by Charles Roe
Part 1 of Business Intelligence 3.0 discussed the introductory analysis of Mr. Krish Krishnan’s “Social Analytics: Measuring the Digital Age” presentation at the Enterprise Data World 2011 Conference. It focused on the importance of the growth of Jeff Howe’s term “crowdsourcing” as a necessary element in the early evolution of Social Analytics (SA) through three primary elements:
- Emergence of user-based communities
- New data collection technologies
- A “Long Tail” retailing strategy
Those three factors have combined to allow business to utilize “the wisdom of the crowds” through the Social Media (SM) revolution and thus gain access to a much greater amount of information and larger audience that traditional marketing and advertising techniques do. SA is similar to traditional Business Analytics in that it uses exiting Business Intelligence applications to collect, process and report on a company’s data. But, SA takes a company’s entire SM online experience from a range of sources like Facebook, LinkedIn, Twitter, blogs, video and photo sites and others, and integrates it with already present, traditional data warehouse assets. This wealth of information allows a business to get a much more comprehensive understanding of market share, reputation, sales and marketing possibilities, as well as grow their brand into areas that did not exist before.
Part 1 then discussed the five major steps necessary to integrate traditional data assets with SA assets in what is called an Enterprise Social Data Evolution. This transformation allows a business to change their traditional mass market model to one where each customer is targeted at a personal level. The steps include:
- Collection of standard enterprise data and unstructured data
- Transformation of that data into useable content though BI applications
- Categorization of content into effective metadata system
- Integration of content into Knowledge Base and other functional applications
- Use of all the data to create a personalized marketing system
Part 2 of Business Intelligence 3.0 discusses the creation of a Social Intelligence Architecture, how a business can implement such a system and gives a concluding snapshot of American Airlines and their successful use of Twitter data to learn about customer complaints during flight interruptions.
Key Features of an Effective Social Intelligence (SI) Architecture:
The integration of an SA system into traditional BI applications is now called Social Intelligence, since no longer can a business attain success though the analysis of traditional enterprise data without integrating social media data into their BI solutions. The Social Intelligence Architecture is an interconnected system of four primary stages:
1. Gather Data: This is “listening stage,” whereby a business creates a plan to collect as much data as possible across the SM landscape. It involves a mining blogs, microblogs, photo sites, video sites, communities, forums and any other forms of social media for any and all information concerning the business, and if possible that of competitors as well so that a more comprehensive assessment can be created. Included in this gathering of data is the collection of web metrics like page views, hover over clicks, widgets/gadgets, shares and any other metrics provided though standard analytics systems.
2. Process Data: The second stage is termed “categorize, context and score.” In this stage of the framework, a given business must decide on a set of processes for the voluminous amounts of SM data collected, break it into categories, add tags, scores and focus on specific personal or professional subject areas that are most relevant to their desired SA goals.
3. Analyze Data: Once the data is gathered and processed into functional units with defined attributes that can be analyzed, a company then begins the real “analytics” process. There are a multitude of different analytics that can be used:
- Context – What is the relevancy and context of all that content to a specific organization? What are the negative or positive impacts of that context during a given measurement cycle?
- Reach – How much of a reach did the content have, both in terms of time and people who it got to? How many times to how many people?
- Channel – What channels did this content enter into? Was it just text, or did it also make it into video or blogs or other places?
- Segmentation – What segments of the market did this content enter into? Was it only a segment the company is currently in? Or did the content reach into new segments?
- Traditional Data Mining: Using standard BI applications this data can be data mined for patterns, apply statistical models to it and summarize it into operational information resources that can be used within a company for predictive forecasting, behavioral /trend analysis and more.
- Content – How good or relevant is the content? What is the effectiveness? How often is it read? How long has it been available?
4. Report Data: Finally, once the first three stages are completed – though a true SA process is continuous – a business can compile reports. The future world of SI reporting is portal driven and will work in similar fashion to reports already used today. But, instead of analyzing internal company metrics, the analysis will be of social media data coming through the entire SI Framework – including sentiment analyses, keyword assessments, post volume metrics, conversation trends and domain/author inquires to name just a few.
The Future is Social: How Businesses are Implementing SA
The successful implementation of SA process requires a focus that begins with a “Blue Ocean Strategy,” whereby a company defines specific SM goals, where they are going and what they want from the social media world, such as increased revenues, establishing leadership in certain online communities, educating constituents, building awareness and recognition among an endless list of possibilities. The next steps include generating demand, participating with customers and using SA to guide the entire process though useable, valid metrics. Apple’s release of the iPod is a prime example of such a successful SM strategy; Apple had specific goals, created and captured demand, engaged directly with their customers and leveraged SA metrics to drive innovation with the iPod. Mp3s were already popular when the iPod was originally released, but Apple implemented a strategy across the social media web that said “iPod is the best” and “if you don’t own an iPod you are missing something;” five tech generations later the iPod is still selling and Apple’s new products like the iPhone and iPad are top sellers in their markets. Apple made the competition irrelevant though pointed application of a social media strategy. They built a vast SM community through their own websites that spread to innumerable other customer-driven sites, so that now Apple sits at the zenith of an immense web of users who constantly interact, review, buy and sell, participate with and link to each other. Through the implementation of a functional SA system, Apple can leverage that web of information and know immediately what is being said, what are pros and cons, how many people are listening/reviewing/communicating about their products and can get the word out to millions of brand ambassadors literally overnight.
The entire SI/SA process leads to a change in traditional data warehousing as well, which requires a shift in the mentality of those in charge of data operations within a particular company. A social media warehouse combines the standard, structured, enterprise data warehouse with volumes of unstructured data collected through the social media experience. The social media is harvested across the Web from the many sources listed above, categorized, processed through textual ETL software and then viewed through the medium of BI/BA reporting features. The reports look similar to standard BI/BA reports, what is different is the content and how that content is represented. Companies now need data stewards, data architects, network specialists and administrators that understand the necessary changes and are qualified, willing and supportive of the entire process.
Social Data Warehouse Concept
Conclusion – A True Snapshot
The people in charge of the SI/SA process must create the Blue Ocean Strategy so that all stakeholders involved in the change know what the expectations are. There must be a path to success laid out, a strategy in place that shows what success means so that all the volumes of unstructured data added into the enterprise data warehouse can be utilized appropriately to aid in that strategy. From April 20th to 30th, 2010, a large US based airlines did a SM study on Twitter concerning flight disruptions during the volcanic eruption in Iceland. They gathered data from about 100 tweets concerning the opinions of customers on airline punctuality, customer service and the effects of the delays on those opinions. They then analyzed them and found out exactly what their customers were saying, who was saying them and the effects that had on their reputation. Granted, this is only a snapshot of 100 tweets and thus does not cover the entire SM world, but for the Airlines it gave them a better idea of what was going on in one small part: overall customers were unhappy about Customer Services and how they dealt with delays, 17% of the tweets concerned delays, delays were reported most during the days of peak traffic and most were attributed to mechanical issues during departure. So, even a small group of 100 tweets gave the company actionable metrics.
Social Analytics is only in its infancy, but over the next three years it will become the driving force in the business world: mass marketing will continue to become more personalized marketing to individual clients, successful social innovation will require changes in traditional BI/BA processes, new data models will need to be built to account for the constant changes in the SM universe, the implementation of Master Data Management and Metadata procedures will need to be perfected, data governance questions of how to govern such massive volumes of unstructured data will need to be further addressed. The entire Data Management industry is undergoing a revolution because of the rise of social media, social intelligence and social analytics – they are Business Intelligence 3.0.