BrandStand » Using Analytics to Increase Online Marketing Effectiveness

April 25, 2009 | Steve Chitwood

Using Analytics to Increase Online Marketing Effectiveness

Tom Davenport, the noted Babson College and Harvard Business School scholar, has written extensively on how companies that effectively employ analytics to leverage business intelligence data into fact-based decisions can build a competitive advantage. His team has researched large organizations such as Proctor & Gamble, Bank of America, Wachovia, and Harrah’s, and makes some points that translate well to businesses of all size and the importance of capturing and, more importantly, using analytics data.

Most companies already collect analytics data, and analysis tools are readily available. In the case of website traffic and online marketing initiatives such as search and email marketing, tools such as Google Analytics and email tracking software are commonplace, but few marketers leverage the data to its fullest extent through sophisticated analysis The differentiator, just as Davenport states regarding large complex multi-national organizations, is the level of reliance on analyzed data when making decisions. Organizations that build an “analytical culture,” where data and sophisticated statistics are heavily relied upon in decision making, will build a competitive advantage in the marketplace.

Let’s apply this principle to a hypothetical case study about online customer acquisition.

Both Company A and Company B have websites, purchase online display ads, engage in search engine marketing (SEM) via Google Adwords, and regularly communicate with customers and prospects through ongoing email campaigns. Both are contemplating a blog to establish several key individuals as subject matter experts (SMEs), have a company LinkedIn page, and encourage employees to participate in online conversations. Both companies wonder if Facebook and Twitter can help them in their efforts. Neither have immediate plans to initiate a mobile device marketing initiative, but wonder if they should. Company A and Company B are very average. They each also collect similar amounts of tracking data.

The Company A marketer knows how many people came to his website last month and that traffic increased by 13% over the previous month. He knows that his monthly email newsletter was sent to a list of 14,000 current customers and prospects, that the open rate and click rate were well within expected thresholds, and showed a slight increase over last month. He congratulates himself on the new design that he conceived. Finally, he knows that he spent $11,000 on a Google AdWords campaign during Q1 and that the click-through rate steadily increased throughout the process. Overall, web-based sales and sales inquiries are trending upward. He is happy, but more importantly his boss is happy and sent him a very complimentary note.

Company B is a different place. The marketing executive has diligently worked to create an analytical culture. She sees that her website traffic is up by nearly 20% and can attribute the spike in activity to a particular segment of users. She executes a targeted email campaign that highlights a specific product line that users in this segment have historically purchased. The predictive model she uses to track trends over the past years suggests that a similar spike is likely over the next several months with a different segment of her users. She confirms the planning meeting for the related campaign.

After lunch she reviews the results from the multivariate testing (MVT) study that she has been running on the various landing pages used in her AdWords campaigns. Noticing that users are converting 2% better on the page that includes a six-word headline and a call to action highlighted in blue versus the eight-word headline and the image she was very attached to, she orders the change. Though the click-through rate on the campaign was acceptable, the conversion rate she had been watching needed to come up. She now knows that her cost-per-conversion will slightly decrease. She plans another MVT experiment for next month.

Finally, at the end of the day, she reviews the samples for the three email newsletters that are scheduled for delivery next week. The first, which goes to past customers of her core product in the New England market, needs a copy edit. Another, going to prospects who downloaded a sales sheet via the website, looks good. The final one, set to deliver to customers who have been identified as targets for the recent release of an updated accessory product, will just have to wait. The CEO has invited her to share her insights with the directors at their meeting in Paris.

Davenport points out that historically, the “business as usual” model has meant taking whatever customers that came and looking backwards on financials to understand what happened last month. Companies who leverage analytics have gotten rid of this definition of business as usual. A distinct competitive edge is available to organizations that decide to use data and sophisticated statistics in decision making and who are no longer satisfied with the same kind of information others use.