Intro to Software Business Valuation

Abstract 🔗

Distinguishing the quality of software business is an important skill for investors, engineers, and executives. As software gets more critical in the industry, value investors cannot ignore this intangible sector anymore and should establish valuation models vastly different from the physical world. Likewise, engineers need to understand how their contributions are tied to business value to be impactful. Of course, valuation is always art rather than science, and you need both quantitative and qualitative judgment, but this post suggests basic aspects that can help build your mental model.

1. Intro 🔗

The list of the most valuable companies clearly shows the importance of software: Apple ($2.2T), Microsoft ($1.7T), Google ($1.3T), Amazon ($1.1T), Tesla ($640B), and so on. Yet, despite the importance, the market is still unsure how to value software business. As a result, in 2022, we are witnessing a drastic change from unreasonable optimism to pessimistic hopelessness (again).

Fundamentally, the software business is “scary” as those companies own very few physical components you can point at. For example, GitLab reached a $15B valuation after IPO, but their S-1 didn’t even have an address because the company is remote-only. The company has negligible assets besides cash: property and equipment are only $5M. It was never profitable with an accumulated deficit >$.6B. You may want to say its code is the most valuable asset, but even that is open to the public. It is hard to predict how this specific business will prosper, but those intangible businesses will surely represent more market value in the next decade (unless nuclear wars push humanity back to prehistory). So then, how should we reason about such businesses?

To give us some calm, it is worth noting that the larger transition toward an intangible economy has been ongoing much longer – ever since the Bretton Woods system ended in 1971, even the mighty US dollar had no physical backing. So, instead of being scared of new dynamics, it’d be wiser to evolve our mental models to embrace changes. The model should help us to navigate through dramatic changes with a long-term focus and not be distracted by continuous noises. Over a decade, I’ve been iterating with my mental model of software business to make critical decisions like investing, immigrating, switching jobs, and designing software solutions. The model is rudimentary but can still provide useful insights, so here we go.

2. Market History 🔗

Strategy is a sequence of 1) interpreting the past, 2) predicting the future, and 3) executing now. So, let’s start with a bit of the stock price history before building a model. Even though the price can fluctuate widely far from the value, it still provides a valuable view because the past is “fixed” and doesn’t require complex interpretation.

At the dawn of the internet business around 2000, the market was overly excited about it and pumped valuations to the sky. Many turned out to be unsustainable and phased out (e.g., pets.com), but even solid businesses were dramatically affected. For example, Amazon’s stock price dropped over 95% from $113 to $5.51 (pre-split), which would’ve been an outstanding stock for the next two decades. Cisco is another intriguing example as it became the most valuable company with a $546B valuation but never recovered the peak until now.

Since 2012, when Facebook IPOed with a $100B valuation and the explosive growth of smartphones became more apparent, the market started to appreciate the power of the software business again. The trend continued until the end of 2021, and the market became highly optimistic – it was a noticeable year with a very high number of IPOs, where 80% of the business was unprofitable¹. Still, it is unfair to say that the market was “unreasonable” with today’s data. It is easy to conclude that the numbers were indicative of significant loss, but the strategy of ignoring 2021 probably would’ve missed the following opportunities as well (and way beyond that):

• Amazon in early 2010: Lehman Brothers collapsed about a year ago, and there were substantial financial uncertainties like the European debt crisis. The financial risk seemed particularly relevant as Amazon was aggressive in international expansion. Besides, the company has been investing heavily in a web service business that seemed unrelated to its core retail business. Still, the stock almost tripled in a year, making PER > 70. It turned out 2010 was the last year you could buy AMZN stock at that price range.
• Tesla in the summer of 2013: Its market value increased by 4x in a few months and was already comparable to other major automobile manufacturers like Ford. It had been only a year since Tesla started to sell their first major model (model S, which was still very expensive), and everything was questionable, including market demand, supply capabilities, product quality, etc. In retrospect, there was never a good time to own Tesla stocks solely based on financials without excitedly embracing its potential to scale revenue with a fixed cost. This intangible value was much more essential than Tesla’s revenues or book values, which were always marginal compared to the market cap.

Stepping back, because of the intangible aspects of software, we can probably find real examples to justify any predictions of financial statements. For instance, if I were to write the same article last year to be cautious of optimism, I would’ve used examples of Groupon, Yelp, or LendingClub. So, quantitative analysis isn’t enough to evaluate software businesses; we need to dig deeper to understand dynamics and plan for the future. Indeed, a successful investment was never solely about reading quantitative metrics, even in Benjamin Graham’s era of the 1930s. So, we’d rather accept valuation as an everlasting challenge that we will never fully master.

