2022 Reflections

I reflect on my journey thus far and how I plan to improve my process going forward

Summary

• After some initial successes, I overconcentrated my portfolio in earlier-stage names in order to capture right-tail outcomes and earn the highest possible IRRs. However, this also exposed me to left-tail risks that I significantly underweighted.

• I’ve reduced my allocations to high-risk companies and plan to reserve my largest 15%+ positions only for rare low-risk, high-reward opportunities. I also plan to diversify my portfolio to 10 to 15 names in order to hedge against left-tail risks.

• I’ve also spent the past year broadening my coverage outside of growth/tech to include more established, predictable companies with attractive shareholder yields, disciplined capital allocation, and a higher margin of safety. I aim to practice greater factor diversification in order to generate more stable returns.

• Finally, I aim to do a better job of identifying and cutting my losers quickly. To accomplish this, I will continue to do pre-mortems and quantify my thesis as much as possible, selling when my thesis breaks. I will also learn to short in order to see things from a new perspective and improve my timing and risk management.

---

With 2022 finally over, I thought now would be a good time to reflect on what has certainly proven to be a challenging year and share my experiences and learnings.

I will start off by explaining my background and investing journey thus far, before diving into my process, my learnings from my portfolio drawdown, and what I plan to do differently in the future.

My Background

Growing up, I had a big interest in strategy games. I enjoyed the competition and non-linear gameplay and would constantly iterate my strategies until I had perfected them. It wasn’t until my third year of university and finally had some disposable income that I discovered the world of investing. Being a naturally curious and competitive person, I marveled at the idea that I could make a career out of applying the right mental models to understand the world more clearly than others. I soon began to develop a passion for researching businesses and continued to iterate my process while putting my own money at risk.

What started as a hobby soon turned into an obsession as I was trying to learn from as many investors as I could to develop my own process. I made mistakes along the way and had to learn the hard way many times. I lost money investing in options, in stocks I didn’t understand, and from selling at the wrong times. However, each experience built discipline and helped me become a better and more resilient investor. 

After graduating from Queen’s University in Canada with my Bachelor’s of Commerce degree in 2019, I worked as a technology consultant. I continued learning about investing on the side and studied the process of other great investors like Munger, Bill Miller, Peter Lynch, and of course Buffett. It was also around this time (fall 2019) that I stumbled on a company called Livongo. It was a recent health tech IPO that had fallen over 50% from its IPO price yet was growing revenues very fast (>100%) and was trading at a significant discount to SaaS comps.

Although the company wasn’t pure SaaS, it had a recurring revenue base from charging a per member per month fee with a high retention rate and NPS. It was a win-win for patients and employers as patients with diabetes would get free support through glucose monitors and coaching, while employers would benefit from 3.7x ROI as patients would file fewer claims. Livongo also had a first-mover advantage and built a highly efficient go-to-market in educating channel partners.

Needless to say, I had high conviction in the company and believed I understood the opportunity, so I decided to write a Seeking Alpha article on it. At the same time, I started a Twitter account where I began posting my thoughts on investing. My goal was to use Twitter to build my network within the industry and learn from others. I wrote a thread explaining my thesis behind Livongo. I believe a few other big Fintwit accounts owned the name as well and retweeted my post and my Twitter account steadily grew.

It was around the same time that I was contacted by Michael Nowacki, who ran Luca Capital, an investment firm in Cleveland that also owned Livongo at the time and had read and liked my article. I took the bus down to Cleveland to meet with him and was introduced to his partner Joe Frankenfield, who ran the Saga Portfolio, which was run independently. It seemed like we hit it off and had similar ways of thinking about investing. We exchanged emails for a few more months and I eventually was given the opportunity to join them in August 2020. I was delighted at the opportunity to turn my passion into a full-time job.

In the eight months between the time of my article and when it was acquired by Teladoc Health, Livongo Health rose 576% and turned out better than I could have ever imagined. It’s impossible to know how well Livongo would have done as an independent company and its returns were certainly aided by the favorable environment for growth stocks, but I believe the combination of a low relative starting valuation and an underappreciated amount of visibility on its 2020 growth due to record bookings and a large Federal contract created a compelling set-up at the time regardless.

