Looking back at Bright's early strategy to understand whether its individual exchange business was always doomed from the start
Obviously Bright led the news last week with their retreat from the individual exchanges and decision to focus on Medicare Advantage via NeueHealth. Obviously given the ACA was Bright’s initial focus and core business, it’s a massive setback for the organization. But what we’ve been curious about is how predictable was this failure at the beginning of Bright? Was this inevitable from Day 1, and if not, what happened along the way that drove this outcome?
To do so, we’ve attempted to journey back in time, specifically to early 2016 and try to understand what the world looked like back then. At the time, Bright looked like a sound investment opportunity. It was taking a solid approach to building a profitable exchange business with a relatively simple strategy and feasible rationale for why it could win in the space - Bright would partner with local health systems to develop narrow network plans that were shown to be profitable, and the big incumbents wouldn’t be able to partner with health systems in the same way. It’s a narrative that makes a lot of sense. Yet the challenge that Bright appears to have had over time is getting away from that simple strategy as it chased the opportunity to scale in order to manage ballooning operating expenses.
Facilitating this exercise is the fact that Bessemer has their original investment memo in Bright from March 2016 posted publicly on its website. The memo does a very nice job providing insight into Bright’s early strategy and the opportunity Bessemer saw in making the investment. What’s most notable to us is how different Bright looks today (even before it exited the exchanges) in terms of its overall strategy as a public company.
We’ll note the purpose of this exercise isn’t to fault Bessemer for where the memo was off-base, but rather to use it as a learning tool to understand where assumptions may have proven faulty, or the strategy deviated to make those assumptions irrelevant, so we can learn about future innovation opportunities. Kudos to Bessemer for putting it out there and opening themselves up to this sort of analysis. It’d be easy for them to take this memo down at any point. The fact that the result has turned out to be wrong doesn’t necessarily mean that the process was wrong, and either way we all learn from the transparency.
We’re going to go through this walk down memory lane starting with the exchange environment at the time, before going into Bright’s beginnings in Colorado and challenges with growth from there. Here are the sections we’ll dive into:
It’s impossible to get into Bright’s history without first starting with Colorado HealthOP, a nonprofit insurance cooperative in Colorado, which is really where Bright found its origins. Colorado HealthOP was one of the state insurance co-ops that was created along with the exchanges, designed to encourage competition in the individual markets. Originally, there were 23 of these nonprofit co-ops across the country, although as of 2020, that number had dwindled to three (link).
The story of the co-ops deserves its own article in many regards, but for now, we’ll just touch on a key piece for this story. A big part of the failure of co-ops was the federal government reneging on its obligation to pay out risk corridor payments that it had committed to making as part of the Affordable Care Act legislation. These payments were an essential feature of the “3Rs”, a critical mechanism for the exchanges to function and allow insurers to participate without worrying about potential adverse selection issues. The risk corridors in particular were important because if an insurer got a relatively healthier population versus others, that insurer would make a payment to the government, and in turn the government would make a payment to insurers that had sicker populations. The only problem is that the federal government decided to pay only a fraction of what it had committed to, at least until 2020 when the case went all the way up to the Supreme Court, which ruled 8-1 that the government was obligated to make the full payment to insurers (link).
At the time, many co-ops were still struggling in their early years and were not yet profitable. The Colorado HealthOP was one of these struggling co-ops. In 2015, the Colorado Department of Insurance (DOI) decided to shut down the Colorado HealthOP, at least in part due to the fact that the federal government decided not to pay Colorado HealthOP the $16 million risk corridor payment that the government had committed to. Instead, the government only paid out $2 million, which caused Colorado HealthOP to be out of compliance with the Colorado DOI’s regulatory reserve requirements, and so the Colorado DOI shut it down (Link). This move wasn’t without controversy, the Colorado HealthOP sued the DOI in October 2015 to keep operating, referencing in the lawsuit that a venture capital loan was a potential solution to keep it afloat (link). But nonetheless, the Colorado HealthOP now had to figure out what to do next.