Investment students need only two well-taught courses - How to Value a Business, and How to Think About Market Prices.

– Warren Buffett (1996 Letters to Berkshire Shareholders)

3. Simplest Valuation 🔗

It is easy to forget, but business value is eventually tied to its capability to generate profit. However, even a simple business can have multiple variables affecting economic dynamics and profitability. For example, let’s say your grandma wanted to sell socks.

1. She had a great vision to sell socks at a high margin with automation. So, she convinced her friends to invest $100k in her company at the exchange of 40% ownership.
2. She immediately converted 100% of the cash to invent a machine that knits socks in her backyard. The machine needs basic fabrics and electricity that cost $5 each, and the product sells at $15.
3. The machine can make 10k socks each year with no errors. The demand is strong, and the goods always sell out ($150k yearly revenue).
4. She pays various fees to handle sales that cost $10k a year ($60k total cost of goods sold and 60% gross margin).
5. Her vision was to make the operation high margin with low cost. So, she’s the only employee to admin all the processes by working a few hours a week and paying herself $30k per year ($60k operating income).
6. The business pays $10k in taxes, leaving $50k as the annual net income.
7. She gives out all the profit to shareholders (60% to herself and 40% to external investors). As a result, investors get $20k each year for a $100k investment.

The outcome is excellent for investors, steadily generating 20% of their investment yearly. If those investors managed to sell their positions at the same initial price after the 5th year, they would’ve more than doubled their assets in 5 years ($20k each year would generate some additional income). Given a solid track record, how much would you pay for the 40% of ownership? The original $100k? How about $200k, which would still generate 10% of your investment yearly (and maybe you can sell out even higher in the future if interest rates are low)? How about 100% of the ownership? Thinking more, it is fascinating that the original price ($100k) is irrelevant, and the current price is all about the future.

This question is the same for public stock investors. Even when we see the fixed past, the future is full of possibilities and unknown unknowns. What if a neighbor sells similar socks at a lower price, and the company needs to lower the price from $15 to $12? What if the socks become out of fashion and the company needs to develop a new style? What if fabric prices increase by $2 due to international conflicts? What if the machine breaks down in an earthquake and takes three months to fix? What if the government enacts a new law requiring more documents for socks sales, doubling the sales fee? What if grandma decides to retire, and a new employee demands $60k, not $30k? Those minor fluctuations can make the $20k flow evaporate instantly. In the opposite direction, grandma may decide to scale the business by borrowing $300k at 5% interest rates and increase the capacity by 4x. A portion of the extra earnings needs to pay back principal and interest, but the net earnings can increase significantly.

4. New Dynamics 🔗

Now, let’s see how software can provide vastly different economic options. Instead of a socks machine, let’s say grandma implemented a program that returns if a company’s stock goes up or down tomorrow with high accuracy. She hosts her program in Cloud and charges $0.1 per request.

4.1. Elastic distribution 🔗

Software is incredibly powerful in delivering service to customers. If grandma wanted to sell socks across different states, the logistics would’ve been costly and time-consuming. Thus, she would need to forecast demand carefully and increase CapEx (e.g., expenses for factories or warehouses) at the “right” pace. If too fast, financing risks may destroy the business, or there can be quality concerns. If too slow, that market can be taken by others and is much harder to penetrate. Cloud enables the global and instant distribution of services with no quality concerns.

• Pokemon Go was widely popular at launch, and the traffic was 50x larger than the company’s expectation. Even though the estimation was way off, the underlying infrastructure scaled quickly to serve customers, and the company made over $500M that year.
• Apple and Tesla sell only a few products at a high margin, and they become much more attractive with accompanying software. Customers pay for hardware, which has lower friction than intangible goods, and the software continuously improves for more customization (e.g., App store) and enhancement (e.g., better self-driving). Those companies can even expand way beyond their initial domains without disrupting fixed hardware (e.g., Apple to finance and Tesla to insurance).

4.2. Focused cost 🔗

The cost of serving one customer isn’t too different from serving 10k customers. As a result, Software-as-a-Service (SaaS) companies often have a gross margin of >50% (e.g., GitLab had 88%). Of course, to serve globally with low latency, grandma needs to evolve the architecture (e.g., CDN). Still, the cost can be skewed to high-yield R&D expenses, not ephemeral per consumption, and compounded for the long runway.

