What is Big Data and How Do We Use It?
Why, with all the data available, are pharma companies so bad at getting true marketing insights to drive their business? Bill Drummy asks how marketers can realize the promise of “Big Data”.
My inquiry began with a simple question: Why are pharma companies so bad at getting true marketing insights to drive their business? At an advisory board meeting for Pharmaceutical Executive magazine, I spoke with Brad Sitler, principal industry consultant at the analytics company SAS. I asked him why, with all the data available in the pharma industry, companies are basing entire multi-million (sometimes hundreds of million) dollar campaigns on flimsy market evidence: focus groups, 1-on-1 interviews, primitive segmentation. Data encoded with potential insights are everywhere — sales data, ad serving data, lab data, social media data — why don’t they use it?
To cite the most top-of-mind example, with all the hubbub about social media, pharma has been slow to exploit the true value of Facebook. This is not ‘likes’ on a pharma brand page; the true value is in the “social graph,” which can be mined for insights about patients, disease states, and behavior patterns.
Brad cited two main reasons: “Pharma has outsourced the analytical talent they need to make sense out of the data; it’s no longer a core competency. The second reason is that, until recently, pharma has never had the economic pressure to fundamentally change their business model. There was no burning bridge.”
Well, in 2012, the bridge is in full flame. The good news is that the deeper I dug into the issue of Big Data, the more I found innovators who are boldly applying imagination, technology, and analytical sophistication to reshape industry practices, in marketing and beyond.
Reshaping industry practices. That’s a big statement. But it’s no exaggeration. Because once you can see beyond the confusion, the potential of Big Data reveals itself to be startling in its power.
What is Big Data?
First, let me deconstruct this latest fashionable tech buzzword. What exactly is Big Data and what does it mean in the healthcare industry? Put most simply, Big Data refers to the tsunami of information that began submerging us with the digitization of knowledge five decades ago. Fueled first by the computer revolution, it has been amplified more recently by the accelerants of mass networking, cloud computing, and the government-mandated move to electronic health records. The consequence: Every patient experience now generates rivers of data which, if pooled intelligently, can trace a detailed portrait of a patient’s health and, when aggregated with other patient data streams, can coalesce into deep reservoirs of knowledge about entire disease states and patient populations.
To comprehend the potential of Big Data in healthcare, imagine it as broadly as you can, and then raise your imagination by several orders of magnitude. In R&D, its promise for finding more effective, targeted, “personalized” therapies is enormous. In the supply chain, the potential for massive increases in market knowledge and efficiency is gigantic. In marketing, the ability to uncover actionable insights about patients, HCPs, or virtually any other audience you want to influence is ginormous. Yes, whatever term you choose to use, Big Data is indeed Big.
And yet, like so many things in what I have been calling the “Speed of Change” era, the ability to harness these potentialities does not simply require deployment of a new set of tools and techniques; nor is it sufficient to just scale up your intellectual framework to accommodate the tiers and tiers of terabytes. More fundamentally, success requires a new way of thinking, a different approach to solving fundamental problems. In fact, tying yourself even to fairly recent approaches to technology virtually guarantees that you will miss the Big Data boat. The companies that are getting it so startlingly right are, almost without exception, led by visionary entrepreneurs who are not tied to older ways of doing things and therefore willing to follow the implications of technology without fealty to older modalities.
In our own work here at Heartbeat Ideas, for nearly ten years now, we have always begun our strategic analysis of a particular marketing problem (whether developing a campaign for patients, HCPs, or payors) by deploying our social media and competitive listening tools to uncover insights about target audiences and the facts about competitive performance in these sectors. I had always assumed that these more advanced marketing techniques, which enable the development of strategies grounded on objective, data-driven market insights, would become commonplace. Yet, for the most part, this does not appear to be the case. Most strategies seem to still be based on ‘old school’, highly subjective analysis built on the thinnest foundations of objective data.
But when I looked around healthcare more broadly — beyond the backyard we play in every day with our pharma marketing clients and partner agencies — I found some truly astonishing work.
