By Heidi Cohen , July 3, 2008
While many e-tailers have enjoyed rapid growth since the 1990s, the recent holiday results show this trend is starting to slow. Combined with the cloudy U.S. economic forecasts and the increasing cost of paid keywords, it's easy to conclude that we're looking at a maturing online market. This is evolving into a market where share often must be achieved at the competitor's expense and consumers are more reticent to part with their cash.
Today, online merchants must cater to increasingly online-savvy, time-crunched consumers. The five top challenges e-tailers face are:
- Competitors are just a click away. When consumers search, they have multiple options available, and many use search to navigate the Web rather than type in or bookmark specific sites.
- Visitors can disappear in 15 seconds or less. Online consumers are goal-oriented shoppers. If they don't immediately find what they're looking for when they reach your site or landing page, they're gone in under 15 seconds.
- Shopping is a multistep process. Online consumers love to browse. Many spend a fair bit of time visiting several sites just to gather information. They may also compare the offerings of several competitors before hitting the "buy now" button.
- Time between initial visit and purchase has increased. Increasingly financially challenged consumers may wait longer before buying.
- Customers wait for merchants' best offer. Having been seduced during the holiday season with free shipping and handling and other price-driven offers, consumers have been trained to wait for a special deal.
What's a Marketer to Do?
Multivariate analysis is one option that can help marketers significantly improve the effectiveness of their site pages, landing pages, e-mail messages, and shopping process. According to Mark Wachen, CEO of Optimost (recently acquired by Interwoven), over 90 percent of clients experienced double-digit percentage improvement in conversion when they initially add multivariate analysis to their marketing tools. This often translates into a seven-figure increase in revenues.
For example, Delta's multivariate optimization yielded over $20 million in incremental revenue, and Lillian Vernon's increased online sales by over $1 million. Wachen notes that changing areas that affect all transactions, such as the credit card input process, often have the biggest impact.
The Benefits of Multivariate Analysis
Multivariate analysis can improve results quickly. While direct marketers have long been well versed in A/B testing and generally incorporate it into every offline mailing they do, multivariate analysis is very difficult to implement offline. However, Internet marketing works well with this powerful tool. It:
- Requires a smaller test population. While a baseline number of test subjects is needed, this population can be used to test hundreds of options rather than just two.
- Yields results relatively quickly. By using traffic more efficiently, more permutations can be parallel-tested rather than serially testing one pair at a time using A/B testing.
- Reduces creative costs dramatically. Multivariate analysis involves testing sometimes thousands of versions of images or copy. Offline, the expense for creating these inputs can be tremendous. In the online world, technology allows you to create these versions at a fraction of the cost.
- Eliminates interactions between variables that can reduce or obfuscate results. A/B testing can only show that one option is better than another. If multiple factors have been changed, it's difficult to discern what caused that improvement.
Five factors that marketers should consider before using multivariate analysis:
- Multivariate analysis is an iterative process. Online marketing content tends to wear out faster than offline collateral. Further, online content tends to evolve over time.
- Not all multivariate methodologies are created equal. You want to select a method that allows you to customize the design to meet your needs. Use a methodology with a modern, optimal design as opposed to a preplanned one.
- What is the statistical significance? Have someone who understands the statistical implications of these types of tests assess the findings. With multivariate analysis, there's an increased chance of false positives (instances where the conclusion is incorrect). This is another reason for having an iterative process.
- You need sufficient technology resources. Since multivariate analysis tests many options, you must be able to implement the identified optimizing changes quickly to get maximum benefit.
- You can bring in outside expertise. You don't need to build or buy the technology, and you don't need to hire specialized staff to get these benefits.
While multivariate analysis may sound complicated and expensive as a marketing tool, the potential results are large and measurable. This approach has the power to make the kind of significant improvements you need in your business in the face of increasingly competitive markets. Why guess at the optimal changes necessary for your marketing and presentation, when the right combination is easily knowable? Multivariate analysis is really the only way for your online marketing to live up to its potential and make sure you're introducing only those changes that will help, not hurt.
Heidi is off this week. Today's column ran earlier on ClickZ.