Wharton@Work

February 2014 | Marketing

Effective Internet Marketing: Here’s How

Wharton Marketing Metrics™: Linking Marketing to Financial Consequences has helped executives understand how their marketing initiatives affect their bottom line for over ten years. During that time, the program has evolved, keeping up with changes in strategy, channels, and customer behavior. Faculty Director and William Stewart Woodside Professor of Marketing David Reibstein says the evolving landscape requires new strategies and measures to assess performance — social media as a marketing force is just one example. “Everybody is chasing after social media, but it is still a mystery to most people. An abundance of firms are diving into the fray of social media without any idea of whether it is productive or not. But there are many new measures available to assess online spending and whether it is delivering value for the firm. Bringing Kartik Hosanagar into the program will help participants better understand this new frontier.”

Wharton@Work recently spoke with Hosanagar, who studies online advertising, to learn about his latest research and what he brings to Marketing Metrics.

Wharton@Work: What are some of the major challenges facing companies as they move from traditional channels to internet advertising?

Kartik Hosanagar: The most important challenge is that the online channels are significantly more data-driven than traditional channels. Online channels are more measurable and deriving the best ROI from the channel requires a data-driven approach to advertising. Many firms find it hard to navigate that shift in approach.

W@W: Is that something you cover in Marketing Metrics?

KH: Yes. I discuss issues related to managing online advertising campaigns, including measuring their effectiveness and increasing user engagement through initiatives such as personalization and customer analytics.

W@W: What are some of the ways in which marketers can measure the effectiveness of internet marketing?

KH: Online marketing initiatives are more measurable, thus giving marketers a great opportunity to measure their effectiveness. For example, we can measure whether users interacted with the ad by clicking on the ad. We can measure the quality of those clicks by measuring how much time those users spent on the website after landing there from an ad, or how many different pages or products they viewed on the website. We can even measure the rate at which these clicks converted to purchases or to repeat visits. All these measures ultimately allow an advertiser to determine which online marketing campaigns are delivering results and which ones are not.

W@W: How can companies work to convert a page viewer into a sale?

KH: There are several approaches. One that I’ll single out is personalization. The opportunity on the internet is that we can track a user’s online behavior — which products they view, which ones they buy, etc. — and personalize their experience based on their interests. There is ample evidence now that personalization increases repeat visits, engagement, and purchase frequency. In some environments, I have seen engagement and purchase measures go up by as much as 40-50 percent post-personalization.

W@W: Some of your recent research concerns social media. What have you learned about what works and what doesn’t?

KH: Given the amount of time users spend on social media, most companies now have social media presence. While these firms actively generate lots of content on their social media accounts in the hope of engaging consumers and improving brand awareness, there exists very little insight on what kinds of content engage or do not engage the consumer. In practice, only a small proportion of posts by companies get any significant engagement. In this scenario, the role of content engineering — the process of crafting a compelling message and fine-tuning the message to the target audience — becomes very important.

In a recent study, we analyzed over 100,000 unique Facebook messages posted on nearly 800 company Facebook pages. Using state-of-the-art Natural Language Processing (NLP) techniques, we content-coded each post into a set of constituent advertising attributes, including product informative (mentions of prices, availability, and product features) as well as persuasive content attributes (e.g., emotional and philanthropic content).

We found that inclusion of persuasive content — like emotional and philanthropic content — increases engagement with a message. We also found that informative content — like mentions of prices, availability, and product features — reduces engagement when included in messages in isolation, but increases engagement when provided in combination with persuasive attributes. Persuasive content thus seems to be the key to effective engagement.