Lyncredible Navigating the tech stack of engineering and management

Capital efficiency

The decision to work for a startup1 can be stressful. Everyone wants to get rich, but it is an impossible task to accurately predict the trajectory of any given company. When I consult my mentors and friends, most of the advice I get is a bit handwavy, as is usually the case. I have spent quite some time thinking about this topic. Here is one important lesson I have learned: examine its capital efficiency when deciding to join a startup.

What is capital efficiency

Capital efficiency, as its name suggests, measures how efficiently a business uses capital. Loosely speaking, capital efficiency is defined as the ratio of money made over money spent during a certain period. It is obviously better to deploy $1 and make $10 in return, than to deploy $1 and make $2. In this context, the definition of capital is limited to tangible assets like cash, while intangible assets like human creativity and brand recognition are excluded.

There are much more formal definitions for capital efficiency. The formal definitions are not required to read this post, but I do recommend the following articles to learn more:

Why I care about capital efficiency

The tech industry is defined by high-growth companies, and I certainly would like to work for a company that has a shot to become the next Amazon or Microsoft. This was why I joined Uber in 2015, although it did not go the way I had hoped. There are many explanations for why Uber has not achieved the high expectations2. I don’t think there is a single correct reason. However, I do believe that Uber’s poor capital efficiency was a significant factor.

My argument is as follows. The high valuations of tech companies are betted on zero marginal cost. Such companies can spend most of their resources on upfront fixed-cost investments i.e., building great products. They can scale to an almost unlimited pool of customers rapidly and inexpensively. A poor capital efficiency, sustained for a long period, suggests that the marginal cost is not zero, but rather very high. This forces the business to pay its attention to the uncapped marginal cost, while the fixed-cost portion that fuels product innovation takes a backseat. That would be unfortunate for tech workers like me, whose skill set is concentrated in building products. Such skill set would not be as impactful and would not be valued as much in a business with high marginal cost.

We can turn our attention to the one of largest tech companies for example: Microsoft. The business model of pre-Internet Microsoft was primarily selling packaged software. The marginal cost of producing a DVD copy of Windows is almost zero, yet Microsoft could charge hundreds or thousands of dollars for a single copy. The profit from the high-margin packaged software business allowed Microsoft to invest massively and expand to many business lines. It is important to note that such investment was concentrated in the fixed cost of creating new products, rather than the marginal cost of replicating existing products. They started by selling BASIC interpreters and gradually moved to operating systems (MSDOS/Windows), productivity tools (Office), databases (SQL Server) and developer tools (Visual Studio), to name a few. Microsoft did all of this with no outside funding3.

On the other hand, capital inefficient companies are forced to spend a huge amount of money to maintain and grow its existing business. For example, Uber and Lyft could not differentiate their product experience, but had to resort to price wars to take market share. The implication here is that the product is fungible, and the tech workforce building the product is also replaceable.

Capital efficiency for startups

Once I was convinced capital efficiency is an important metric, the next question was how to measure capital efficiency for startups. Recall the loose definition above: capital efficiency is the ratio of money made over money spent during a certain period. Startups are private companies and do not report earnings. How could I know how much money a startup spent or made in any given year?

The answer is of course a big NO without insider knowledge4, but there is a reasonably good approximation: the fund-raising dilution ratio(DR). Imagine a startup raises $M in funding at a pre-money valuation of $V. All existing shareholders before this funding round is diluted by M/V * 100%5, and that ratio is a proxy for the company’s capital efficiency.

Let’s look at a couple of hypothetical examples:

  • Imagine Company A raises $1B at a pre-money valuation of $10B. The DR is 1B/10B * 100% = 10%, which means that all existing shareholders are diluted by 10%6.
  • Imagine Company B raises $100M at a pre-money valuation of $10B. The DR is 100M/10B * 100% = 1%7.

Both companies are valued at $10B before their respective funding round, but I would argue that Company B likely has a higher capital efficiency than Company A, because Company B’s DR is lower. Why is that? A company’s valuation is a function of the amount of money it expects to make in the future, while the amount of money being raised is an approximation of the money it will spend. Put the two together and you have a reasonable proxy to the company’s capital efficiency.

We can also go one step further and measure the Venture Capital Efficiency Ratio(VCER), which is defined as, for any given round, the pre-money valuation over total capital raised before the round8. It is the same idea but measured over all previous funding rounds cumulatively. Imagine a company is valued at $10B pre-money for a new round, and it has raised a total of $2B over all previous funding rounds, then its VCER is 5x at that round.

Here is how I think about the two metrics:

  • DR is a forward-looking metric for a single round. It shows how much of the company ownership is given away to new investors in exchange for the money raised to fuel future growth. DR predicts capital efficiency in future business growth. Lower DR is a proxy to higher capital efficiency.
  • VCER is a backward-looking metric over all previous rounds. It shows how much money the company has raised and possibly spent to grow the business and reach the current valuation. VCER validates capital efficiency in previous business growth. Higher VCER translates to higher capital efficiency.

