A review of Babylon's recent Investor Day
Babylon’s investor day made for an interesting session to listen to, as it did a good job storytelling about its vision of leveraging a digital first model to provide better care via value-based contracts, but it was also very light on details as to how exactly it is executing on that vision.
You can leave a session like that either very positive or negative on prospects here. Either you’re optimistic about the future vision Babylon shared, thinking that Babylon’s current state is just one step along a broader path, and believing in leadership’s ability to execute on that vision. Or, you’re concerned about the current state of technology and clinical model, the way the story glosses over important details, and lack of consistent data being shared by Babylon leadership. We’re very much in the latter camp, personally.
For us, the session invites a broader question about what we should expect of public companies. Many parts of the session felt a lot like a pitch that one might expect from an early stage VBC startup talking about the potential impact of the technology platform and clinical interventions they’re building, and as you’ll see below we were underwhelmed by that. We find ourselves expecting more from a public company, and that it’s not enough to just share a great vision, but we want to see tangible results justifying a company's ability to execute. Babylon just doesn’t seem to have that yet.
Ultimately, what really matters here is whether Babylon is providing better care to people. If it can provide care to people who otherwise don’t have access to care, and do so in a way that improves outcomes, all of these concerns will fade away. Yet it is really hard to tell from this session if Babylon has actually figured out a way to do that, or if it has just figured out a good way to acquire patients in VBC contracts from payors. Only time will tell on this front, but it would be nice to see more data from Babylon on how Babylon’s clinical model is actually driving better outcomes for patients. Without more data, we find it hard to trust that the big vision is anything more than a play to improve Babylon’s stock price, particularly given the recent management sell off at the expiration of their SPAC lock-up, which drove a steep decline in Babylon’s stock.
Our Summary Takeaways:
If you want to believe in the team and the vision, Babylon has done a really nice job telling a big vision that is easy to get excited about. But if you want a solid data-driven story about how exactly Babylon is going to execute on this vision, it simply isn’t there. The result was that we found ourselves asking this question a lot throughout the session:
“this sounds really nice, but is any of it actually real?”
Let’s take a look at what we learned from the presentation:
Perhaps the highlight of the session is that Babylon does a nice job of creating a story arch highlighting the potential of a digital first model in healthcare. At the beginning of the session, they spend time telling a patient’s story, highlighting the inconvenience of traditional, in-person care, particularly in rural settings where access to providers is a major challenge. This plays into Babylon’s overarching theory slide:
The patient story and framework for driving value was a nice way to set up the day, yet the story at the same time also highlighted some of the weaknesses that Babylon currently has in its current state. The impact of Babylon’s service for this patient was that she’s used Babylon’s virtual service a few times, and that it was convenient. To be sure, providing access to patients who otherwise wouldn’t have any is a good outcome by itself, but it invites questions as to why that is the most meaningful impact story that Babylon can share on one of its patients. It is clear Babylon provided better accessibility via its virtual offering, but it is not clear it provided better affordability or clinical quality in doing so. This sets the tone for the session in many ways - a good high level narrative that is begging for data supporting it.
Babylon shared the table below as a way of illustrating how it intends to grow the business. It expects to get to profitability once it achieves $3 billion to $4 billion in VBC revenue, with an additional $150 million - $200 million of licensing revenue, assuming Gross Margins on the VBC book of business around 10% - 15%.
This all makes sense, in theory, at least. Babylon views its operating costs as staying relatively flat over time, which if that holds, means that if Babylon is generating gross margin on each additional new member, it should be relatively easy to calculate when that reaches profitability, as Babylon appears to be doing here. The result is Babylon thinks that at a Gross Margin of 10% - 15%, it needs about $3 billion in VBC revenue, with $150 million - $200 million of licensing revenue, to generate enough gross profit that covers off on its operating margin.