Developing my Process

Over the next year, my work gave me a lot of freedom to source my own ideas, but I also did research on companies that we owned, and this exposed me to a wide variety of industries. I continued to find success, owning concentrated positions in names such as CrowdStrike and The Trade Desk. I also was able to correctly identify risks in existing portfolio names at Luca such as GoHealth, Hims, Redfin, and Meta among others, which allowed us to exit at with profits/minimal losses compared to current prices. 

Throughout 2021, I was able to recognize some mistakes early and sell. The most notable of which was that my conviction in Teladoc was misplaced. After a dramatic deceleration in Livongo member-adds post-acquisition, expert calls highlighting some difficulties in past Teladoc acquisitions, as well as the sudden departure of the Livongo exec team, my conviction dwindled, and I eventually sold out in May 2021. My ability to rethink my original thesis led to Livongo still turning out to be a great investment for my personal portfolio as well as for my fund.

Going into 2022, I felt I had a good sense of what my process was and had a high level of conviction in my portfolio and the research that I had done. I like to look for unique, quality businesses that have strong customer value propositions, durable and growing moats, long runways for growth, strong/improving unit economics, shareholder-aligned management that thinks long-term, and are misunderstood in some way that I can disprove using data.  

Where I believe my process differs from many traditional value investors is that I place a larger emphasis on generating qualitative insights to inform my estimate of intrinsic value rather than investing off optically low multiples on historical earnings. While I absolutely need to find quantifiable data to support my thesis, I don’t believe relying on past earnings to be the best predictor of future value, which is also why I rarely use screens. For examples of great companies that have subsequently underperformed, look no further than the case studies in books like “Good to Great” or “Built to Last”. Business success is difficult, and often fleeting, and while having a long track record of profitable growth and intelligent capital allocation raises the base rates of continued success, even the most established companies are not guaranteed to continue to outperform.

Investing is all about forward-looking expectations, so I instead try to approach each potential investment with a fresh perspective, unclouded by prevailing narratives. I try to develop a thesis on why a company is positioned to win using first principles, often with the help of expert calls with former employees or competitors to dive into what makes a company tick. I find that this helps me avoid the “halo effect” where people may attempt to force-fit moats and other buzzword terms like “flywheel” or “vertical integration” to form a narrative about strong recent stock performance. I was not immune to this, and I found myself justifying Teladoc’s long-term durability, for example, based on variables that ultimately didn’t matter like a data advantage. Some qualitative factors that I look for when assessing whether a company has a moat include when companies have majority market share within their niche and high win-rates, when barriers to entry are high and industry market share shifts are rare, when a well-financed and larger competitor tries and fails to break into the sector, or when a company is able to raise prices without impacting churn.

When approaching valuation, I attempt to encapsulate a wide range of expectations based on my qualitative understanding of a company and its historical financials. I try to keep my assumptions realistic, but conservative, acknowledging base rates that indicate very few companies are able to maintain >20% growth over long periods of time. I also attempt to size the opportunity and predict a company’s market share trajectory over time. For companies that are unprofitable such as Carvana, I attempt to predict their margin trajectory by analyzing unit economics of past cohorts, breaking down the assumptions embedded in management’s target margin structure, and working backwards from profitable, established comps like CarMax keeping in mind my understanding of Carvana’s qualitative advantages.

At the end of this process, I would only invest if I could gain conviction in my ability to predict their earnings power 10+ years out. I attempt to think from the lens if this were a private business; would I be comfortable buying 100% of the business today and owning it forever for its future FCF. Finally, I would try to search for disconfirming evidence and would frequently put my ideas in public in order to get different perspectives. Naturally, very few companies make it past my filters, and I typically own a concentrated portfolio of 5 to 10 of my highest conviction names.

My Journey to Self-Improvement

While I continue to believe that forming verifiable theories about why a company will win ex-ante is the right way to approach business analysis, I overallocated to some early-stage companies because I believed I had an informational edge. In reality, even if you understand a company perfectly, there will always be things that will blindside you and this is especially true for early-stage companies where a lot can go wrong.

One example was my concentrated position in GoodRx, a prescription drug discount marketplace. When GoodRx reported their Q2’22 results, they disclosed that their largest pharmacy partner, Kroger, accounting for 25% of GoodRx’s revenue, stopped accepting GoodRx coupons. This came as a surprise as I believed, from GoodRx’s own website to calls with competitors and management, that pharmacies were forced to accept GoodRx through their PBM contracts. This development ultimately caused me to sell my position as it was inconsistent with my original thesis.