As we’ll get into in a bit, the Colorado HealthOP was offering Coloradans two plan design types, a PPO (“preferred provider organization”) and an EPO (“exclusive provider organization”). The EPO plan is essentially another term for an HMO that is governed by the Department of Insurance in a state, and it functions the same way as an HMO in that its a narrow network plan design. In total, Colorado HealthOP served ~80,000 Coloradan’s in 2015. This is relevant as we dive into Bessemer’s investment thesis below, which mentioned that it would “relaunch” a narrow-network product that had 40,000 members and an MLR of ~65% - 75%. Connecting the dots here, it seems pretty clear this was Colorado HealthOP’s EPO membership. The EPO book of business was apparently quite profitable on an MLR basis at least, but it was being dragged down by its PPO membership, which presumably had a MLR well-north of 100%.
It’s also important to remember that the health insurance exchanges were in a very different place in 2015 - 2016 while Bright was being ideated, with much more existential dread about the future of the exchanges. Whereas today, the big insurers are racing to add markets in the exchanges, back in this time period, it was a bloodbath. The large carriers were all running for the exits, concerned over the viability of the exchanges and the political battle at hand. Aetna’s CEO went so far to say publicly that the exchanges were in the dreaded “death spiral” (link). Of course, this demise ended up being overstated, as many of the big insurers have figured out how to operate profitably and are now growing their exchange businesses substantially.
Rather astutely, Bright’s founding team and Bessemer saw a massive financial opportunity in this environment to build a profitable insurer on the exchanges leveraging what the co-ops had already learned - that with a narrow network product you could attract a profitable book of business, leaving the losses behind. And then also note how completely different that strategy is than that of Bright-as-a-public company, which we’ll come back to later.
The original investment thesis for Bright was to take the profitable EPO membership and narrow network product design from the Colorado HealthOP and leave behind the unprofitable business. As Bessemer noted in the document:
Bright Health will “relaunch” a narrow-network product in Colorado that obtained 40K members in one year (~$110M in premiums revenue) with an impressively low unadjusted Medical Loss Ratio (MLR) of ~65-75%
That sentence from the memo tells you everything you need to know about what Bessemer thought Bright was doing - resurrecting a failed local plan that had an EPO product that was working and using VC funding to grow a viable business moving forward. If anything, it’s a relatively unremarkable investment idea that wasn’t about fundamentally changing the industry. Rather it was about finding a profitable opportunity in a large market that other incumbents were struggling, and growing that business profitably by partnering with a large local health system that could help drive membership volume to Bright while negotiating lower rates with the health system in exchange for driving them the membership volume. Realistically, that sounds a lot more like a Private Equity play than a Venture growth strategy, and the exit outcome scenarios Bessemer highlight (which we’ll get to more below) further give that impression.
Check out this section of the investment memo to see how Bessemer viewed the crux of the opportunity:
It provides a compelling and straightforward case at the time as to why Bessemer expected Bright to be able to operate an insurance plan profitably. Bright would:
Note for a second how the technology platform is only briefly described as essentially a secondary priority for Bright - in addition to them offering a low cost, profitable plan via the mechanisms described above, they will invest in technology to drive a better experience. The word “technology” only appears twice in the investment memo, once in the paragraph above, which gives some indication of how much diligence Bessemer felt it had to do on the technology platform. And the second mention of technology? This sentence:
“The company is currently recruiting a technology lead and has a large number of really interesting candidates with strong consumer tech and health backgrounds.”
The technology platform was so unimportant to the original business case that it didn’t concern Bessemer in the slightest that there wasn’t a technology leader in place at the time of the investment. Compare this with Centura, the health system Bright partnered with, which was mentioned nine times in the memo and that relationship appears to have been diligence extensively. It should be pretty clear that this business was never intended to be about the technology play, it was all about partnering closely with a local health system. The technology centric focus we hear from Bright today has been an added piece of the story around Bright's IPO.