I refer you to just two of dozens of innovators, each of which is working to solve messy problems of data glut that portend significant market-shaping consequences: Medivo is taking a type of healthcare data that is ubiquitous but largely ignored (laboratory records) and mining it for patient and disease insights, and improved marketing; and Integrichain is performing astonishing acts of real-time analysis of pharma sales data to dramatically re-balance the entire pharma supply chain.
Both of these companies show great value and investment potential. Moreover, they demonstrate to the entire pharma industry how the problem of the data flood can be tamed and turned into empowering knowledge. Finally, they provide proof points for pharma companies looking to truly harness the dynamics of digital disruption before their business models are entirely submerged.
Medivo: Diagnosing a data chasm
Like a lot of great ideas, this one is shockingly obvious — once someone else has figured it out. Sundeep Bhan had been a founder of Medsite, but after selling the company to WebMD in the mid-00s, he grew restless. So, Sundeep looked around the industry for a new opportunity. He eventually settled on uncovering value from a kind of data that nearly every patient generates, that ubiquitously populates the healthcare system, but that was being dramatically under-exploited. The humble, homely laboratory record.
So in 2010, Bhan co-founded and became CEO of Medivo, focused on bringing that data together in a simple, easy-to-use system that would help physicians see patterns in a patient’s disease, as well as the patient’s response to treatment. Their system also makes it easy to flag patients whose results fall outside of desirable ranges, in comparison to benchmark population health reports. In addition to giving doctors a better view of their patient populations, Medivo also offers doctors the service of directly contacting their patients to encourage compliance and return office visits.
In so doing, not only does Medivo provide a valuable compliance service to doctors and patients, it also enables pharma companies to become smarter marketers. By reading the tea leaves in the lab data, marketers can identify physicians who may be the most appropriate dispensers of their therapies, but who may have been missed using traditional market-targeting techniques.
For example, the de-personalized data may uncover a group of hep-C patients who are not responding to treatment, making their doctors ideal targets for a manufacturer with the right hep-C medication. Or in diabetes, where typically 75% of patients are not well-controlled, the lab data can help to focus doctors on problem patients and enable them to manage the worst cases more closely.
The value of this insight to pharma companies is equally clear: “Lab results help sales reps identify doctors they didn’t call on before,” says Bhan. “And these are doctors who clearly have real needs for their products.”
Since starting in 2008, Medivo has consolidated data from some 4,000 patient service centers, covering 176,000 physicians and more than 50 million patients. The sheer scale of data they are accumulating on a daily basis is extraordinary. “Every time we add a doctor, we add one gigabyte of data per year.” So, with 176,000 doctors and five years of data for each doctor, Medivo is currently managing a data pool of 8.8 terabytes.
So what is Bhan’s take on pharma’s use of Big Data? “They are ahead on R&D side, but behind on the marketing side,” he believes. “Diagnostics need to be more integrated in their marketing.”
Using Medivo (and other companies who have begun to harness the flow of Big Data in marketing), a pharma company can get a much clearer picture of their best targets and participate in programs to target these prescribers more efficiently.
“We can send hyper-targeted messages like Google does,” said Bhan, “We can start to do that with clinical data.”
Integrichain: “This is a tectonic shift in technology”
Four years ago, Kevin Leninger had an epiphany. “I would be talking to senior financial people in large pharma companies and I would ask them where their products were in their supply chains, and nobody could tell me. And these were multi-billion dollar brands, so knowing whether they had actually been sold, were still in a warehouse, or were being returned to their loading docks was not trivial.” So Leninger, with a physics degree from Iowa State and a deep appreciation for the importance of keeping track of billions of things, dollars especially, co-founded Integrichain, with the initial focus of bringing supply in line with demand. The ins and outs of pharma supply chain management are arcane, to say the least, involving area code naming conventions like “852” inventory data and “867” sales data. Suffice it to say, the system was a mess, and the fallout from that mess included massive cloudiness about what was happening in the pharma sales channel, and chasmic lost opportunities for understanding how marketing actually can drive sales.