In doing a post-mortem, I discovered that a year ago, GoodRx had implemented a universal BIN (ID) at Kroger that would automatically select the lowest-cost pharmacy benefit manager (PBM) for a given drug. This made it easy for Kroger to shut down GoodRx directly. In addition, GoodRx’s largest PBM partners had negotiated very attractive fees and abnormally low drug prices with Kroger, whose prior management was willing to accept lower margins in return for higher foot traffic. This, in turn, caused GoodRx to direct a lot of traffic to Kroger and because Kroger was selling a dollar for 80 cents in many cases, Kroger was losing more and more money as a result. When prior management at Kroger had left last summer to join GoodRx and Kroger’s contract came due this year, new management took the opportunity to renegotiate the contract and stop accepting GoodRx in the process.

While Kroger later announced they would start accepting GoodRx coupons again, they would be doing so at a higher price, more in-line with the prices available at the other pharmacies. This clearly invalidated my original thesis that GoodRx prices would continue to beat competition as a function of its PBM marketplace, because it shows that GoodRx doesn’t have leverage over Kroger. Though GoodRx will likely continue doing fine as a business due to its brand and cash-generative ability, I misjudged the strength of its competitive position against pharmacies, and this opens up the opportunity for other pharmacies to renegotiate prices and take rates in the future.

I was aware that Kroger typically had the lowest pricing which was potentially unsustainable but the benefit to chain pharmacies via incremental foot traffic seemed clear as each customer would purchase an average of $40 of additional items on each visit. I was also under the impression that Kroger did not have much leverage, given they only had 3% market share of retail pharmacies, a far cry from the 25% share of GoodRx’s prescription revenue which was not previously disclosed.

From my original purchase price, I believed I could earn a 15% IRR if GoodRx managed to grow at a 30% CAGR over the next 5 years while maintaining FCF margins, at which point they would still have only captured 23% of what I estimated to be their serviceable addressable market in the US. With GoodRx growing revenues at a 50% CAGR between 2016 and 2021, and over 40% market share of the growing cash card market, this did not seem like an unreasonable assumption at the time. However, the combination of Kroger’s departure along with an increase in competition and their market being more saturated than I had originally expected contributed to a rapid decline in growth in 2022, meaning I overpaid.

My lesson is that I underweighted the risk that GoodRx would lose a key pharmacy partner, even though it seemed like a low-probability event based on all the data points. Going forward, I will be more cognizant of the risks of companies with concentrated client bases. I also should have been cautious about underwriting high growth for long periods of time due to the low base rates of that happening. Excessive concentration in early-stage companies can be dangerous, even if they are profitable and had a decently long track record like GoodRx.

Another example of a company where I bought too much too early on was Carvana. Carvana’s wipeout in 2022 was due to a range of factors, including management making an ill-timed purchase of the used car auction business ADESA funded through the issuance of high-interest debt, logistics issues, management expecting COVID/stimulus-fueled growth to continue, and used car industry volumes plunging to 2008 levels due to poor affordability from supply shortages and volatile interest rates. It was the perfect storm of both macro headwinds and over-optimism by the management team.

However, unlike GoodRx, I underestimated the risks and even added as the stock declined. I believed that if Carvana could simply make it past this bad macro, they would be positioned to take additional share from subscale online competitors and even CarMax due to their ADESA purchase. I rationalized management’s aggressive expansion as the model is viable only if they reach scale because of the significant fixed costs. But once they reach scale, I believed it would be incredibly difficult for anyone to replicate and they would achieve better unit economics than CarMax due to superior operating leverage and value proposition.

I was aware that they were highly exposed to macro with the additional leverage, but after doing liquidity projections across a range of scenarios, speaking with IR, and leveraging expert calls and alternative data to keep track of units, I believed the market was overestimating the odds of bankruptcy. During mid-2022, units were softening but still stable, and Carvana seemed to have a logical path to cutting costs. Carvana already proved its ability to scale opex by reaching target EBITDA margins in mature cohorts and proved during the pandemic that it could grow while cutting costs. It seemed as though the path to returning to prior SG&A levels was relatively straightforward once they shifted from all-out-growth mode to all-out-efficiency mode. However, while I expected units to decline, I didn’t expect units to fall quite as much as they did going into Q4’22. Carvana retail units declined 8% in the third quarter and are expected to decline in 25% YoY in Q4. Furthermore, while management did cut SG&A in Q2 and Q3, it seemed they were over-optimistic on a recovery in H2 as well and did not cut enough.

