Understanding How Profits are Shifting Across the Computing Value Chain

This post was originally featured on TechZulu.

There’s a famous scene in Pirates of Silicon Valley (a movie that chronicles the origins of Microsoft and Apple) in which Bill Gates is meeting with IBM executives. The IBM executives agree to license software from Microsoft because “there’s no money in software anyways.”

The IBM executives weren’t stupid. They failed to recognize an inflection point in technological history in which profits would shift from hardware to software. They didn’t value the application layer because up until then, there weren’t any profits in the app layer. All profits had resided in the hardware and configuration layers.

Today, Google, and cloud computing companies more broadly, are doing to Microsoft what Microsoft did to IBM. Profits are moving up the value chain from on-device software to software-as-a-service in the cloud. This is manifesting as OS and software licensing revenues are shrinking as cloud-hosted SaaS apps are growing.

In the past few years, Microsoft has come to realize this, as reflected in Steve Ballmer’s repeated assertions that Microsoft is a “devices and services” company. Microsoft has acknowledged that their traditional profit layer of the value chain is evaporating. On-device software used to be an end – a valuable service that customers would pay for – but on-device software is evolving to be a means to a new end: cloud-host SaaS apps.

Although the profitability of the on-device software layer is eroding, on-device software is more important to profits than ever before. Software is more tightly coupled to the layers below and above it – hardware, and cloud services. Effectively, on-device software is becoming a fixed cost of the associated hardware and cloud service solutions.

Why have profits migrated away from on-device software and toward both the cloud and hardware?

Profits are moving up the value-chain because the benefits of delivering cloud apps outweigh the costs. The most significant costs of delivering apps through the cloud are performance and bandwidth requirements. Because of Moore’s Law, these costs are effectively trivial for most cloud apps. The benefits of moving apps to the cloud are enormous: always up to date, maintaining a single instance, opportunities for A/B testing, unlimited computing potential, and many more.

But why are profits simultaneously moving down the value chain to companies such as Apple, Samsung, Jawbone, FitBit, Pebble, Inevitables, littleBits, Looxcie, Parashoot, and robotics companies? During Microsoft’s peak, hardware standardized around a few form factors (ATX, mid-ATX, micro-ATX, laptops, etc). But in the modern mobile computing era, virtually no two computers share an identical form factor. As such, hardware is far less of a commodity today than it used to be. Perhaps more importantly, mobile devices present new opportunities to experiment with novel hardware functions that are closely bound to associated software functions. Microsoft’s traditional OS-licensing model simply doesn’t support a model in which hundreds of companies are creating novel hardware form factors.

Will profits shift again? If so, why?

Although I feel naive saying this, I don’t think profits will materially shift within the foreseeable future. Economics dictate that cloud computing is more power and cost efficient than local computing given sufficient connectivity. As systems-on-a-chip continue to shrink, we’re going to continue to see computers appear in all kinds of new places, particularly in the realm of wearable technology and smaller form factors. Local software is a stepping stone to the power of the cloud. In essence, local software is losing significance in the profit chain as it amplifies the layers of the value chain above and below it. Local software no longer dictates the price, nor plays a role in the device’s total value, but rather, is powering new hardware and connecting that hardware to the cloud.

The new hardware-to-cloud era is here.

Looking Through the Crystal Ball: FutureMed

This post was originally featured on HIStalk. 

Last week I spent four days at the exclusive FutureMed conference in San Diego.

FutureMed is, as the name implies, a conference focused on understanding the future of health and care delivery. There were speakers from virtually every sub-vertical in healthcare, and even traditionally non-healthcare verticals such as 3D printing and robotics. I was particularly pleased by the fact that very few of the speakers at FutureMed spoke about the hottest terms in health IT today: medication adherence, population health management, and patient engagement.

My single most important takeaway from the conference was delivered during Daniel Kraft’s opening keynote: the human brain is hardwired to think local and linear, and that we live in an exponential and global world. Evolution groomed us to learn to survive hour by hour, not to think critically about abstract systems that we cannot possibly “see.” This is a profound notion, and indeed explains organizational bureaucracy and resistance to change. People look at their immediate surroundings, see nothing wrong, and continue as they always have. Daily observations lack a larger sense of context.

Although there was enormous amount of content on stage from a variety of disciplines, one of the overriding themes of the conference was convergence of new technologies stemming from exponential and global thinking. Previously disparate technologies are converging that are creating completely new opportunities to deliver better care at a lower cost to more people than ever before:

Flying robots that occupy no more than half a cubic foot are delivering medications to remote and underserved regions. Although the obvious use cases are in areas that lack roads and other basic infrastructure, these flying robots could even be used in sprawling cities. As the payloads and batteries improve, I’m curious to see how McKesson will respond. Although flying cars probably won’t happen, I can totally see a future with thousands of mini flying robots delivering small sensitive payloads.

