An Investigative Approach to Social Media Content Analytics

Two hundred million active Twitter users send 400 million tweets per day. Six-hundred-thousand pieces of content are shared on Facebook each minute. Seventy-two hours of video are uploaded to YouTube every minute.

The ever-growing streams of social media data hold the secrets you need to reach and engage your customers. Yet, according to a Harvard Business Review study, only 12 percent of companies using social media content believe they are using it effectively.

To date, social media analytics focused on the content of posts — the actual text of a tweet, for example — to measure consumer opinion. While sentiment analysis is important, companies also need to dig deeper into the data to glean actionable insight.

Whether companies are trying to engage better with customers, improve brand awareness, or boost lead generation, intelligent social media analysis is critical. And to get there, companies are increasingly embracing a new approach: investigative analytics.

Evolved analytics: Asking iterative, open-ended questions

Data analysis is no longer just the domain of the CIO and the IT department. Data-driven decision-making is the realm of marketing, sales, and business development — and anyone else who has a stake in the organization’s and customers’ success.

However, traditional analytic tools don’t let users interrogate fast-moving, highly diverse types of high-volume social data. As data connections and dependencies grow exponentially, it’s no longer possible to capture actionable information in a rigid set of KPIs and canned reports. To manage content, brand and customer engagement in a social world, companies should consider performing richer, real-time data analysis — with (it’s true) far fewer resources.

Enter investigative analytics, where users ask a series of quickly changing, iterative questions to figure out why something did or did not happen, and determine how to optimize a particular outcome. Compared to traditional analytics, which lack flexibility because they are tied to rigid KPIs and reports, investigative analytics yield insight into questions that haven’t even been dreamed of yet.

For example, traditional analytics may help companies answer a question like, “How many leather handbags did we sell via Pinterest last week?” Investigative analytics ask more meaningful questions, combining social media with other data points (e.g., campaign information, click-throughs, conversions, etc.), to enable more flexible data interrogation.

An investigative analytics series of questions may follow a path like this:

1. “Where are we getting the most sales conversions on leather handbags?” The answer: 25 specific Pinterest boards provide the most leads.

2. “Is there a particular product-marketing image those 25 Pinterest boards are using?” (e.g., Is it a static image with lots of white space, or a woman modeling it at a chichi restaurant? Is it black or vibrant red?)

3. “Are we selling to a younger demographic when a specific color bag is pictured?“

4. “Is there a common conversion thread for consumers who paused a video ad but then resumed play? Did they convert better or worse than those who never paused the video?“

5. “Do more targeted consumers convert from a mobile device?” If yes, figure out how to optimize the campaign for mobile users.

This open-ended process of investigative analytics allows anybody looking at the data to constantly evolve the questions they ask and query data in real time, regardless of data volume. As a result, companies can figure out how to develop customized campaigns that deploy the right messages via the most receptive channels. Not only will they be able to increase efficiency of campaigns and how they target consumers with content, but they also will uncover up-sell targets and inform product development road maps as consumer preferences are revealed.

Navigating investigative analytics

Following are key considerations to put investigative analytics into action:

1. Embrace “frictionless inquiry:” Investigative analytics are all about frictionless inquiry, where the path between the question and answer is void of rigid structure. Frictionless inquiry depends on ad hoc query capabilities and simple analytic tool administration. So, when you reach the aha moment, you’ll have all the information you need to ask the next question or dig deeper into data without having to call IT or the help desk to create a new query.

2. Define, but don’t limit your universe of data points: The lines between web, social media, and advertising analytics are blurring as these modalities become more and more interdependent. To uncover and respond to consumer insights, it’s critical to analyze when and where a consumer said something, what else he was talking about, and what he did as a result. This correlation between engagement and conversion requires attention to a slew of data: demographic background; geospatial and time tracking; which social network and device was used; who opened a video, who paused it, who resumed watching after the pause. While it’s important to determine the data points you can mine, it’s equally important to always add and refine.

3. Ask why, not what: Gone are the days of simply knowing what happened in a campaign. Today, you can — and need to — know why: patterns of behavior or insights to capitalize (either in the moment or in the future). With investigative analytics there are no wrong questions. Instead, by letting the results of your ad hoc queries drive the next question, you’ll be asking the right ones.

4. Take advantage of trending topics: To benefit from trending topics, it’s essential to extract intelligence — whether it’s A/B split testing or consumer likes and dislikes — from social networks as soon as it’s generated in real time. Can we increase engagement if we push out another tweet two hours after it posts? Five hours after? Five days after? Insights gleaned through investigative analytics help you take immediate action to optimize social engagement and drive further traffic and revenue.

5. Don’t ignore historical data: Take advantage of real-time data, but don’t forget the past. For example, it’s important to know if, historically, you’ve been successful with pushing a contest out on Facebook on Wednesdays followed by successive tweets on Friday so that you can replicate your success. Year-over-year trends are also important when planning for annual events, such as Black Friday shopping.

By harvesting the massive amounts of social media data with the power of investigative analytics, businesses will be able to determine exactly what people want, when they want it, and through what social network — which results in true competitive advantage.

This article originally appeared in the February 2014 issue of “Chief Content Officer.” Sign up to receive your free subscription to our quarterly magazine.