As categorized in <Strategy Architecture>, only some actions add value, while others exist to stay afloat. When executed well, software businesses can direct their expenses toward the value-addition – mediocre managements often spend resources that seem related but not core competency (distraction is easy, but focus is hard). For example, grandma can focus on making the prediction algorithm better; as time passes, others will find it harder to catch up. With Cloud and SaaS, grandma has more options to outsource non-essential parts rigorously (e.g., DevOps to GitLab, identity to Okta, payment to Stripe, etc.).

• Stack Overflow and WhatsApp served millions of customers with a small team. For that, they had a simplified but broadly compelling product that needed little operation complexities.
• Google invested heavily in its core technologies while keeping its interface simple (Google search looks identical to the early days). With its focus, Google could rise as the outstanding winner even though there were already established companies (e.g., Yahoo, Lycos). Moreover, the capabilities provided substantial advantages to expand the company swiftly in the wild internet (e.g., Gmail, YouTube).

4.3. Competitions 🔗

The software provides powerful capabilities to everyone everywhere. Thus, anyone from any background can disrupt grandma’s business from diverse directions. As a result, disruptive innovations arise faster, but not all structures provide a long runway to be relevant in the market. Ironically, the best way to win is to focus on customers, not competition. Of course, she needs to be alert to see other ways to serve customers better (even if that’s very different from how her product runs), but she shouldn’t chase competitions because she will never run out of players. Similarly, distraction is a more significant threat than the competition as it will prohibit the business from growing exponentially with a compounding impact.

• Amazon has been relentlessly focusing on customers from day one and ruthlessly cannibalized itself before others. For example, it introduced Amazon Prime and Kindle to serve customers better even when they were costly in the short term, making other competitors irrelevant.
• Facebook provided great ways for people to connect online. Google Plus was irrelevant even when a stronger player backed it because it didn’t exist for customers. However, a 13-employee startup Instagram was a bigger threat as it provided other compelling ways to connect people.

4.4. Vulnerabilities 🔗

Competitions aren’t the only risks for business. While software businesses can worry less about earthquakes, new concerns like cyber-security or critical bugs arise. For the same reasons software is so powerful, vulnerabilities are hard to combat, and their power grows as the company scales. Another hidden vulnerability is productivity. Even when visible products look identical, internal architecture or culture can yield very different productivities per employee.

This topic is particularly subjective because those vulnerabilities are hard to measure and take a long time to improve. A common mistake is to think everything is measurable and to rely exclusively on a top-down approach. Instead, companies should accept that there’s no silver bullet and empower employees to contribute. While top-down can be effective at reacting to incidents, bottom-up can be more effective in proactively identifying possible problems and tackling the most impactful issues. For example, removing a feature can be much more valuable to the business than adding one. Unless companies appreciate such important activities (e.g., employees can get promotions by removing code), vulnerabilities can turn into a catastrophic time bomb.

• Knight Capital Group was the largest U.S. equity trader in 2012 but dissolved quickly after losing $440M in 45 minutes. The main problem was that obsolete code got unintentionally triggered. Adding smarter algorithms to earn $1M more would’ve been rewarded, but probably no one cared to remove outdated logic that cost the business dearly.
• A widely popular open-source Log4j had a severe security vulnerability that could hurt any business immediately. It was a scary event as many expected open source to be more secure with more eyes to verify. The event showed that to truly enjoy the benefits of free open source, the dev team should still keep their system decomposable and updatable, which isn’t free and often unappreciated.

5. Old Wisdom and New Practice 🔗

The power of intangible business isn’t new wisdom. Companies like Coca-Cola, Disney, and JP Morgan established great businesses by leveraging the intangible value of recipes, content, or credit. Still, the software allows intangible assets to be much more actionable, providing a new ground to grow a dreamy business.

A dreamy business offering has at least four characteristics. Customers love it, it can grow to very large size, it has strong return on capital, and it’s durable in time – with the potential to endure for decades.

Jeff Bezos (2014 Letters to Amazon Shareholders)

The wisdom is timeless but challenging to realize and maintain. As time passes, even the most successful software businesses get bulkier, losing the benefits of nimbleness and flexibility. Like entropy increases naturally, those mighty companies find it easier to expand than focusing their cost structure on their core competency, gradually turning mediocre. ∎

Be formless, shapeless, like water. You put water into a cup, it becomes the cup. You put water into a bottle, it becomes the bottle. You put it into a teapot, it becomes the teapot. Now water can flow or it can crash.

– Bruce Lee