Companies are delivering telemedicine solutions through every form factor conceivable. Many companies are betting on the fact that telemedicine will dramatically improve access and thus quality of care all over the world. Although there were about a dozen BEAM robots rolling around the conference, iRobot is developing an autonomous telemedicine robot for ICUs, others are putting iPads on flying drones, and my startup Pristine is putting Google Glass on surgeons and CRNAs in the OR. It makes sense that so many companies are going after telemedicine: telemedicine breaks perhaps the most basic assumption of care delivery, that one needs to be in place X to deliver value in place X.

Patients are increasingly self-diagnosing and self-treating with at-home technologies: Cellscope, AliveCore, 23andMe, Bio-Meme, Scanadu, the Muse, and others. Telemedicine will reinforce and accelerate this trend.

Nurses are 3D printing stents, bone replacements, and other sensitive medical equipment at the point of care that are customized to the perfect size, shape, and fit for the patient. This makes enormous sense for orthopedic replacements, where the replacement should be the exact size and shape of the body part being replaced. This will also dramatically reduce inventory costs, theft, and the need to track assets.

Pharmaceutical companies are looking for ways to invert the drug development and clinical trials processes. They are beginning to rapidly prototype new drugs using algorithms and simulations, and companies such as PatientsLikeMe and ePatientFinder are trying to invert how patients are recruited into clinical trials.

Each speaker at FutureMed was offering his or her own opinion about the future of healthcare delivery. It was amazing to hear from each of them to learn their perspectives. But what’s also clear based on my surface-level analysis above is that no one can predict what’s going to happen. There are simply too many pieces moving in too many segments of a highly fragmented value chain.

The future is going to be ridiculously exciting. I can’t wait to watch it unfold.

The Proliferation of Healthcare Marketplaces

This post was originally featured on HIStalk. 

Some of the fundamental tenets of capitalism are that that markets are free, open, and transparent. Healthcare is notoriously opaque, with little to no price competition within a given region. As such, many have pushed for marketplaces in healthcare.

The marketplace proponents argue that marketplaces will increase transparency and competition, and in doing so, create value and usher in a new wave of innovation. The Obama administration recognized this some time ago, and gave up a lot of political capital to push for federal and state health insurance exchanges as part of the Affordable Care Act.

We are in formative era of marketplaces in healthcare. Intrinsically, marketplaces only work if they can connect parties that were once disparate. With approximately 70 percent (that number can easily be +- 15 percent depending on who you ask) of providers using EHRs and the proliferation of smartphones and the Internet, the technological infrastructure has been laid to support dozens of marketplaces throughout the multi-faceted, multi-partied, fragmented healthcare system.

The following list illustrates just some of the opportunities for marketplaces within the healthcare ecosystem:

Allerad – radiology readings.

2nd.md – second opinions.

Teladoc and Ringadoc (and other similar behavioral / mental health marketplaces) – consult with a healthcare professional from home.

Healthtap – healthcare Q&A.

New Choice Health – procedures.

ePatientFinder – provider education, patient education, and clinical trials.

Informedika – providers and labs. I believe Informedika is looking to connect providers to diagnostic imaging centers and other free-standing, non-provider medical facilities.

It’s interesting to note that there are two distinct types of marketplaces: those such as Allerad and Teladoc/Ringadoc that provide a platform to deliver care services, and the rest that simply try to connect existing players in the healthcare system.

This post of course begs the question: what other marketplaces are to be had? Here’s a quick brainstorm. Some of these aren’t intended to be real businesses, but simply outline the number of opportunities to connect disparate healthcare parties along different dimensions.

• Concierge medicine services.
• Games and books (i.e. a lending library) for patients stuck in hospitals.
• ERs. Perhaps the marketplace could aggregate and show wait times, congestion and other KPIs for the nearest 2-5 ERs. Within an integrated health system, this could be particularly significant for resource allocation and disaster recovery.
• Find nursing home for parents.
• Find at-home care givers for parents.
• Rehab and therapy services – smoking, alcohol, physical therapy, cancer treatment.
• Plastic surgery, including of course lots of pictures.
• Fertility services.
• Cosmetic and optional medical services.
• Renting at-home sleep diagnostics.
• Clearinghouses. Although this sounds redundant, it’s actually not providers because have very few resources to understand and select the best clearinghouse to contract with.

It’s interesting to consider that many of these marketplaces work on the premise of helping providers and medical facilities fill up previously unused time (ZocDoc is the most prominent example). Many of these marketplaces are predicated on a fee-for-service model. How will they hold up in an ACO world?

The Irony of the Modern Conference

Over the past few months, I’ve attended over a dozen conferences. I’ve noticed a trend: the majority, and in some cases the vast majority, of people don’t pay attention during lectures. This is particularly ironic because the most cited reason people attend conferences is to listen to subject matter experts share.