Two elements of this seem tricky:
If the answer to those questions are yes, Babylon has a really large opportunity in front of it. If the answer to either of those questions is no, the entire model seems to crumble. As we’ll dig into more in the next two sections, Babylon doesn’t provide much confidence in the answer to either of those questions being yes.
One of the bigger concerns we have with Babylon is that it appears in an effort to simplify its story and tell a consistent narrative, it glosses over many details or tells them in an inconsistent manner. Here are a few examples of this:
In the opening remarks, Babylon’s CEO Ali Parsa repeatedly remarks that the problem with healthcare spending is the people, and that until you automate what those people do, you won’t be able to impact healthcare spending in this country. Parsa cites that ~66% of healthcare costs in various countries are labor. Parsa ponders the challenges that healthcare faces due to its human-driven nature in questions like this during the investor day:
“[It] take[s] 17 years for best practice in medicine to become common practice. Last year I think 11,000 papers were published on dermatology alone. How do you know that that doctor that you go to in that fancy office, how up to date are they?”
We can debate all day how reasonable it is to think we can automate care delivery, but it’s worth keeping in mind as we think about the technology that Babylon has built to date and its ability to build a profitable business. The entire premise of Babylon is that it can automate many of the costs associated with people providing care delivery, effectively making it so that technology can be used to deliver care to more and more people at a marginal cost of $0. This is why some of the details glossed over above are so important here - does Babylon’s IPA in California understand the general strategy and expectation that Babylon is going to have over time? Does it have any ability to execute on that strategy? It is worrisome at best.
Leaving those unanswered questions aside, it is clear that executing on such a vision would require a sophisticated technology infrastructure to replace human labor. Let’s dive into Babylon’s technology and what they can demonstrate currently on the technology front.
Babylon spent a good deal of the session talking about its technology capabilities and providing examples of the technology in action. Given the critical role technology will play for Babylon in keeping operating costs down while achieving margins on VBC contracts, let’s take a look at what they show on how they’re progressing in that area:
Babylon went through a demo of the AI platform it is building, which was actually pretty interesting to see live. And when you peruse the slides or watch the live demo, it certainly looks like a cool platform they’re building. Here’s the slide that shows the Health Graph platform that Babylon is building:
But as always, the questions about AI in healthcare pertain to how usable any of this actually is to influence outcomes. As the saying goes, Garbage In = Garbage Out, which is the current state of most healthcare AI. How is Babylon fixing this to make the AI usable to drive better care delivery?
Babylon attempted to answer that by sharing an anecdote about how they’re building a predictive model that assesses risk of a member being admitted to the hospital in the next twelve months that can then provide recommendations specific to the patient on actions to take. Having this capability theoretically allows Babylon to intervene early and prevent health issues from occurring.
Sounds great, right? So let’s take a look at a screenshot of the output they showed for a member with Type 2 diabetes, hypertension, and a history of smoking (this is ~57:00 into the video):
See that “recommendations” section in the middle, which is the output of the AI that is at work here? That appears to be suggesting that this patient do the following two things to stay out of the hospital:
I mean… c’mon? This is the best example that Babylon can provide at its investor day of how the incredible AI platform it is building will automate care delivery? Telling a high risk patient to stop smoking and eat better? Sometimes I feel like I am taking crazy pills when I am listening to these sorts of presentations and having companies suggest that this is the underlying technology that will enable them to perform on VBC contracts in a differentiated manner.
Babylon appears to be investing a ton of time and resources into an AI platform that today has little to no value at the point of care at the moment, despite how cool it sounds to be working on an AI platform that automates care delivery. The challenge here of course is that you need to start somewhere on building these technologies, and just because Babylon isn’t there today doesn’t mean they can’t get there in the future. It’s just unfortunate that Babylon is a public company telling this story, as it feels much more like an early stage company selling a vision to investors without much proof they can get there. But this current state is really far away from being able to automate care delivery.