While I don’t believe annualizing Q4’22’s depressed quarterly unit sales to be appropriate given seasonality and improving affordability and management still has levers to pull such as a leaseback deal on the real estate and debt restructuring, the liquidity situation through 2024 is going to continue to be challenging. It is difficult to imagine how Carvana will survive without an improvement in industry volumes or very significant dilution. While the Garcias still have every incentive to recover their equity stake and alternative data indicates they’ve made significant progress cutting costs through Q4, I do believe the probability of bankruptcy has increased beyond what I would have estimated in Q3’22 and the thesis has devolved into a macro bet. Given the deteriorating unit sales, I sold a large portion of my commons to bring the allocation to a lower level where I’m comfortable with either outcome. I still maintain exposure to LEAPS due to the binary nature of the bet and will continue to monitor units and cost-cutting efforts over the coming quarters.

Overall, I made a risk management mistake by allocating as much as I did to Carvana and overestimating its ability to cut its burn rate when units are declining. A leveraged, cyclical, and unprofitable business is highly exposed to macro, and past a certain point, I was exposing myself to too much macro risk which I can’t predict. The long-term is just the aggregation of short-term periods and if a particularly volatile period like 2022 causes Carvana to fail, then that can change the long-term picture just as much as a shift in structural factors such as what happened with GoodRx.

My Learnings

I’ve been careful not to take away the wrong lessons from this experience because I know doing so can be more costly than the original mistake itself. A generation of investors were scarred by the Dot Com bust and chose not to invest in tech which was a mistake, and another were scarred by 2008 and focused too much on predicting macro and trading in and out of the market which was also a mistake.

However, I have identified several areas that I will adjust my process that I believe will allow me to do much better as an investor going forward. In summary, I remain confident in how I approach analyzing businesses, but I grew overconfident in my ability to forecast the future of early-stage businesses and I overpaid and overallocated as a result. I believe I would benefit from applying the same philosophy in analyzing profitable, slower-growing businesses where I have a greater margin of safety and balancing my long-term thesis on the business with a greater focus on the short-term direction of the stock. Along the same lines, I believe I would benefit from a much higher degree of diversification, enabling me to be less attached to any one position and better manage idiosyncratic risk.

1.     Sizing

My first lesson is on position sizing. Just because you understand something the best or have the highest conviction in doesn’t make it the best investment. People talk about how diversification for the sake of diversification can dilute returns but concentration for the sake of concentration is much more dangerous. Highly concentrated positions (15%+) should only be reserved for very rare low-risk, high-reward opportunities.

Scuttleblurb on Twitter mentioned that excessive due diligence and poor judgment are a deadly combo and I agree. I have often thought I knew something the market did not and gotten attached, doubling down when I should have acknowledged that there will always be things I don’t know that I don’t know and that the future is messy. I intend to keep my holdings at ~10-15 names going forward; enough to minimize idiosyncratic risk but not too much that I won’t be able to keep track of them. I have not gotten up to that point yet, but I have already raised cash and brought my concentration to a reasonable level by limiting speculative positions below 15% allocations and diversifying into LGI Homes, Floor and Decor, and Alphabet. I expect to continue to diversify my portfolio over the coming months and remain patient for attractive entry points.

In addition to stock price volatility, I would say I underestimated the amount of actual business volatility associated with early-stage, unprofitable growth names and especially the probability of left-tail risks actually playing out in a bad macro environment. Let’s say the risk of Carvana going bankrupt was 15%. Well, the true probability may indeed be 15% but they could still go bankrupt in a particularly adverse macro environment such as 2008. While the expected value of the bet may be very much in my favour, I simply don’t have enough shots on goal within a concentrated portfolio in order to have my outcomes match the true probabilities.