When we look at that original investment thesis, it looks like something that should have worked really well! The patient population was already proven to be very profitable in an environment where the exchanges were generally unprofitable for payors (as the memo notes, PPO products were routinely experiencing MLRs above 100%). Bessemer was anticipating that Bright would be paying rebates every year in order to hit the 80% medical loss ratio mandate on the exchanges, suggesting the plan was to use the narrow network product to attract healthy risk. Beyond partnering with Centura, Bright went so far as to hire folks from Colorado HealthOP to lead the launch of the health plan in Colorado and de-risk the transition from the co-op to Bright. In many ways, this was a slam dunk opportunity. It should have worked, profitably, from close to day one. After all, it was already profitable inside Colorado HealthOP prior to Bright getting involved.
Understandably, Bessemer didn’t see much risk in this approach. When you look at how Bessemer modeled potential outcome scenarios, it only applied a 20% probability to the company going belly-up. Here is a Series A venture investment where 80% of the time, Bessemer thought it would at least return its investment. The remainder of outcomes ranged from Bright being a modestly successful health plan in Colorado to becoming the next UHG. Seems like easy money for Bessemer, no?
This chart highlights how the “risk” associated with this investment doesn’t exactly look like that of a typical venture investment. The range of outcomes here, and associated financial model below, feel much more like a private equity-style cash flow generating business.
To wrap the investment memo, Bessemer outlined four key risks: 1. It was a large investment with little track record, 2. Health insurance margins are small, 3. Customer Acquisition Costs (CAC) in the exchanges were unknown, 4. Political / regulatory risks for the exchanges. Note what sticks out to us here was Bessemer’s following question related to the first risk, and it’s something we’ll come back to later in this paper:
can Bright Health actually keep medical costs down and run a profitable health plan through 2017?
We’ll come back to it because it seems to represent the interesting conflict Bright’s investors faced after launching in 2017, assuming that the 2018 data is a good proxy for 2017. Medical costs looked quite good, but the plan was significantly more unprofitable than expected, driven by significantly higher than expected operating expenses. Note that operating expenses was a topic that was not touched on heavily in the investment memo. It seemed that because of the UHC ties, Bessemer assumed that the operations side of the insurance business would be straightforward to manage. This in many ways appears to be the Achilles heel of Bright - underestimating the complexity, cost, and focus required to stand up a health insurance plan. Underestimating that then required Bright to fundamentally alter its growth strategy, seeking to expand quickly to cover off on its cost base.
Let’s dig in further to what the financials look like.
You can see Bessemer’s investment thesis starting to go awry in the financial model when you compare what Bessemer expected to be the case versus what actually materialized between 2018 and 2021 based on Bright’s SEC filings. The two different sets of financials paint a very different picture:
Looking through the financials reveals a number of interesting insights:
To put a finer point on two of the bigger issues in those financial models - look at these two charts highlighting the differences in actual vs projected MLR and OpEx as a % of Revenue:
It goes without saying that the Actual results in those two charts are deeply problematic for an insurer, and that Bessemer’s projections here ended up being very different from what actually happened.
When Bright pitched Bessemer back in 2016, it appeared that Bright envisioned (and Bessemer bought into) Bright as a regional payor that would be deeply integrated with a handful of leading health systems in their local markets. The Centura partnership provided a great test case for that. Keep in mind Bessemer was thinking that Bright would be operating in only three markets, attracting both individual and MA members in all markets, per the memo. Even in the top 1% “home run” outcome Bessemer forecasted, Bright would only be generating $1.5 billion of revenue in 2021, versus the $4 billion Bright ended up generating. It should tell you something that a leading VC’s “home run” revenue scenario actually ended up only being ~38% of what actual revenue ended up being within five years.
Bright was originally founded on the notion that it was going to focus on building narrow network products with a committed provider that is a leader in the local market. By building these deep relationships and slowly expanding into new markets, Bright would be able to repeat the financial success that the Colorado HealthOP had with its EPO product offering with Centura. These relationships would allow it to build upon the three factors that differentiated Bright: better rates, data sharing with providers to intervene with high cost members, and risk adjustment. This pitch is a pretty reasonable one, and one that had the benefit of Colorado HealthOP’s experience and relationship with Centura in Colorado.
