Making Data Relevant: The New Metrics for Social Marketing

by Prashant Suryakumar

This article first appeared on MASHABLE.

Social media has come of age. Marketers now have the ability to augment their traditional marketing approaches with rich behavioral and activity-based targeting that should increase marketing ROI significantly.

However, businesses are facing an uncomfortable truth: There are no “best practices” for measuring a successful social media campaign. Crowd behavior is dynamic and context-specific, and it is difficult, if not impossible, to build a “one size fits all” solution.

A structured approach to capturing, measuring, analyzing and refining marketing strategies in near real time is essential to executing a successful social campaign. Initially, however, companies need to invest in infrastructure to make such a learning cycle possible.

Invest in Data

Measuring the impact of social media campaigns is systemically different from that of traditional marketing campaigns. Since the medium touches all the aspects of the customer purchase cycle, a holistic measurement of awareness, transactions and brand impact is essential.

Additionally, social media is a two-way communication medium and businesses need to invest in listening capabilities that capture the activities of their existing or potential customers online. Several paid and “freemium” tools that monitor online chatter can be found online.

While data is abundant, it is by nature unstructured. Integrating listening data with internal web behavior metrics captured by JavaScript tags, customer care logs, brand surveys and transactional data can enable a business to get a 360 degree view of the activities of customers across all of the purchase touchpoints.

Real-Time Monitoring

A typical online conversation has a life span of about one to two days. As a result, it is imperative for companies to respond to conversations in nearly real time. During this short window, they not only need to understand the context and content of the conversation, but also create an effective response mechanism. All of this underscores the need for real-time monitoring and analysis.

Companies like Dell and Best Buy are adopting different strategies for listening to InternetInternet chatter. These investments help keep a finger on the pulse of every conversation active on the networks.

Sentiment Analysis

Text mining and sentiment analysis are the flavor of the season for social media analytics and a common complaint is that the current tools are not able to classify a high percentage of the comments about your brand.

Step back and think about a conversation you had in the last 30 minutes. How many statements in that conversation were unambiguously positive or negative. Not many, right? Getting a 20% sentiment mapping for individual comments is a very high number.

On the other hand, think about the same conversation; Was the overall sentiment of the conversation positive or negative? That is far easier to cognitively classify. If businesses shift their focus to a conversation-based, rather than a comment-based sentiment analysis, they will be able to get a far better read on the aggregate sentiment of online chatter.

New Metrics

The need for improvisation and identification of new metrics is high. Currently, three categories of metrics need to be developed to enhance our understanding of social activities.

  • Metrics that help understand conversations and engagement (e.g. aggregate sentiment, conversation heatmaps),
  • Metrics to spot influencers in a community (e.g. influencer score, Klout score), and
  • Metrics that help in measuring holistic impact of social media activities on the business.

The Interplay Between Buzz, Branding and Sales

Measuring the impact of increased chatter for your brand might not always translate to more revenue for the business. Measuring cause and effect between buzz, branding and sales might show different dynamics for different product groups. For example, the Old Spice social media campaign saw an 800% increase inFacebookFacebook interaction and a 107% increase in sales. The numbers are related, but not necessarily 1:1.

Testing Mechanisms

Social media is a fertile testing ground, and businesses need to appreciate the importance of a robust testing protocol for social media-based actions. Having a mechanism to measure the effectiveness of comments will ensure that businesses can learn quickly and adapt to the social dynamics.

A key point to remember is that the instance and context of the test is as important as the test itself due to the temporal nature of conversations.

Some of the tests that can be conducted are:

  • Who are the right “influencers” to target for a particular product or service?
  • What is the right time to message these influencers?
  • What is the impact of competition activity on our buzz?
  • What is the impact of traditional marketing on social media and vice versa?
  • What are the type of comments that work for selling a product?
  • What are the type of comments that work for selling a service?
  • What are the right pricing strategies?
  • How should the business tap into current affairs?

Behavioral Segmentation

Behavioral targeting dramatically changed with online advertising, and now social media can take this effectiveness to new heights. Activity-based segmentation is far different from traditional demographic segmentation, and this is typically driven by a difference between the purchasers and the consumers of a product. Businesses can draw parallels from traditional marketing (targeting kids so that they can influence their parents) and build a unique social targeting mechanism.

Crowd Behavior

Businesses have tried to artificially stimulate a conversation by mettling in their own communities or creating artificial hype. This approach usually fails miserably. They need to understand that social networks emulate real-world interactions, and excessive policing of user generated content can be detrimental to the natural growth patterns of a network.

Math, business technology and behavioral sciences are the key ingredients for good decision making. Understanding organizational dynamics, flock behavior and complex adaptive systems are all directly applicable to social media. Integrating analytics with a deep understanding of how humans interact in a sociographic and psychographic sense can help a business stimulate a conversation within a community, or trigger flock behavior amongst customers.

Integration Into Existing Business Models

Once companies understand the impact of lead indicators, like buzz, on transactional metrics, like revenue, they can include such metrics into their forecasting models and predict short-term revenue with greater accuracy. Additionally, since a good social media campaign will improve the brand health, the long-term impact of these campaigns can be assessed.

While every business wants to understand the impact of its social media spend, it might not be so easy to integrate that into a media mix model. A good social media campaign might manifest itself in increased brand scores or customer loyalty and will impact the lifetime value of the customers more than the immediate transactional metrics. Including indirect metrics like buzz or sentiment might be one way to capture social behavior.

Product Design

Social media can be a direct line of communication with the end user of your products. Businesses can leverage this very effectively in product design by soliciting input from the end user on what features they prefer in the product. Getting feature specific intelligence from the customer can help in building a product that caters to most of the population and also helps in building a sense of loyalty among the user base. Good examples of this include IdeastormVitamin Water and Fiat.


The framework above is the first step in helping companies understand the who, what, when and where of social targeting. The obvious next step is to integrate all this knowledge into traditional marketing and CRM.

Prashant Suryakumar is a Social Media Engagement Manager at Mu Sigma and is currently focused on social media analytics. This post was co-authored by Dhiraj Rajaram, the founder and CEO of Mu Sigma.