In approaching the problem, Integrichain realized that even ‘modern’ relational database systems built with Oracle would be inadequate to handle the problem. “The prevailing way of dealing with data, the so called ‘data warehouse’ approach, was just not going to cut it,” says Leninger. “You know what Google calls data warehouses? The place where good data go to die.” The data are indeed stored, but hardly ever leave the warehouse.
This is what I meant when I said that Big Data challenges require a new type of thinking. Integrichain set about building a cloud-based system that would allow massive amounts of pharma inventory and sales data from virtually every manufacturer and ‘class of trade’ in the supply chain – pharmacy, specialty pharmacy, mail order pharmacy, and hospital — to be held in memory, so that calculations can be performed in almost real time — without the need to retrieve bits from the warehouse. Anyone with even a slight understanding of how a desktop computer works should be staggered by the audacity of that approach. “We now can give our customers a report that shows where every single product is in the supply chain — in three hours. You just try to do that — execute 50 billion queries against an Oracle database. It would melt down.”
The implications of finally having 100% channel visibility are wide-ranging. Companies can identify distributors who may be holding on to product too long. They can better target high-value customers and channels (especially pharmacies), perform much more accurate sales forecasting, and finally begin catching up with other industries in seeing the relationship between marketing activities and actual sales. With Integrichain’s system, pharma companies can tell reps which customers have the highest unrealized sales potential, and so better focus their selling efforts. In pure marketing terms, it allows pharma marketers to finally behave more like ‘real’ marketers.
“When I talk about how pharma does marketing now, the P&G and Unilever people just laugh their asses off,” Leninger says. “It’s all about test and control right? Find out which promotion, which coupon, which co-pay card is actually working. We don't have to guess anymore. We can give them the answer in three hours.”
It’s hard to over-emphasize just how ‘revolutionary’ the Integrichain system is in the pharma industry. Both because it has centralized previously disconnected data, and because of the way in which it is doing it. Their system takes full advantage of technology “on the edge” to solve absolutely central problems. “This represents a tectonic shift in technology,” says Leninger. “We’re running the largest data cloud in the industry, and we are able to run the entire transactional history of an entire pharma company’s sales — in memory.”
Integrichain and Medivo are just two examples. Many other companies (including my own) are applying bold thinking to other aspects of the Big Data challenge: using behavioral media data to identify appropriate customers, mining social media for true insights, tying together Health Information Exchanges, pooling the many rivulets of customer information inside pharma companies. The opportunities for high value innovation are legion.
Beyond Big Data
Yet the entire question of making the most of Big Data is a small part of a much bigger challenge for the industry. Pharma companies currently have not taken full advantage of Big Data, either because of the talent outsourcing problem Brad Sitler pointed out, or because they are not structured to exhibit the necessary boldness to truly follow the technology or customer needs to their logical conclusions.
To address that, pharma needs to cultivate the skills and — more critically — the courage to change its management and compensation systems so that risk-taking is rewarded, “playing it safe” is penalized, and latent talent is unleashed. Then smart people can finally think and act more like the people working in organizations that represent the new value-creators; those able to mine the Big Data landscape for unique customer insights, market efficiency and true competitive differentiation for their brands. Big companies can do this too (look at Google and Apple), but they won’t do it if their people aren’t incentivized to take risks and allowed to make mistakes.
The transformational power of such an embrace of risk-taking would apply far beyond Big Data. Indeed, it would bring change and dynamism into every corner of the commercial organization, finally giving marketers the permission to be bold -- even within the well-known constraints of a regulated industry -- turning risk-averse commercial organizations into risk-taking companies. It sounds radical, I know. But isn’t it just the kind of change needed to race ahead of a burning bridge?
About the Author
Bill Drummy is CEO & Founder, Heartbeat Ideas and Heartbeat West.
How Digital Medicine Can Pinpoint Dosing Regimens to Optimize Drug Efficacy and Safety