2.     Diversification

Second, I learned the importance of factor diversification. I did not pay as much attention to this before as I should have because I thought that over the long term the value of a business is its discounted future cash flows so I should buy the businesses trading at the largest discount relative to their estimated future cash flows, especially when I could handle the short-term volatility. However, being in the right sector at the right time can mean all the difference in terms of your short-term returns. It would have been easy to outperform the market in 2020 if you owned growth but difficult in 2022. Since I am primarily a bottom-up investor and still trying to learn top-down, it makes sense to diversify across factors to properly manage risk.

Going forward, I expect to increase my allocation to wide moat names with lower returns but a much narrower range of outcomes to balance out my allocation to earlier-stage names with higher returns but a much wider range of outcomes. Picking stocks is difficult enough as it is but investing in speculative companies is arguably the hardest area to be in as an investor because the range of outcomes is so wide, both in terms of the business and the stock, and this is reflected in their low base rates of success. I understood this fact well and actually viewed it as an opportunity because I believed that this uncertainty would create the most opportunities for mispricings for investors that were willing to do the work. However, by only focusing on companies with the potential for right tail outcomes that met my high IRR hurdles, I was also exposing myself to significant left tail risk that I underestimated at the time.

While I still intend to be a quality-first investor, I will place a higher emphasis on companies with attractive shareholder yields (dividends and buybacks) as it is much easier to predict than revenue growth and therefore, I will have a better margin of safety. In addition, I will have greater peace of mind during drawdowns, assuming the companies are self-funding, as they will be afforded the opportunity to buyback more shares, thus increasing the value of my stake. When a company has a long track record of profitable growth, a strong moat, and is distributing 10% of its market cap back to shareholders each year, it is difficult to lose 50% of your money on it permanently. The same cannot be said for a speculative growth company that is not profitable and is trading based on what it might earn 5-10 years down the road.

Though my circle of competence had been built up in tech names when I started, I continued to broaden my coverage over the course of 2022 to encompass more non-tech compounders like TransDigm, Copart, Constellation Software, Credit Acceptance, Blackstone, FleetCor, LGI Homes, Perimeter Solutions, Floor and Décor, and Brookfield Asset Management. I also keep a watchlist of over 220 names in my investable universe as part of my work and monitor valuations for opportunities, doing deep dives on the ones that I find the most interesting in terms of both the qualitative and quantitative criteria I mentioned previously. I’ve learned to reject companies quickly and I try to turn over as many rocks as possible in order find the best opportunities. I’ve also found that studying a wide variety of businesses helps me better put opportunities in context as it’s easy to miss the bigger picture when you only compare companies/strategies within the same industry.

3.     Selling

Finally, I want to do a better job of identifying my losers quickly and selling when warranted. Although I did do pre-mortems and would quantify my expectations via quarterly KPIs, I found myself often justifying misses by arguing that lower expectations were already priced in and that the stock had overreacted. Going forward, I aim to be more intellectually honest and sell when my thesis breaks regardless of price, being very careful about buying or averaging down when KPIs are headed in the wrong direction, and do a better job of assessing macro cycles when it comes to investing in cyclical companies.

I won’t be right at timing all the time, but I hope to make it a larger part of my process and balance my long-term views with short-term considerations, especially for companies that are dependent on capital markets for funding. Feedback cycles in investing are very long, so it can be easy to not realize you're wrong for a very long time, especially when you engage in long-term investing. However, the liquidity of the public markets can be an advantage. I can always sell when in doubt and revisit when I have a clear head because it's so easy to continually shift the goalposts and cling to a broken thesis.

I will also try to develop my ability to short stocks to be able to see things from a different perspective. Since shorting has the potential for unlimited losses, understanding the direction of sentiment, managing risk, and being able to time moves is critical to being a successful short seller. For example, while the liquidity analysis I did on Carvana after their Q1’22 earnings indicated that the chance of bankruptcy was low at the time, a short seller may have recognized that the combination of deteriorating fundamentals and a high debt load going to a potential recession would likely cause the market to overweight the probability of failure, and sentiment is the primary driver of near-term stock movement for a stock like Carvana with no valuation support.

Overall, by focusing on controlling downside risk through selling at the right times and not losing too much money in bear markets, I stand a better chance of earning higher returns over the long-term.

For more information on my background and qualifications, please check out my LinkedIn.