Insights from MSI
What Are the Right Benchmarks for Your Social Media Metrics?
User-generated content on a brand’s social media can offer valuable information about brand performance relative to competitors, as well as inform strategic and marketing decisions. The challenge yet to be solved, however, is how to establish comparative social media benchmarks that provide a useful indication of performance. “The absence of such benchmarks,” says researcher Wendy Moe, University of Maryland, “makes it difficult for a marketing manager to interpret the meaning of, say, a brand receiving 50,000 Twitter mentions in a week, of which 30% are positive, 50% neutral, and 20% negative.”
How should a marketer interpret 50,000 Twitter mentions in a week: 30% positive, 50% neutral, and 20% negative?
Brands have unique personalities, positioning, and target customers, even within the same industry, and these produce markedly different social metrics, notes her colleague David Schweidel, Emory University: “Simply choosing one brand’s metrics at random as a benchmark for another brand would be highly inappropriate and not particularly useful regarding strategic and marketing decision making. To give such comparisons value, we need to first establish how a brand’s social media presence relates to that of others in the industry.”
Visual grid of brand presence
Moe and Schweidel devised a solution to this “mapping” issue, which they describe in their study “Integrating Social Media Metrics.” “What we did,” Schweidel continues, “is use the social media metrics to map a brand’s presence onto a latent perceptual space, that is, a visual grid that represents the ‘voice of the consumer.’” The distance between brands on the map reflects the level of differentiation in their social media metrics. Thus, the metrics of brands that fall close together on the map can be used as performance benchmarks for each other. Using the information that emerges from this map, marketers can more accurately monitor changes in consumer sentiment as a reflection of their brand’s performance.
The metrics of brands that fall close together on the perceptual map can be used as performance benchmarks for each other.
To test their model framework for perceptual mapping, Moe and Schweidel gathered two years of Twitter data for the top 20 quick service restaurants (QSRs), based on revenue, as listed by QSR magazine. In analyzing the data, for each of the 20 QSRs they identified (1) comments that mention the target brand, (2) any positive, neutral, or negative sentiments expressed in each comment, and (3) comments that mention the target brand and one other top 20 QSR brand. They then aggregated these comments into metrics they could graph onto a two-dimensional latent perceptual space.
Building on previous marketing research using perceptual maps, in Moe and Schweidel’s model, the distance from the origin to the brand reflects the volume of comments for the brand — for example, in their sample data, McDonald’s received nearly three million social media comments whereas Burger King received fewer than four hundred thousand. This would put McDonald’s much farther “out” from the perceptual map’s origin point (the 0,0 coordinates) than Burger King. The total mentions for Pizza Hut and Dunkin’ Donuts were similar to those for Burger King, which puts those brands at a similar distance from the origin and suggests they are more comparable as benchmarks for each other.
Co-mentions point to competitors
Wendy’s received a slighter higher number of mentions than Burger King, over four hundred thousand, but Burger King received many more co-mentions with Wendy’s (over twenty-six hundred) than it did with Pizza Hut or Dunkin’ Donuts (both fewer than three hundred). This puts Burger King and Wendy’s closest together of these four on the perceptual map and suggests the two are more appropriate performance benchmarks for each other.
The researchers also found that the farther a brand falls either up and/or to the right of the origin point on the perceptual map, the lower the overall “positivity” of the sentiments expressed toward it in the social media comments. For example, the authors note that a brand like McDonald’s attracts a high volume of social media activity along with “a healthy dose of negative comments.” A niche brand like Chick-fil-A or Sonic has a lower volume of comments and the comments are more positive on average than those for McDonald’s. A manager comparing these two metrics might conclude that McDonald’s was not performing as well as Chick-fil-A. This comparison would not be appropriate, however, given the substantial distance (read “brand dissimilarity”) between the two in the latent perceptual space.
Moe explains: “What the model tells us is that even though the social media sentiment around McDonald’s is more negative, it is still in line with, if not outperforming, the expected sentiment given its volume of social media mentions and co-mentions compared to those for other brands.”
For brands less differentiated from each other — Wendy’s and Burger King, for example — the degree of differentiation correlates highly with sales revenues.
Moe and Schweidel’s research revealed one other observation of value to marketing managers. Specifically, they found that for brands with less differentiation from each other — Wendy’s and Burger King, for example—the degree of their differentiation correlates highly with their sales revenues. In short, their results suggest that an increase in differentiation from the other brands in this group will have a positive effect on sales revenue.
The ability to map where their brand stands in relation to others in terms of the frequency of mentions and co-mentions and the average sentiment expressed toward it in those social media comments may prove useful in a number of ways. “For one, this approach can provide marketers with a more nuanced ‘voice of the consumer’ than metrics alone,” Moe says. Employing the latent space approach might also give marketers insight regarding the effect of their actions on the social media activity around their brands. “In short,” Schweidel observes, “our latent space framework, in conjunction with other data collection mechanisms, may provide brands with a more complete perspective of their current and potential customers.”
By Kim Alan Pederson
Integrating Social Media Metrics
Wendy W. Moe and David A. Schweidel (2015) [Report]
2014 Best Paper and Top Download:
Social Media Intelligence: Measuring Brand Sentiment from Online Conversations
David A. Schweidel, Wendy W. Moe, and Chris Boudreaux (2012) [Report]
Social Media Intelligence (2014) [Video]
3 WAYS to GET CONNECTED
Employees of MSI Member Companies enjoy the benefits of complete online access to content, member conferences and networking with the MSI community.
Qualified academics benefit from a relationship with MSI through access to msi.org, conferences and research opportunities.
The public is invited to enjoy partial access to msi.org content, a free e-newsletter, selected reports and more.