Reengineering Business Intelligence Amplify social signals

Overview

Marketing executives already recognize the power of social data to yield insights into customer behavior and expectations.¹  Over the course of just a few years, social data became invaluable to them as customers readily revealed information about themselves and their interests, friends, and purchasing decisions.

But when it comes to strategic decision making, social data is just one piece of the puzzle. At the same time, it has become bigger than marketing for some. Businesses are beginning to see its value in areas such as risk management, product development, reputation management, and supply chain operations. Still, nearly half (44 percent) of senior executives consider social data alone inadequate to attain the strategic insights they need to guide their organization (figure 1).²

Senior executives increasingly recognize the need to account for a regularly changing context in their decision making. By combining social data with other data sets both inside the enterprise (such as financial, enterprise resource planning, and business intelligence systems) and outside (such as traditional media, insights from industry analysts, and human intelligence), they’re able to lend new context to their insights. Plus, using this approach they are able to regularly assess their strategic decisions so critical issues don’t have to wait until the next budget cycle or ad-hoc market research study.

Senior executives are developing next-generation business intelligence capabilities by rebuilding the engines that deliver insight. Along the way, they’re making the transition from decision making that is aligned with traditional budget cycles to near-real-time decision making. These engines blend social media with other external and internal data sets, along with refined analytics, to help anticipate strategic risks and opportunities.

To reach these bold goals, businesses are rebuilding their “insight engines.” They are expanding data sets, deploying advanced analytics tools, and using different types of human expertise to answer increasingly complex questions. As a result, business leaders should be equipped to make swift decisions as situations unfold.

This isn’t a “bolt-on” analysis capability. Companies are embedding signal detection and analysis across their organizations. For example, Burberry, a luxury retailer, has accelerated the identification of emerging trends around its products throughout its operation. The company achieves this by combining enterprise information (e.g., SAP data) with social data (e.g., customer social data feeds and employee communications via Salesforce® Chatter®) to make timely adjustments throughout its supply chain, including changing product design.⁴ Such initiatives, however, are relatively nascent, as early adopters learn what is required to deliver these new capabilities effectively.

What’s driving this trend?

In 2008, Barack Obama’s first presidential campaign team made pioneering use of social networks to raise awareness, generate funds, and encourage voter turnout. They doubled down on this approach in 2012. The day after Obama won the 2012 election, Time magazine trumpeted the role of big data and data analytics in Obama’s historic win. The team was able to reduce many unknowns about voters and their behaviors by combining knowledge about people from the party database, the campaign’s interactions with people, and people’s reactions to campaigns (including their opponent’s campaign) through social media and other media sources. Mash-up analyses of these data sets allowed them to monitor the changing context, consistently identify specific targets for special campaign efforts, and make strategic investments in certain states over others, beyond the so-called “swing states.”⁶

Social data—part of a big(ger) data trend growing intensely 

In a recent survey, 93 percent of C-level executives indicated that their organizations are losing revenue opportunities by not fully taking advantage of the data they collect. Nearly all of the respondents say their organizations need to improve information optimization soon. Nearly half feel that the ability to translate information into actionable insight is the most important area on which to focus.⁵

Many companies that are rebuilding their insight engines have a similar goal. They too want to remove the unknowns from strategic decision making. In the corporate world, strategic decision making has often relied on a combination of experience, intuition, and, more recently, business intelligence derived from analysis of enterprise and market data. Previously, business leaders were often unable to detect the early signals of change when executing their strategy. For example, competitor moves, new investment patterns, and changing stakeholder behaviors that had the potential to increase organizational risk may have taken months to identify. By that time, the business context may have already shifted.

Analyzing both social and traditional media can deliver strong signals about emerging developments in the market. But unless those signals are identified and integrated into the decision making process, their strategic value can be limited.

Lessons learned: What works and what doesn’t

Until recently, the idea of harnessing both external and internal data sets to provide executives with early signals was more fantasy than reality. But that’s changing. There are already several hundred companies involved in developing solutions in this space—driven by available, mature, and sustainable technologies.⁸ The capabilities being developed can be as diverse as the data they analyze.

