In our last post we mentioned a fascinating statistic--
While 56% of marketers say the Internet is key to their business, only 15% of them actually measure results.
A common misconception about marketing is that it is all art and no science. In reality, most types of marketing programs are measurable to one degree or another. True, the ROI for a billboard designed to enhance brand recognition might be difficult to quantify. But at the other extreme, search engine marketing is highly quantifiable--one reason for its popularity and rapidly improving effectiveness.
There is no shortage of data available for analyzing Web site traffic, user demographics, customer behavior, and pay-per-click campaigns. To make full use of Web analytics, companies should employ techniques that drive the right kind of data. For example--
1. Newsletters always seem like a good idea, but how do you know if they work? Try sending out two versions in an A/B split test and see which one produces a higher conversion rate. Repeat the process with every newsletter release.
2. Use A/B split testing for PPC campaigns, creating multiple landing pages to test which message is more effective.
(When split testing, take care to change only one variable at a time. Otherwise, it becomes difficult to determine what caused user behavior to be different.)
3. For Web pages having a steady stream of fresh content, offer RSS feeds to encourage subscriptions (that can be tracked and analyzed).
4. Speaking of Web pages, make sure each one has a specific purpose, and as much as possible, only one specific purpose. If users hit a particular page with high frequency, it will be hard to determine why if the page contains several types of information.
Apply a little "scientific method" to the creative work of building Web sites and search campaigns. The customer insights you'll gain down the road are indispensable!