More than ever before, companies have at their disposal immense streams of customer data. How can companies, particularly those in retail, make use of such data to capture market share?
Kartik Hosanagar, an Operations, Information, and Decisions Professor at the Wharton School, highlights in his paper “How Do Recommender Systems Affect Sales Diversity?” the benefits of using personalization, social media, and rich metadata to attract customers in an online retail environment. Not only do these tools help companies target the right customers, but they also significantly improve the customer experience. And a social media presence alone is no longer good enough; companies have to get smart about producing the right type of content to boost customer engagement.
Professor Hosanagar sat down with the Mack Institute to discuss his findings. A transcript of the interview is below.
One of the things that the internet and digital media in general have enabled is access to vast amounts of customer data. Customers leave behind digital traces on websites and increasingly even in the brick and mortar world. When they leave behind that kind of data, the question is how can managers and how can businesses use it efficiently? My focus really is on looking at both what we can do with customer data and also what kinds of quantitative tools we can apply to use that customer data.
I also look at how to apply the same kind of data to efficiently find customers of interest through online marketing channels. For example, on sponsored search, say a search engine like Google, what kinds of key words might be of interest to a business? How should they participate in auctions run by Google? How could they use online marketing channels more efficiently and allocate their ad dollars across social media sponsored search and other digital media more efficiently?
Generally, a lot of my research shows that customer analytics, big data and analytical machinery can provide significant value. For example, we find that the use of these kinds of tools can improve product or service fit for consumers by a really significant amount. If we look at online personalization, we find a significant increase in customer satisfaction from the use of personalization tools. That increase in customer satisfaction also ultimately results in better customer retention or higher sales. We find that in e-commerce, a lift in sales from personalization systems can be as high as 33 percent. This is a significant increase in value for both the firm that deploys this technology and also for the consumer that ultimately benefits from such technology.
“We find that personalization systems can change who are the winners and losers in terms of the market share battle.”
Another facet of my research is looking at the application of these tools for online marketing. Again, we find that the use of these kinds of tools can help improve the return on investment from marketing dollars by as high as about 33 percent. That happens because marketing messages are targeted to the right person and the right kind of message is shown to that person. For the consumer, the value is they don’t see messages that are ultimately irrelevant for them.
Finally, my research also finds that not only are these kinds of tools and algorithms impacting consumer choice, but it also has a pretty big impact at the market level. We find that personalization systems can change who are the winners and losers in terms of the market share battle. For managers, it raises interesting questions about how best to use these systems and how best to plan your strategy so your products are likely to get discovered by personalization algorithms, consumer reviews and ratings, social media tools and other digital mediums.
For managers, one of the implications is that it’s not sufficient to merely get the product built and get it out there. A lot of discovery, whether it’s for media or for products, is happening through digital tools for consumers, whether it’s social media or search engines and personalization algorithms. For managers, they have to increasingly incorporate digital strategies into their overall marketing strategy and get savvy about how to leverage digital channels. They need to be really thinking about how their choices will allow their products to get discovered. That might mean sometimes providing rich metadata about their products so that personalization algorithms might be able to pick up these products. To get discovered by these similarity algorithms, you need to provide a lot of rich data so that they can figure out what products your product is similar to.
“Now that everyone is there, just being on social media doesn’t cut it. Now [companies] have to ask themselves what kinds of messaging works on social media.”
My research on ratings for products on online websites shows that product ratings tend to reinforce themselves. Consumer ratings are not often independent. What we find is that when people are about to post a rating for a product, they get influenced by prior ratings. This creates a certain path dependency where your first few ratings impact future ratings. This again implies that while managers acknowledge that product reviews and ratings are important, it really emphasizes the need to seed this the right way, to focus heavily on getting your best customers to provide early reviews and ratings. That helps really set the sentiment and tone for future reviews, as well. There is a sense of optimal timing in terms of when you want to focus on reviews and ratings on your website.
Similarly, on social media, one really needs to think about what kinds of messages work in social media. A lot of focus in the last few years for a lot of firms has been to really establish a social media presence, but now that everyone is there, just being on social media doesn’t cut it. Now they have to ask themselves what kinds of messaging works on social media. We find, for example, that a lot of companies tend to discuss their products a lot on Facebook and Twitter and these other social media platforms, and those kinds of posts have the least engagement. It’s instead when they showcase their brand personality, when they have emotional appeal to their products, and when they can really go from being a product to having a certain personality associated with the brand, that’s when people engage with them on social media.
Find more interviews about Mack Institute research here.