said. He suggested embedding web analysts in other teams of the organization “to find out what those teams want, so you can do better tracking and analysis.”
6. Not tasting the soup. Greco recited a joke from the Eddie Murphy film Coming to America. In the scene, a restaurant patron who’s been served soup says to his waiter, “Taste the soup.” The waiter asks if something’s wrong with the soup, to which the patron says, “Taste the soup.” The waiter now wants to know if the soup’s too hot.
Patron: “Taste the soup.”
Waiter: “Is the soup too cold?”
Patron: “Taste the soup!”
Finally, the waiter says he’ll taste the soup. “But where’s the spoon?” he asks.
Of course, the patron’s “taste the soup” request was his way of pointing out the waiter had failed to give him a spoon. Greco said he sees this scenario played out in web analytics. “We try to solve the problem before we understand what the problem really is,” he explained.