Want to know what this approach looks like in action? Here are four recent examples of innovators that have shaken up their approach to insight delivery, and are already beginning to reap the benefits.

• UBS: Reputation risk management. Brand reputation management is a big issue for boards and the C-suite, which are beginning to establish capabilities for integrating and analyzing news media and social media with their internal compliance systems, giving them the ability to regularly monitor and assess emerging risks. UBS, a global financial services firm, analyzes thousands of external sources in combination with its internal compliance database to detect controversial information and compliance issues at the organizations with which it does business. UBS uses this information to vet clients, determining that its reputational risk profile falls in line with the bank’s risk exposure.⁹
• GE: Operations. GE developed a tool called Grid IQ Insight that mines social data, including geotag data and attachments such as photos, for mentions of electrical outages. The tool is designed to help electrical utilities identify and validate service outages in a timely manner, determine the resources needed to address them, and accelerate the repair process.¹⁰
• Walmart: New product development and demand prediction. Walmart Labs acquired Kosmix, a $300 million unstructured data analytics company, to build its Social Genome Platform. It uses a wide variety of data, including social media updates, blogs, transactions, images, media check-ins, and location, to help business users more effectively predict product demand and launch products.¹¹ For instance, Walmart’s private-label brand introduced new spicy chips in California and the Southwest based on consumer preference insights from Walmart Labs, which combined social chatter analysis and sales of branded spicy chips carried by Walmart stores to identify a new geography-specific opportunity.¹²

DoD: Strategic investments. The US Department of Defense (DoD) and Central Intelligence Agency (CIA) use temporal analytics (trend analysis over a period of time) and other analytics technologies to pick up “predictive signals” amid the clutter of the World Wide Web and social media data. The DoD is using these capabilities to guide future science and technology investment decisions. Similarly, the CIA uses this approach to track protests around the world to predict threats against the interests of the United States.¹³

On what makes a data scientist successful: “Think of him or her as a hybrid of data hacker, analyst, communicator, and trusted adviser. The combination is extremely powerful—and rare.” —Tom Davenport (author, professor, and senior advisor to Deloitte Analytics) and D. J. Patil (data scientist in residence, Greylock Partners)⁷

For companies that have been experimenting with social data analytics or semantic analytics tools, there are some lessons learned from efforts that have failed to deliver the desired degree of insight:

Analysis without defined goals: Exploring—sometimes in real time—data sets that you never had access to before may offer new insights. However, starting without strategic questions, clear metrics, and hypotheses around insights generally leads to poor utility of those insights.

Frameworks without context: Using off-the-shelf social data analytics tools that track key words, volume, and sentiment allows you to listen in to external conversations. However, these tools do not include your business context and almost inevitably lead to an incomplete analysis.

Right problem, wrong resources: Deploying analytics or market research resources who have experience with mining large volumes of qualitative data, pattern discovery, or anomaly detection capability is only half the solution. Neglecting to connect these resources with people who have a deep understanding of your business and strategic decision making process can position them to fail. They may struggle, for example, to raise insights to the right levels in the organization.

Looking ahead

In many companies, merely capturing and managing enormous volumes of data, much less analyzing them for insights, can seem like a virtually insurmountable task. But technology can help—along with a recognition of the role that humans and organizations have to play in the process. Here are some likely short- and long-term implications of this trend on business.

 Technology will likely continue to evolve at an accelerated pace

Already, academic and corporate scientists are continuing to advance the state of the art of data science in areas such as natural language processing, computational heuristics, and semantic systems. And they’re being spurred along by a more intense public recognition of, and interest in, such technology. Big events like elections only serve to keep such technologies in the public eye. Meanwhile, practical applications of these technologies continue to mount. It all adds up to a self-perpetuating system that could drive analytics capabilities to new heights in coming years.

Talent and organizational strength

Developing these capabilities often requires the skills of three specific leaders: data scientists, change agents, and executive champions. Data-gathering and analytics technologies receive plenty of attention, but capitalizing on the opportunities they introduce ultimately requires people who can examine, understand, and interpret the data, then present it in a way that the organization can use it effectively.