To its credit, Babylon acknowledges during the session that this is a relatively simple insight and that they're working on more advanced ones, but this demo invites all sorts of questions about what it would actually take to get to the more advanced future state that Babylon envisions.
Babylon also provided a demo (around 1:04:00 of the video ) of how it is using NLP in video visits to automate clinical documentation, which is reducing documentation time for clinicians by 23%. This seems like a meaningful improvement for clinicians if Babylon has implemented this capability at scale. But again, it would be helpful to see more on how much it is being used by Babylon clinicians to understand how meaningful that stat is.
Babylon also does a nice demo on its patient facing app (starting around 1:09:00) and how it integrates various data sources on patients - Redox, Validic, and direct integration into EMRs were cited as data sources in the analyst Q&A (you can see some of those sources on the left side of the chart above). It’s worth checking out the demo video for an example of a patient facing app with lots of bells and whistles.
Yet it is a bit disconcerting to see Babylon insert this footnote on the bottom of every slide in the technology section:
Note: Some features or elements of features may be under active development, have not been commercialized and we cannot guarantee if / when the product will be delivered to members.
It’s worth keeping in mind some context here for Babylon’s business as you look at the app demo, specifically that Babylon is being assigned patients by payors, which Babylon has had no previous interaction with. Whatever data is getting into the app that Babylon is describing is either through the sources described above (Validic, etc) or the user is inputting it. All of this seems completely unrealistic and makes it seem like a demo that is not being used. We would need to see data on questions like:
Without sharing data like the above, Babylon has to know that all of this seems hard to believe. You expect to share a demo and have folks believe that rural Medicaid members in Missouri are going to have all of that data being tracked in your app, are engaging with the various prompts you are sending them, and changing the way they behave based on that? Given the AI section above, if you’re sending suggestions to people like “hey, we think you should probably consider not smoking and eating healthier,” I am not putting much weight on this.
In January 2022, Babylon announced the acquisition of DayToDay, a patient engagement app designed to support patients recovering at home from major health events. It is interesting to note this acquisition given all the capabilities Babylon spent time describing above, both in terms of its AI and its patient facing app, that it is building in house.
Despite all of the internal work on technology, there is only one slide in the clinical model section of the investor day on how technology is being used, and it’s a slide about how technology is supporting complex chronic patient initiatives… via the DaytoDay app?
Inadvertently, this again seems to highlight the reality of the current technology stack at Babylon. It seems concerning that Babylon’s core engagement tool for chronic condition management is not actually the Babylon app, but rather a startup that it acquired five months ago and operates on an entirely different tech stack than Babylon (going back to one of the core questions above of what app are members actually using?). It is hard to place much weight on anything they’re saying from a technology perspective when you see things like this in the slide deck.
Let’s revisit that question: “this sounds nice, but is any of it actually real?”
The answer from the technology section in investor day appears to be a resounding no. Despite the nice demos, we’ve been shown advanced AI capabilities that result in high risk patients being told to stop smoking and eat their fruits and veggies along with a patient-facing app that apparently isn’t even in production and Babylon is cautioning that it might not make it to members. On top of that, the technology that is cited as being used in the clinical model is the DayToDay app, which is unrelated to Babylon’s tech infrastructure. All of this makes it even more disconcerting when Babylon shares virtually no data on the impact the technology is having on care delivery or how many people are using it.
One of the oddities of the story telling in the investor presentation is that there is almost no overlap between the sections focused on VBC story and the technology platform story, despite the overarching narrative that the technology is going to be what enables performance in VBC contracts. There are certainly some areas where you can infer that technology is making things more efficient for Babylon, but that is largely left up to the listener to infer.
What is talked about in the VBC section is largely what you’d expect from a VBC organization that is thinking about implementing a virtual care model, but it’s also a reminder of how far Babylon has to go in turning this into reality.
