Working with Data

In the days of traditional media, actionable data was a highly desired but scarce commodity. While it was possible to broadly understand consumer responses to marketing messages, it was often hard to pinpoint exactly what was happening and why.

In the digital age, information is absolutely everywhere. Every single action taken online is recorded, which means there is an incredible wealth of data available to marketers to help them understand when, where, how, and even why people react to their marketing campaigns.

This also means there is a responsibility on marketers to make data-driven decisions. Assumptions and gut feel are not enough; you need to back these up with solid facts and clear results.

Don’t worry if you’re not a "numbers person." Working with data is very little about number crunching (the technology usually takes care of this for you) and a lot about analyzing, experimenting, testing, and questioning. All you need is a curious mind and an understanding of the key principles and tools.

The sections below outline the data concepts you should be aware of.

Performance Monitoring and Trends

Data analytics is all about monitoring user behavior and marketing campaign performance over time. The last part is crucial. There is little value in looking at a single point of data—you want to look at trends and changes over a set period.

For example, it is meaningless to say that 10 percent of this month’s web traffic converted. Is that good or bad, high or low? But saying that 10 percent more people converted this month, as opposed to last month, shows a positive change or trend. While it can be tempting to focus on single "hero numbers" and exciting-looking figures ("Look, we have five thousand Facebook fans!"), these data really don’t give a full picture if they are not presented in context.

Big Data

Big data is the term used to describe truly massive data sets—the ones that are so big and unwieldy that they require specialized software and massive computers to process. Companies like Google, Facebook, and YouTube generate and collect so much data every day that they have entire warehouses full of hard drives to store it all.

Understanding how it works and how to think about data on this scale provides some valuable lessons for all analysts.

  • Measure trends, not absolute figures—The more data you have, the more meaningful it is to look at how things change over time.
  • Focus on patterns—With enough data, patterns over time should become apparent. Consider looking at weekly, monthly, or even seasonal flows.
  • Investigate anomalies—If your expected pattern suddenly changes, try to find out why and use this information to inform your future actions.

Data Mining 

Data mining is the process of finding patterns that are hidden in large numbers and databases. Rather than having a human analyst process the information, an automated computer program pulls apart the data and matches it to known patterns to deliver insights. Often, this can reveal surprising and unexpected results, and tends to break assumptions.

Data Mining in Action: Target

The US retail chain Target uses data mining to market specific products to consumers based on their personal contexts. Target gathers a range of personal, psychographic, and demographic data from customers and then analyzes this against their shopping habits. They analyze their data and predict the products customers might be interested in, based on their lifestyle needs and choices (Duhigg, 2012).

Each customer is given a unique code called the Guest ID number. This is linked to all interactions the customer has with the brand, from using a credit card or calling in to the help line, to opening an email. They then gather data, including the following:

  • age
  • marital status
  • whether the customer has children, and how many
  • estimated salary
  • location
  • whether they’ve moved house recently
  • which credit cards they use
  • which websites they visit

Target is also able to buy supplemental data on their customers from other companies, including the following:

  • ethnicity
  • employment history
  • favorite magazines
  • financial status
  • whether they’ve been divorced
  • which college they attended
  • their online interests
  • their favorite coffee brands
  • political leanings

For example, Target markets baby- and pregnancy-related products to expectant moms and dads, from as early as the second trimester. How do they know to do this? Using this data, combined with customer shopping habits, Target has created a pregnancy-predictor model, determining whether a woman was pregnant (sometimes even before she knew it herself).

One father of a teenage girl complained to the store about sending his daughter marketing materials of this nature, though he apologized to the retailer after discovering that his teenage daughter was in fact pregnant, indicating that the predictor model did actually work (Duhigg, 2012). However, Target has decided to be a bit subtler in how they use their valuable insights!

A World of Data

Another consideration to keep in mind is that data can be found and gathered from a variety of sources—you don’t need to restrict yourself simply to website-based analytics. To get a full picture of audience insights, try to gather as varied a set of information as you can. Some places to look include the following:

  • online data—Aside from your website, look at other places your audience interacts with you online, such as social media, email, forums, and more. Most of these will have their own data-gathering tools (for example, look at Facebook Insights or your email service provider’s send logs).
  • databases—Look at any databases that store relevant customer information, like your contact database, CRM information, or loyalty programs. These can often supplement anonymous data with some tangible demographic insights.
  • software data—Data might also be gathered by certain kinds of software (for example, some web browsers gather information on user habits, crashes, problems, and so on). If you produce software, consider adding a data-gathering feature (with the user’s permission, of course) that captures usage information that you can use for future updates.
  • app store data—App store analytics allows companies to monitor and analyze the way people download, pay for, and use their apps. Marketplaces like the Google and Apple app stores should provide some useful data here.
  • offline data—Don’t forget all the information available off the web. such as point-of-sale records, customer service logs, in-person surveys, and in-store foot traffic.

References 

Duhigg, C. (2012, February 16). How companies learn your secrets. The New York Times. Retrieved from http://www.nytimes.com/2012/02/19/magazine/shopping-habits.html?pagewanted=all

Licenses and Attributions

Chapter 18: Data Analytics from eMarketing: The Essential Guide to Marketing in a Digital World, 5th Edition by Rob Stokes and the Minds of Quirk is available under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported license. © 2008, 2009, 2010, 2011, 2013 Quirk Education Pty (Ltd). UMGC has modified this work and it is available under the original license.