An Epic Dragon Hunt

Most of my Twitter followers know that Nuance is one of my favorite stocks to talk about. For those not familiar, the following Tweet is an extremely brief summary of the setup. 

The most exciting aspect of Nuance is their Dragon Ambient Experience (DAX) product. This blog post will dissect my views on the size of the total addressable market (TAM) of DAX, the emerging competition for this TAM, and why I think DAX is best positioned to win the market. DAX is currently in the lead, but other companies have embarked on an epic hunt to overtake Dragon.

For those not familiar with the DAX product, I highly suggest reading or skimming the whitepaper linked in the following Tweet. 

It is clear from the whitepaper that DAX is a revolutionary product. It increases physician efficiency, improves quality of care, and decreases physician burnout. For these reasons, DAX has the potential to meaningfully increase Nuance’s revenue. Nuance estimated that the US TAM of DAX was $10 billion dollars in their 2019 Investor Day (over 5 times larger than current Nuance revenue levels). This was determined by assuming a price of $10,000 per US physician (~1 million total physicians). Nuance used $10,000 because it represents roughly 10% of the estimated incremental revenue generated by increased physician efficiencies in the early trials. Data from ongoing trials portrays significantly higher than expected levels of incremental customer revenues attributable to DAX. This implies the possibility of a larger US TAM than the one presented at the 2019 Investor Day.

My channel checks on DAX pricing have ranged from $12,000 to $20,000 per user. Based on these prices, US TAM could be 20% to 100% higher than Nuance‘s initial estimates. These channel checks more than support Nuance’s estimated DAX TAM and suggest that there could be upside. I think it is also prudent to consider the ~1 million physicians in Europe (in countries where Nuance Healthcare currently operates) to be a reasonable incremental addition to the TAM. Although the DAX product is not currently available outside of the US, Nuance is aggressively expanding sales efforts of prequal cloud products, such as Dragon Medical One (DMO), across Europe. For those not familiar, DMO is cloud based software that transcribes orally dictated healthcare notes for medical record purposes.

There are significant regulatory barriers to entry in Europe surrounding healthcare data. Many counties require data to be stored in specific manners and locations. Nuance currently operates five data centers in Europe to satisfy these requirements. As Nuance expands its infrastructure, it will also expand its European salesforce. These steps provide the groundwork for Nuance to unlock the DAX TAM in Europe down the road.

I think about the DAX TAM as 2 million physicians (1 million from the US and 1 million from Europe in countries where Nuance Healthcare currently operates) at $12,000 per user (the low end of my channel checks) or $24 billion. Surely Nuance is not the only company attacking this massive market.

From everything that I have read, DAX is the leading product on the market for fully automated clinical note creation. Additionally, Nuance has recently acquired the distant second place contender, Saykara. I like the Saykara acquisition not only for the talented research and development professionals that were brought onto the Nuance team, but more importantly because their other competitor, M*Modal, cannot acquire Saykara’s technology. In this sense, the acquisition is both offensive and defensive.

M*Modal has a medical speech dictation product that directly competes with DMO. The two companies nearly have a duopoly in physician transcription software. Based on testimonials from several implementation consultants, DMO is a far superior product because it is faster and more accurate. Even though DMO is ~30% more expensive, customers that churn to M*Modal’s cheaper solution generally are quick to switch back to DMO.

M*Modal is in a tough competitive spot. Their current product is inferior to the Nuance counterpart, and Nuance has already built the best next generation offering with DAX. An M*Modal VP of Product Solutions directly mentioned that they are working with Epic Systems (Epic) to create this product on LinkedIn, as captured below. (To reiterate a point from my last post, LinkedIn is an excellent tool for insights that cannot be picked up in SEC filings or conference calls.)

Earlier this year, Epic’s founder, and CEO Judy Faulkner said “Epic is still at least two years away from launching an AI component that can listen in on the conversation between the clinician and patient and then draft orders, place a plan of care and close the visit". In order to close a visit within Epic’s electronic medical record, the note must be one hundred percent complete. This means that Epic could be attempting to build a product that competes with DAX.

Epic is the leading electronic medical record provider in the US. The company is incredibly well run and possesses a unique and driven culture. Faulkner has vowed to never go public and never acquire another company. Epic seldom partners with other software companies and generally tries to build software components from scratch, whenever possible. Nuance and M*Modal both currently partner with Epic, which puts both companies in an exclusive club.

Based on a conversation with a knowledgeable health tech product executive, Epic sometimes likes to create competition to prevent adjacent software providers from achieving monopolies. Epic does not have core competencies in speech recognition. This makes constructing a product from scratch to compete with DAX incredibly difficult and time consuming. Since Epic already partners with M*Modal, and M*Modal is a speech recognition company that is being outcompeted by Nuance, it makes sense for M*Modal to take any help that they can get.

It is still too early to tell what Epic’s end game is. One possibility is that this is just a side experiment for Epic. It could also be an attempt to prop up M*Modal to try and prevent a DAX monopoly. A DAX monopoly could shift electronic medical record industry control away from Epic and even put them in a very tough spot if say Nuance merged with Cerner (Epic’s largest competitor). The final possibility that comes to mind is that Epic could be seriously internally attacking the speech recognition market. It is engrained in Epic’s culture to work tirelessly to develop adjacent software inhouse when possible.  

If Epic can create a DAX alternative within their electronic medical record, that could harm Nuance, since Nuance’s products must integrate and function on top of the client’s medical record systems. Many of Nuance’s largest clients use Epic. The scenario where Epic is merely assisting M*Modal to create a DAX alternative is significantly less threatening to Nuance since distribution into Epic would not be undercut. Even if Epic and/or M*Modal can create a DAX alternative in two years, DAX will have a strong technological advantage.

DAX uses artificial intelligence (AI) and machine learning (ML) to automate the creation of a medical record. The AI and machine learning rely on data to learn and function correctly. Data is collected each time that DAX is used for a patient encounter. Each at bat that DAX gets, the product improves incrementally with ML. I have read one account of a DAX user receiving instant medical record inputs upon the conclusion of the patient interaction. To my knowledge, DAX is the only current solution for a completely automated note based solely on the patient interaction. This firmly established Nuance as the current market leader. Not only does DAX have a technological moat today, but anyone trying to compete is chasing a moving target that improves itself every single day. Importantly, the scale that Nuance has already won means DAX will improve the fastest because it has the most users and is getting the most at bats.

The moving target moat that DAX has, and will continue to grow, means that anything Epic or M*Modal can build over the next two years will likely be inferior to DAX. Customers are willing to pay a meaningful premium for DMO over M*Modal’s inferior solution, and there is no reason that this behavior will not continue in the next product generation with DAX.

Today, DAX is still not perfect. It is an extremely complicated solution and needs many more at bats. Some DAX medical record outputs still need to be manually reviewed to ensure quality. This will likely not be the case forever. Based on conversations with users over the last year, it is clear to me that DAX is rapidly improving. The accuracy has drastically improved in the initial group of specialties where DAX first launched. These five specialties, including Dermatology and Orthopedics, were selected for their limited relative vocabulary banks. DAX has since expanded to eleven other specialties including primary care, the specialty with the largest vocabulary bank. In the current product state, DAX provides significantly more value to physicians that spend more of their time conversing with patients compared to physicians that spend more of their time in surgery. I expect that incremental product improvements over time, such as machine vision, will help to bridge this gap and justify the pricing of DAX to as many physicians as possible. I am excited to continue to follow the DAX product evolution and hopefully watch DAX survive the Epic Dragon hunt. My money is on the Dragon, literally.