In the presentation, Babylon highlights how it is building a virtual care team that supports the needs of members, and that care team is dictated by the needs of the member - it can include PCPs, RNs, mental health support, health coaches, etc. Exactly what you’d imagine for a next gen care delivery company to be talking about as it describes a virtual first care delivery model. And the clinical playbook for managing its VBC contracts is also what you’d expect, as the slide below depicts:
Those appear to be the set of interventions that essentially anyone managing VBC contracts would be implementing for a population. So, it’s good to see that at a high level.
But again, the concern here isn’t that Babylon doesn’t understand the story it needs to tell, but rather there is no data provided to support what it is actually doing on any of these initiatives. It seems like repeat behavior where Babylon is making capabilities sound real when they are very much in development. For instance, on the slide above, it looks as though Babylon is currently managing 11 chronic diseases. But then you look at the slide below and realize that it has only deployed three out of eleven of these programs - diabetes, mental health, and substance use. The other eight programs are currently in development, with estimated deployments in Q4 2022.
It is yet another reminder that while Babylon might be telling the right story and have right ideas, the execution is woefully inadequate at the moment, at least for what we’d expect from a public company.
One of the more interesting slides from the Investor Day is this one below (slide 43) highlighting the most common clinical needs for Babylon’s Medicaid populations. As you can see below - childbirth is the highest spend driver PMPM, while anxiety is the most common across Medicaid members.
It is worth thinking about these condition areas as it relates to the clinical model that Babylon is rolling out and the challenges it faces in doing so. Babylon has noted that a majority of the initial VBC business is Medicaid, and so it makes sense to have focused thus far on these clinical initiatives. But the initiatives highlighted raise two key questions:
It all feels a bit like Babylon is taking the approach of, do whatever it takes to get a VBC contract from a payor in order to drive growth, and then figure the rest out once you have the contract in hand.
As we mentioned earlier, it’s helpful that Babylon shows an example of a case study of how a patient has interacted with Babylon’s virtual platform, highlighting how a member in Missouri received virtual care through Babylon and received needed care that they had not previously received.
This is a good example of the promise of Babylon, and the opportunity it has in front of it if Babylon can figure out how to execute. The existing system we have all too often lets patients like Amy fall through the cracks and not receive needed care, particularly in rural settings where access to care is a challenge. So if Babylon is able to better serve patients like Amy, and continue doing so while it scales, that would be great.
Let’s again revisit that question: “this sounds nice, but is any of it actually real?”
It appears in this case that some of what Babylon is talking about here is indeed real, although very much a work in progress. Babylon at least starts to demonstrate here that it knows what it needs to have in place to execute on VBC contracts, even if it doesn’t give much confidence that it has built the ability to do so, in the current state anyway. Babylon continues to provide very little data on how the clinical model is working (aside from the slide discussed in the next section on margins and performance), and it’s not exactly helpful to hear that they have a very disciplined approach to how they’re measuring clinical performance, but then not sharing any of it during the investor day. Though the patient stories do provide nice anecdotes of the impact that Babylon is having today and provide some insight into the ability of virtual models to impact outcomes.
Again, it appears that Babylon has much work in front of it for the reality of the business to catch up with the vision being sold at the investor day.
Babylon’s investor day helped shed some light on the business model as it leans more heavily into value-based care. As they mention during the day, there are two other business models here, a licensing model and FFS care delivery. But given the emphasis on VBC in the future, we’re going to focus there.
It appears from the Investor Day that this is the general approach for signing up a VBC contract with a payor:
For instance, it appears this is exactly what happened with Home State Health in Missouri, one of Babylon’s first VBC payor contracts, which apparently has assigned Babylon ~20,000 members. The press release confirms this general approach with this sentence: “Babylon has started outreach to Home State Health members that have been assigned to them.” If you’re a payor, there isn’t a ton of risk in sending Babylon a list of your hardest to reach members and attempting to have them engage those members via the virtual first platform. The payor gets certainty in its medical expense on a hard to reach population, and if Babylon is able to engage that population successfully, it captures the upside.