Large businesses today typically employ strategic analysts to examine competitive and market intelligence, conduct financial analysis, and create forecasts. Often, these professionals are focused on transactional analysis or solving huge problems that take years to address.

As they begin to triangulate social and traditional forms of data, businesses should expand and accelerate their analytical and pattern-detection capabilities. This demand is elevating the role of the data scientist, “a hybrid of data hacker, analyst, communicator, and trusted adviser.”¹⁵ Data scientists plunge into the volumes of data to identify patterns, extract insights, and then apply and present their findings in the context of business problems. These professionals often have strong mathematical skills and investigative capabilities, are adept at pattern recognition, and are able to understand and articulate business problems.

Change agent roles are also important. They can be filled by people who understand the business, have analytical capabilities themselves, and are able to create targeted messaging to help improve acceptance and adoption of the data findings across the enterprise generally and within discrete areas specifically.

Finally, executive champions can play an essential role in driving the development of and budgeting for insights capabilities. In a recent survey, strategic corporate leaders—CEOs, presidents, and managing directors—were almost twice as likely as CIOs and CFOs to say that social business (and by inference the data underlying it) is important to their organizations.¹⁶ Providing top executives with the solution to a thorny problem can help build momentum for insight initiatives.

Data tailoring

Bite-sized tweet, meet the 1,000-page enterprise report. As executives begin to factor in social data and other enterprise data sets for fact-driven decision making, some will be more comfortable with a graph of tweets or other summary messages, while others will demand more detailed analysis. The new insight engine should have the ability to do both, allowing users to peel the layers of the onion.

Contexualization

Companies should have the ability, on demand, to mash up social data with enterprise data from ERP systems and other sources for timely contextualization, much like Burberry has done. Framing the resulting insights within broader issues and trends that are relevant to recipients of the information can facilitate understanding and decision making.

Workflow integration

Over time, more users across the value chain will likely need to be able to consume the new insights as part of their normal workflow or risk being left behind. The process of disseminating insights should factor in the ability of the organization to absorb and respond to them.

Speed

The devices and visuals used to access timely contextualized insights will likely become increasingly important. Passive output such as retrospective status reports will likely give way to consistently refreshed and collaborative data vehicles such as mobile applications and alerts.

Some companies are already surging ahead in the race to harness the potential of insight engines fueled by social data and big data. But no clear winners have emerged, and the technologies and processes continue to evolve in a timely manner. An important factor for the achievement of a company’s goals will likely be its ability to transition from decision making based on traditional survey and budget cycles to near-real-time strategic decisions, without sowing organizational anarchy. Rebuilding insight engines, deploying human expertise in new ways, and effectively integrating resulting insights into the existing workflow can help make this happen.

My Take

Don Springer, VP Product Management/Strategy, Oracle Cloud Social Platform 

Until 2012, I was CEO of Collective Intellect (CI), the social media analytics company acquired by Oracle. We helped businesses track, understand, and use social data to evaluate consumer opinion, measure buzz, identify customer sentiment, and manage corporate reputations. Now, as part of Oracle, we’re augmenting large enterprise data with social data to deliver insights on everything from emerging company risks and new product demand to customer purchases.  And we’re doing it in part by blending unstructured data from social media with structured enterprise data to create real-time signal detection capabilities. The result is a whole new set of insights for the C-suite, business unit leadership, and even frontline workers.

Along the way, interesting use cases are emerging. In the financial services sector, we’re creating next-generation global wealth management solutions that combine research about a company with consumer demand signals from social media on that company and its products. This allows research analysts, almost in real time, to make more informed interpretations, which in turn can drive smarter decisions. We’re also helping some large retailers that are looking to intercept customers who are already in the store, but are using their mobile devices for comparison shopping and price checking. We want to influence the decisions of those customers while they’re in the store making up their minds. Ten minutes after they leave the store could be too late.

C-level executives should understand that they have a window of 12–18 months to complete the shift to real-time signal detection. After that, they will likely be well behind their toughest competitors. To do it, they should attack this challenge at the enterprise level, shift business intelligence and insights, and begin making strategic decisions in alignment with their changing context.