|
|
Evaluating, Implementing, And Learning New Analytics Tools
By Cecily Robyn Lough
Expert Author
Article Date: 2009-12-29
With both large real and hidden opportunity costs associated with evaluating, implementing, and learning the quirks of a new analytics tool, it only makes sense to ensure that the tool one chooses will be able to keep pace with one´s future requirements.
So in order to evaluate tools correctly people need to be aware of the changes that are occurring in the online world as this will influence what data will be important to collect and leverage in the future. In this article I wanted to outline some of the top trends that are impacting the online marketing world so that when someone makes an analytics tool evaluation they are creating a framework for the future.
Part 1 Setting the Scene: Data Trumps Everything
We can even take a step back before this evaluation process to ask on a high level whether the benefits in general of implementing an analytics solution outweigh the costs. The answer becomes evident if we look at the events that have happened in the last year; in 2009 more people than ever agreed that data driven organizations are much more effective than those based on HiPPO (Highest Paid Person´s Opinion) or other strategies.
A March 2009 Accenture Survey of executives working at U.S. companies with annual revenues of $500 million or more found that high performance businesses are five times more likely to use analytics than lower performers. Additionally, 70 percent of high performers identify analytics as a significant decision support tool.
The overarching example is Google: Marissa Mayer, EVP, has said that "we let the math and data govern how things look and feel" and this policy has made them extremely successful. The extent to which this is true can be seen by an article on the front page of the business section of the New York Times on May 9, 2009, that quotes employee Douglass Brown saying that he could not even decide whether a line on a web page should be 3, 4, or 5 pixels wide until he had tested all 3 versions. In the end, Brown left Google since the engineering culture was "not friendly to designers" since all of his decisions were asked to be backed up with data. This was only newsworthy because Brown argues against what has become acceptable practice at Google as well as other high performing businesses: data trumps everything.
In fact, in the last few years, data driven decision making has been validated for all size organizations. For a larger organization with significant amounts of traffic, making even the smallest changes can have quite a large positive impact on revenue. "More is different" due to the large volumes of traffic involved. Therefore, it is important for a high traffic site to find an analytics tool that can handle billions of transactions while presenting granular data in order to make decisions and analysis based on more detailed reporting. For example, it might be interesting to have more detailed knowledge of all of the marketing initiatives that contributed to a sale besides what happened during the last click in the last session before a purchase. Or it might be interesting to know exactly which customers put which items into their shopping carts and then abandoned them, in order to send follow up email marketing campaigns.
On the other hand, a small, start up organization can also use an analytics tool with equal effectiveness. One way would be for the product manager to measure how much additional traffic is gained by adding different features and later doing some comparative testing to understand which feature sets are optimal. Then as one works to grow traffic, the marketing manager might also use analytics for tracking the effectiveness of email campaigns and ad inventory. In fact, Andrew Chen (a former Entrepreneur-in-Residence at Mohr Davidow Ventures) recommends viewing analytics as a necessary tool to be used during your start up product development process, no matter how small your company might be, in order to confirm or deny assumptions you have made on why people stay engaged on your site. He asks, "in reality, is it better to build 10 features where you do not know what worked and what did not, or is it better to build LESS features but have a clear sense of why it worked?"
Comments
About the Author:
Cecily Robyn Lough has over 15 years experience in pulling actionable insights from online marketing data. She is currently Director of International Sales at Webtrekk GmbH, a leading Web Analytics company based in Berlin, Germany. Please contact her through LinkedIn or at cecilyspeaks@gmail.com.
|
|