As Babylon notes during the investor day, they’ve changed their strategy to signing up VBC contracts as they’ve entered the US - originally, Babylon expected that payors would only sign up VBC contracts after having FFS contracts in place, but now Babylon is seeing payors want to move straight to VBC contracts. That seems unusual for payors, and seems to indicate that Babylon is offering them a no-risk solution, which aligns with what is described above. It appears to be working for Babylon, as they shared during the session that growth is not its challenge. Babylon has ~150 payors in the pipeline, and expects to be in over 20 states shortly.
This is one area that does provide a positive indicator as to the Babylon experience, as payors ultimately do care about their member experience and you would assume they are doing their diligence in partnering with Babylon. It’s hard to imagine a payor handing over 20,000 members to a partner like Babylon without having confidence that the member experience will be ok. So it’s a positive sign that Babylon is signing these contracts. What will be interesting to watch over time is whether these contracts essentially amount to pilots where payors are testing to see if Babylon can engage evidence otherwise hard-to-reach Medicaid members. It’s not hard to imagine why payors would be willing to sign up for a contract like this with Babylon - presumably these are high cost patients who are hard to reach, and therefore represent financial uncertainty for a payor’s MLR. Signing up those members to a provider willing to take a capitation payment and associated financial risk reduces that financial risk for payors and transfers it to Babylon. If Babylon’s clinical approach successfully reduces costs, it’s a win for both parties. If it doesn’t, Babylon is left holding the bag.
The major risks here to keep an eye on over time of course are whether Babylon can set an appropriate baseline with the payor on what the costs of these populations are, and whether Babylon can engage the members that the payors are sending to it.
Given this approach to signing up VBC lives, it makes sense as to why Babylon places such a high emphasis on engagement in various contracts - Babylon’s ability to influence cost of care is entirely dependent on its ability to reach members.
You can see why Babylon places such a large emphasis on increasing this number. In Missouri, the results indicate that after 20 weeks, almost half a year, Babylon had “signed up” (it is not clear exactly what that means) what appears to be 1% of the 20,000 members they were assigned in Missouri, meaning ~200 of those members.
In other words, six months into a VBC contract managing 20,000 lives, Babylon had successfully had an interaction with 200 of those people, and there were 19,800 for whom it was unable to interact with. Call me crazy but that does not exactly seem like a scenario for success if your belief is you can better manage the total cost of care for those 20,000 people.
By the end of year 1 in Missouri that grew to about 12%, which would be around 2,400 members, still leaving 17,600 members unengaged. And note that this makes the assumption that this chart is referring to *all* members that Babylon is assigned, which probably isn’t the case given the title refers to “high risk prospects”. So, it’s likely an even smaller number of members that Babylon has engaged here. Given their approach to VBC contracting, this has to be concerning that they’re taking full risk on a population whom they’re not even engaging.
The rapid growth in Georgia / Mississippi versus New York / Missouri drives questions as to what exactly Babylon is doing differently, how they are driving such a substantial improvement in engagement rates, and what they’re learning that is repeatable. Again, Babylon doesn’t share much tangible information about what is actually happening here, and given their results elsewhere, we have questions as to whether this growth is sustainable. Babylon hasn’t shared the mechanism as to how they know this is a repeatable result, which would be helpful to hear more about to generate confidence that Babylon is able to do this in other markets. And then, even if it is, you’re still talking about engaging 16% of a population that you are responsible for managing total cost of care on.
To its credit, Babylon shared the slide below as an update on their performance thus far in VBC contracts and included a MLR metric in the conversation:
The Medicare Advantage margin here actually is pretty impressive - at 14.3% in its longest tenured cohort, that would be a pretty solid MLR for a primary care group taking risk on a Medicare Advantage population. But again, Babylon glosses over details with these results and until we start seeing them consistently report on performance (numerators, denominators, and size of populations) across the book of business, it’s hard to trust the numbers.
