Facebook IPO: Data Rich, Information Poor, Analysis Absent?

Michael Rappa's analysis of the circumstances surrounding the Facebook IPO reveals just how easy it is for a large and seemingly bulletproof enterprise to fail miserably at Business Intelligence, Business Analytics, and Big Data:

Facebook has mountains of data on a scale few companies have ever seen before. Shouldn't it have unparalleled insights into its users and their value to advertisers? Reading the prospectus, I was surprised to see the IPO rested as heavily as it did on the weight of a single trend line ‑ growth in active users over time ‑ and not on data-driven insights into its user community. If Facebook is making good use of its data, it’s not evident to investors.

It seems to be a classic case of selling the sizzle rather than the steak, and the market figured it out pretty quickly, though not quickly enough to prevent a lot of people losing a lot of money. Has the dot-com bubblereally faded from our collective memory that quickly?

It's already hard to remember what we did before social media, but the question about social media, as with the dot-com ventures, has always been how to monetize its great wealth of community and data.

But the Facebook story is now also a Big Data story, because the information Facebook has about each of us is vast, and comes in many formats. It has volume and variety, two of the three defining qualities for Big Data.

The third defining quality for Big Data is velocity: It's not enough to analyze the available data; the analysis has to be as close to immediate as possible. If Big Data offers competitive advantage, it is in the velocity with which insight can be delivered and monetized.

Risk management, management by strategic objective, and public company governance find common cause in Big Data. Jill Dyché says as much in her Harvard Business Review article, "Data: One Antidote to Risky Behavior":

Savvy managers understand that weaving data-driven decisions into the fabric of corporate governance can obviate organizational infighting and drive progress. By establishing clear accountability measures, managers can determine whether and how corporate goals are being achieved, and hold people accountable for how they are achieving those goals. This motivates business people to rely less on hunches and more on hard data.

When we apply what we know, when we know it, to the realization of strategic objectives, we mitigate risk, and increase shareholder value. The attraction of managing by "gut feelings" has always been its immediacy. Providing the hard data to support or counter gut feelings with the same kind of immediacy is the Business Intelligence/Business Analytics/Big Data challenge.

Healthcare Outcomes and Business Analytics

In 2001, the Institute of Medicine (IOM), an arm of the US National Academy of Sciences, released a report detailing the many failings of health care provision in the US, and laying out a plan to fix health care. The plan was to become more proactive and less reactive in engaging patients and families to manage their healthcare, improving the overall health of the population, improving the safety and reliability of the healthcare system, coordinating patient care amongst multiple agencies, delivering palliative services, eliminating abuse, maximizing access, and improving the healthcare system's information infrastructure.

In fact, the focus on healthcare IT at the IOM goes back even further. In 1991, they published "The Computer-Based Patient Record: An Essential Technology for Healthcare"(revised 1997), a report heralding computerized patient records as the best hope for higher quality of care.

In the Fall 2010 issue of the Journal of Healthcare Information Management (a publication of the Healthcare Information and Management Systems Society ‑ membership required), Judy Murphy writes about the progress that has been made in healthcare since the IOM's push for better healthcare IT began over twenty years ago:

Robert Wachter, author of two books on patient safety and editor of the federal government's two leading safety Web sites, gives efforts an overall grade of B-, a slight improvement from his grade of C+ when he performed a similar analysis five years ago. Wachter says that overall, the past decade has seen progress in hospitals' responses to accreditation requirements, regulation and error reporting, but health IT has lagged behind, with research in the area slowly advancing and remaining underfunded.

As Judy Murphy notes, progress has been at best mediocre:

Unfortunately, the attractive claims linking health IT and quality outcomes rest on scant empirical data. Several studies and system reviews published in 2009 and 2010 have demonstrated some evidence for cost and quality benefits of computerization at a few institutions, but with little evidence of broader application.

And it seems that the long-term strategic objectives of this initiative have been obscured by the shorter-term tactical objectives:

The modest quality advantages associated with computerization are difficult to interpret, and are clouded by the fact that the quality indicators used today often reflect care process metrics rather than patient care outcomes. In other words, we are measuring how many patients receive smoking cessation counseling or prescriptions for beta blockers; we are not measuring how many patients quit smoking or what their reinfarction rates are.

The bright spot in all of this is the use of clinical decision support tools:

...it also seems clear that implementing and adopting health IT is not enough. The evidence points out that, unless you specifically use systems with clinical decision support tools and paired with practice changes, you are unlikely to improve quality and patient safety and unlikely to achieve overall reductions in health costs.

Before computerization of healthcare records, we said that healthcare was data-rich, but information-poor. Post computerization, it seems healthcare IT is information-rich, but analysis-poor. In other words, we have the information we need to make a difference, but haven't yet applied the appropriate analytics tools and mindset to the larger strategic objectives.

Clearly, budget is a large part of the problem, but in the age of doing-more-with-less, asking for a larger budget is probably a non-starter. So business analytics managers in healthcare need to look at ways to liberate resources from repetitive administrative tasks so they can spend more time adding value to outcomes via better decision support capabilities. You can't focus effectively on the larger issues if you spend all your time resolving the smaller ones.

Big Data Analytics Skills Shortage

A recent article at SearchBusinessAnalytics cited "raw and user-unfriendly technology" and lack of "skilled experts" in these technologies as the biggest challenges for large enterprises seeking the considerable benefits of big data analytics. As enterprises become more information driven, and business intelligence and analytics become critical to competitive advantage both strategically and operationally, there just aren't enough skilled personnel to handle the development of the needed predictive modeling and predictive analytics applications.

The greatest need is for more data scientists -- people who have post-graduate educations in statistical analysis. As demand grows, and supply remains relatively constant, individuals with these skills will command larger and larger portions of corporate business analytics budgets.

While this "crisis" seems worrisome, it is a problem that the average business intelligence platform manager would love to have. They would all like to be pushing the data analytics envelope to provide proactive and progressive solutions that meet or exceed their enterprises' information needs. Instead, these managers are dealing with smaller curative or preventive issues that won't go away, and which occupy altogether too much of their budget, and their resources' time.

The oldest (okay, maybe second oldest) service-type business proposition in the world goes something like this: If I can make (or save) you $50, will you pay me $5? In IT, such a proposition speaks to both return on investment (ROI) and total cost of ownership (TCO): you make back the money you invest in the service, and you save on the cost of operating the system.

That's the promise of APOS well managed BI solutions. You conserve the time of your resources, and you can move away from fixing and preventing problems, and toward providing progressive business intelligence solutions for your information consumers.

Tweet