If you go back to the scenario analysis chart, Bessemer viewed this outcome, where Bright was a successful plan in Colorado and a few other markets, as being the most likely outcome. Bessemer modeled that outcome as occurring 36% of the time. Obviously, Bright’s ambitions ended up going after the “home run” scenario. So what happened here? Let’s now revisit that statement we highlighted above from Bessemer calling out the key risk for Bright investment:
can Bright Health actually keep medical costs down and run a profitable health plan through 2017?
Looking at the financials from 2018, it appears the answer to that question was likely a version of “definitely maybe”, which is really problematic. Bright posted a solid MLR in 2018, and presumably also would have in 2017. But Bright also had incredibly high operating expenses in 2018, likely associated with the costs of standing up a new health plan. So Bright was faced with a decision point. Given the amount of operational expense it was incurring, the original slow burn growth strategy seemed no longer viable - Bright had a massive profitability problem driven by operating expenses, not by MLR. So how do you solve that problem?
You spread the operating expense you’ve incurred over a larger membership base and start expanding much quicker than originally expected. You can see this playing out in Bright’s press releases over time from 2016. Bright expands both geographically and strategically quite quickly.
In June 2017, Bright announced an additional $160 million in Series B funding, which Bright used to expand into three markets in 2018 - Colorado, Arizona, and Alabama. This still sounds similar enough to the Bessemer memo, right? Then, in mid 2018, Bright announced that it would triple its geographic footprint, expanding into nine new markets across four new states. Bright followed this expansion announcement with another $200 million of funding announced in November 2018.
Over that period, Bright launched its new markets with care partners in:
You might have noticed from the list above that Bright began expanding by partnering with systems on MA products, not IFP products. This might seem like a minor shift, and generally in line with Bessemer’s memo in terms of where Bright was expecting membership growth, but it belies a more important point: it indicates that Bright likely wasn’t forming as deep of relationships with these health systems. Very practically, these deals would have less volume going through the system, which implies that the health system would both be less likely to negotiate favorable rates with Bright, and also less likely to spend the time deeply integrating with Bright. Note that essentially negates the entire business case Bessemer laid out.
If Bright had grown membership significantly, perhaps this issue could have been overcome. But then Bright’s organic Medicare Advantage growth looked like this:
You can imagine how these relationships with health systems could deteriorate even further when seeing such low membership growth. What health system is going to spend the time on a partnership that is netting them a few hundred members? No, they’re going to spend their time with the other payors who are actually insuring a meaningful portion of their patients.
Our best guess of what happened around this time - forming these deep relationships in new markets was proving quite challenging and at the same time Bright started to see that growth was not nearly what it expected it to be in each market. This dynamic undermined the original approach of slowly entering markets in a profitable manner. The way to counteract slow growth that seemed to be threefold:
Whether Bright was intentionally pursuing a fourth strategy - underpricing its insurance product in order to drive IFP growth, is up for debate. Certainly given the MLR results Bright has seen, as an outside onlooker it certainly seems like the answer to that question is yes, even if Bright maintained publicly during recent earnings calls that it was taking a disciplined approach to pricing. So perhaps it wasn’t intentional mispricing on Bright’s part, but instead the organization just became so operationally complex so quickly that its leadership was unable to manage all of the distractions and just made fault assumptions in pricing the IFP product. Either way, that MLR trend speaks for itself.