As a good example of that, it comes up in the analyst Q&A that Babylon isn’t actually taking full capitation on the commercial population, as it turns out that is largely the Meritage book of business in California where they have professional services cap agreements with payors in Northern California. So while it is good to see a healthy 19% MLR on that book of business, it’s not clear that we yet have the total story of what is going on here, or how Babylon’s business will perform over time.
And of course the Medicaid margin is concerning, which goes back to how are you actually managing that population virtually, which will be worth watching over time.
Babylon articulates below the “formula” for achieving success in VBC contracts, building on the clinical model claims savings, with revenue generation from coding and optimizing the cost of care delivery.
This is yet another instance of Babylon saying all the right things from a storytelling perspective, but providing little evidence to back it up. Particularly disconcerting in this section was the number of times that Babylon mentioned the emphasis it has placed on measuring the savings generated from the claims savings programs and the rigor it has in place internally, with targeted savings amounts for each of the six programs on the slide. Yet none of that is shared here. Why not? Again, it doesn’t provide much confidence that Babylon has actually set targets it has confidence that it can achieve.
Babylon shared this rather astonishing chart - highlighting that its providers in the US are only at 59.1% capacity.
It’d be helpful to know what the numerator and denominator is for this percentage - how many providers does Babylon have in the US at this point? I would be curious to know if the California IPA is included in this number, but I would suspect that it is not, and that this is only a number for virtual-only clinicians.
This appears to be what is driving down Babylon’s operating cost currently, which should naturally happen as Babylon scales membership to meet capacity of its clinicians. But, what happens when Babylon reaches capacity for those clinicians, both in terms of clinical and financial outcomes? It is yet another sign how early we are in the ramp up of the business here.
Let’s again revisit that question: “this sounds nice, but is any of it actually real?”
Given the contracts Babylon has been signing up and the future pipeline it is citing, this part of the business does indeed seem real, at least in terms of revenue Babylon is bringing in. Babylon seems to be finding success signing up lives to VBC contracts and growing revenue associated with it. Yet it does seem like a particularly risky strategy when Babylon is responsible for total cost of care for members it is struggling to engage a vast majority of. How is it going to impact medical spend if it is only engaging 30% of high risk patients?
Babylon’s narrative, which glosses over key details like the role of the California IPA in growing the VBC business, still invites a number of questions as to what will actually drive repeat performance on VBC contracts moving forward.
As you may have picked up on by now, we have some major concerns with Babylon as a public company. While the vision of the potential impact of digital first models is great, Babylon provides very little justification actually supporting the story with meaningful data demonstrating its ability to execute on this vision - this is true for the technology (where the AI is underwhelming and the patient app seems like a mockup), the care model (where chronic condition management is still being built), to the business model (where Babylon needs to engage members and prove it can manage costs). All of these things would be acceptable for an early stage startup, but it seems like we should expect more of public companies.
To be sure, Babylon does a good job sharing a vision about how it is going to use virtual care and AI to usher in a new era of healthcare, one that improves accessibility, improves outcomes via standardized care pathways, and reduces costs. Even if we might quibble a bit with whether AI can actually replace / standardize care delivery, we can’t argue with the impact a model like this could have if successful.
Frankly, it seems like this is one of the unfortunate byproducts of the SPAC craze over the past two years. There is so little information to evaluate this company on, you’re left choosing whether you believe their narrative or not, much like an early stage venture investor who is investing in a team with a big vision. But the reality is that Babylon is a public company, not a venture backed startup, and this investor day happens with the backdrop of management selling so many shares at the expiration of their lockup that they single-handedly drove the stock price down by 50%.
We’re going to wait and see more data from Babylon on how this model is working, and until then, are going to be very skeptical about any of the claims they’re making. They’ve shown they can grow revenue but not much beyond that.