Economically, this is highly irrational. People are spending in excess of a thousand dollars (including airfare, hotel, and transportation) for just a few hours of education. When they arrive to be educated, they don’t pay attention. Why does this happen? There are dozens of reasons, but I’ve identified what I perceive to be the top few:

1) Studies have shown that people value purchases more immediately after paying for them, and that purchases tend to lose value over time. Effectively, time disintermediates price and value. Intuitively, this makes sense. I believe this phenomenon explains conference behavior. People buy tickets and make travel accommodations well in advance of the conference. By the time they attend, they no longer feel the financial cost of attendance. As a result, they don’t feel guilty or have a compelling reason to pay attention. All of a sudden, the email that arrived 5 minutes ago is more important than the overpriced conference.

2) Structure. The fundamental structure of conferences hasn’t changed in over 20 years, if not longer. Although we’ve added PowerPoint presentations, videos, and made conference schedules available on smartphones, the fundamental structure remains the same: show up in a room, listen to someone talk, and if you’re lucky, ask questions at the end. Mobile technologies are distracting people from the very lectures they’re supposed to be listening to. The conference of the future will embrace mobile technologies so that they don’t distract attendees from the analog lecture of yesteryear. This structure can and should be inverted using mobile technologies, particularly smartphones, tablets, Wi-Fi, Bluetooth, iBeacon, and other proximity based technologies. I’m not sure how exactly this will manifest, but I’m convinced there’s a radically different, better way.

Modern conferences are clearly broken. Both of these problems tend to grow as the size of the conference grows – as it becomes easier to get lost in the noise. I’d like to tackle this problem someday if it’s still a problem. For now, it’s not a priority.

The Impending Hospital Wi-Fi Collapse

This post was originally featured on HIStalk. 

BlackBerry’s collapse is one of the greatest business learning opportunities of the past few years. It’s peppered with invaluable lessons.

This is perhaps the best analysis I’ve seen yet. As I was reading it, a particular quote struck me: "’How did they get AT&T to allow access to the full desktop web?’ Mr. Lazaridis recalled in the interview at his Waterloo office. ‘It’s going to collapse the network.’ And in fact, some time later it did."

Although I can’t find them, I recall seeing graphs quantifying that statement in 2010. I’ve provided a mockup that effectively shows the same thing. Note that the orange line is actual data and the black line is an approximation that I drew in.

The key mistake that telecom executives made was that they failed to foresee the explosion in growth of data due to iOS and Android. Through 2007, smartphones had eschewed the desktop web because they didn’t have the processing power to drive modern apps. Telecom executives assumed that the mobile apps would continue to emulate BlackBerry’s stripped down mobile web of the pre-2007 era.

The telecom executives were wrong. iPhones and Android powered a new generation of apps that were exponentially more data-hungry than their BlackBerry predecessors. This explains why the general public complained about AT&T during the early years of the iPhone. The AT&T network literally collapsed under the weight of the iPhone. This also explains why our cell phone bills continue to skyrocket: the telecoms are investing in far more infrastructure to support data growth than they had originally planned five years ago.

I foresee the exact same thing happening on hospital Wi-Fi networks, if it’s not happening already. CIOs are forecasting intra-hospital data requirements based on last year’s usage models, not next year’s apps. Most wireless communications in hospitals today are asynchronous: EHRs, CPOE, asset management tracking, pagers, mobile texting, etc. But we’re on the cusp of a major wave of real-time, synchronous communications in hospitals.

Apps that push real-time data to mobile devices are growing rapidly. AirStrip, iSirona, and Capsule Tech are piping raw data to wireless devices in real time. Vocera and Voalte are piping copious volumes of VoIP data over Wi-Fi networks. Telemedicine is exploding (although many of the telemedicine carts are wired and the rest have onboard routers to compensate for weak Wi-Fi infrastructure in hospitals). ORs, which demand stutter-free 3D HD video, tend to operate on their own wired networks that are independent from the rest of the hospital’s wired infrastructure.

Telemedicine is growing rapidly, with estimates between 18-22 percent year-over-year growth for the next few years. Telemedicine is going to require massive investments in Wi-Fi infrastructure as midlevels continue provide an increasing percentage of total patient care relative to physicians.

Moreover, we’re on the dawn of the WebRTC era (a secure protocol for real-time audio / video / image / text based communications on the web), an era that will be characterized by the commoditization of all forms of IP-based communication. Commoditization will lead to a massive wave of innovation and adoption of real-time communications apps.

These communications apps will be better, cheaper, faster, and ultimately more ubiquitous. But they will share the same Achilles’ heel: they will depend on and be bounded by the ability of mobile devices to seamlessly connect to and push data over Wi-Fi networks.