And in many ways, this is the most confusing part of the Bright story, where it “jumped the shark,” so to speak. At some point along the way, it appears Bright unlearned how to price risk in its IFP business, which is a really bad thing to unlearn if you’re an insurer. I say the word “unlearn” deliberately there. Going back to the original idea of Bright, they knew very well the importance of aggregating profitable membership - it was a core part of the founding story in taking only the profitable EPO plan design membership from Colorado HealthOP and leaving the unprofitable PPO membership behind. Bessemer also demonstrated they knew this in the original investment memo - check out this statement from the risk section:
Moreover, while revenues are really big in health insurance plans, margins are small! The slightest variance in medical loss ratios will make a huge difference in the performance of the company (and their need to raise additional capital). Additionally, Risk Adjustment, which forces plans to pay out if their membership base is less risky than average (and in return, plans with more risky membership get paid), complicates the actuarial analysis here. All of this makes the actuarial analysis and premium pricing critical to successfully running a profitable plan. We do think the Bright team is the best equipped we have seen to be successful in this effort. Milliman’s work has given us more comfort on the variance related to risk adjustment, but this will be a key focus for the company
The common trope said about Bright and others is that VCs don’t know that insurers shouldn’t be valued on revenue. But from the paragraph above, it sure seems like Bessemer was quite aware of that dynamic, with Bessemer explicitly calling out their diligence work with Milliman on this topic. Clearly they knew the importance of pricing the product correctly, and even had Milliman telling them the approach was sound back in 2016.
Our best guess is that as Bright incurred more operational expenses than expected, it had to raise more venture funding than intended. As it raised larger and larger rounds from outside VCs, got boxed into chasing the “home run” exit scenario, and found it had created a flywheel that was resulting in the organization losing more and more money that it couldn’t get out of: Increased Losses -> Increased VC Investment -> Increased Growth Appetite -> Increased Losses. This certainly explains the MLR chart over time.
It’s easy to wonder what would have happened here had Bright and its investors held to the more private equity style growth envisioned in the original investment memo - would Bright have been able to make the business work on the exchanges had it prioritized slow, yet profitable growth? Who knows. What we do know is that while Bright certainly demonstrated it was able to grow membership faster than it or Bessemer seemed to expect, but it did so at the expense of pricing and operational discipline, as evidenced by the financials above.
The most salient example of Bright’s operational mismanagement is the $1 million fine that Bright received earlier this year from the Colorado Department of Insurance, after the DOI received over 100 complaints from consumers and providers in 2021 about various operational issues inside of Bright. It’s remarkable that five years after launching in Colorado with a deep partnership with a leading health system, that Bright had botched the insurance plan so badly that it was being fined because of consumer and provider complaints. It’s a real head-scratcher.
Ultimately, the result is that the company described in Bessemer’s investment memo no longer exists, and the biggest thing the company has going for it now is that investors have plowed too much money into this thing at this point to let it fail completely. Bright’s core business is now its services arm NeueHealth, a business that is largely a handful of acquisitions that Bright has made over the past few years. There’s some irony that Bright is now fundamentally a technology business, when Bessemer’s original diligence of the company regarded technology as an afterthought. It goes to show how far this version of the business was from the minds of Bright’s leading investors when it raised its first round of funding only six years ago. Of course, since 2016 a number of companies have shown that variations of the Neue model can be financially successful, particularly when focusing on the Medicare Advantage population, including Optum, VillageMD, agilon, Alignment, and others.
With the IFP failure now largely in the rearview mirror, Bright has the opportunity to hit the reset button moving forward and reinvent itself as a new (get it? Neue!) Medicare Advantage business. Given how badly it appears to have botched the original opportunity it had on the exchanges over the past few years, it will need to.
“The number one driver for consumer purchasing behavior on exchanges has been price. Consumers also consider whether their physicians are “in network”, but price has proven much more important as long as the network is considered adequate. Additionally, new health plans have been successful at getting market share and consumers haven’t demonstrated a strong preference for national brands.”
Bessemer is highlighting what would have been Bright’s challenge it would have inevitably needed to face it it continued raising more money - if an insurer raises prices to get to profitability, do they retain any members, or just shift to another plan? You can see Bessemer's answer to that question above - they'd shift away. Fundamentally, this seems like why Bright no longer has an exchange business. People are purchasing the product because it is cheap, and if prices go up, they will move to another product. Bright has now resolved that challenge by exiting the exchanges and focusing on greener pastures in Medicare Advantage. We’ll be keeping a close eye on Oscar as it faces the same dilemma, and it will be interesting to watch how it seeks to resolve it from here.