Real-world examples


Date Round Amount Valuation Total Raised DR VCER
Aug 2009 Seed 200.0K   200.0K    
Oct 2010 Angel 1.3M   1.5M    
Feb 2011 Series A 11.0M 49.0M 12.5M 22.45% 32.7x
Dec 2011 Series B 37.0M 300.0M 49.5M 12.33% 24.0x
Aug 2013 Series C 363.0M 3.5B 412.5M 10.37% 70.7x
Jun 2014 Series D 1.2B 17.0B 1.6B 7.06% 41.2x
Dec 2014 Series E 2.8B 40.0B 4.4B 7.00% 24.8x
Jul 2015 Series F 1.0B 50.0B 5.4B 2.00% 11.3x
Dec 2015 Growth Equity VC 5.6B 62.5B 11.0B 8.96% 11.5x
Jan 2018 Late VC 1.3B 69.0B 12.3B 1.88% 6.3x
Aug 2018 Late VC 500.0M 72.0B 12.8B 0.69% 5.8x
Apr 2019 Late VC 500.0M 78.8B 13.3B 0.63% 6.2x
May 2019 IPO 8.1B 74.3B 21.4B 10.90% 5.6x


Date Round Amount Valuation Total Raised DR VCER
Oct 2011 Seed 1.0M   1.0M    
Jan 2012 Seed 6.9M   7.9M    
Jul 2012 Series A 17.0M   24.9M    
May 2013 Series B 40.0M   64.9M    
Oct 2013 Series C 150.0M   214.9M    
Dec 2014 Series D 355.0M 4.6B 569.9M 7.72% 21.4x
Jun 2015 Series E 434.0M   1.0B    
Mar 2016 Series F 430.0M   1.4B    
Oct 2016 Series F 260.0M   1.7B    
Jul 2017 Series G 5.2B 20.0B 6.9B 25.80% 11.8x
Jul 2018 Late VC 500.0M   7.4B    
Nov 2018 Growth Equity VC 3.0B 42.0B 10.4B 7.14% 5.7x
Jan 2019 Series H 1.0B 47.0B 11.4B 2.13% 4.5x
Mar 2021 SPAC IPO 1.3B 9.0B 12.6B 14.26% 0.8x


Date Round Amount Valuation Total Raised DR VCER
Jun 2011 Seed 3.0M   3.0M    
Jan 2013 Series A 9.0M 24.0M 12.0M 37.50% 8.0x
Sep 2013 Series B 6.5M 48.6M 18.5M 13.37% 4.1x
Feb 2015 Series C 30.0M 200.0M 48.5M 15.00% 10.8x
Jan 2017 Series D 100.0M 1.0B 148.5M 10.00% 20.6x
Apr 2019 IPO 356.8M 8.8B 505.3M 4.03% 59.6x


Date Round Amount Valuation Total Raised DR VCER
Mar 2011 Seed 2.0M   2.0M    
Feb 2012 Series A 18.0M 100.0M 20.0M 18.00% 50.0x
Jul 2012 Series B 20.0M   40.0M    
Jan 2014 Series C 80.0M 1.8B 120.0M 4.44% 45.0x
Dec 2014 Series C 70.0M 3.5B 190.0M 2.00% 29.2x
Jul 2015 Series C 100.0M 4.9B 290.0M 2.04% 25.8x
Nov 2016 Series D 150.0M 9.2B 440.0M 1.63% 31.7x
Sep 2018 Series E 245.0M 20.0B 685.0M 1.23% 45.5x
Jan 2019 Series F 100.0M 22.5B 785.0M 0.44% 32.8x
Sep 2019 Series G 850.0M 35.3B 1.6B 2.41% 45.0x
Mar 2021 Series H 600.0M 95.0B 2.2B 0.63% 58.1x

A few notes regarding the data

  • Amount is the amount of money raised in the round.
  • Valuation is the pre-money valuation for the round.
  • DR is the dilution ratio for the round, i.e., the ratio of Amount over Valuation.
  • Total Raised is the cumulative total capital raised up until that round.
  • VCER is the Venture Capital Efficiency Ratio of Valuation for the round over Total Raised until the previous round.
  • All funding and valuation numbers are in US dollars.
  • The data is sourced from Dealroom and Crunchbase. Although I have worked at Uber and Stripe, I have only referenced public data when putting together this post. There is no guarantee if the data is accurate.
  • Only primary equity rounds are included. Debt rounds and M&A rounds are omitted. Secondary equity rounds are excluded because no new shares are issued in such rounds.
  • Extension rounds at the same valuation as the original rounds are combined into the original rounds to make DR calculation reasonable.
  • To simplify the result, I decided to omit data for Uber China’s funding rounds. As a result, Uber’s fund raising amount below is understated (because money raised for Uber China is not in the table), and hence Uber’s capital efficiency numbers are overstated.

My takeaways

  • It is typical for DR to start high - about 20% or higher for at least three of the four examples above. We don’t have valuation data for WeWork’s early rounds, but it is safe to assume the DR is also in that 20% range or higher. This makes sense because an early-stage company needs to make large upfront fixed-cost investments to create the product.
  • It is also typical for DR to trend lower in subsequent rounds. A low DR in these later rounds suggests that the company’s marginal cost to scale their business is low. That is a sign of a healthy business that is achieving economies of scale. The trend is most clear for Zoom’s data.
  • Conversely, a high DR in later rounds indicate high marginal cost. It implies problems in reaching economies of scale, as is shown in the trajectory of Uber and WeWork after the colossal rounds with billions of dollars raised. Note the absolute number of the DR can still be lower since the valuation is typically higher at these rounds. For example, Uber’s “Growth Equity VC” round is a good one to examine.
  • VCER is a more reliable metric over the long term. A VCER that consistently trends downwards is a sign of a business that has trouble reaching economies of scale, as is illustrated in Uber. A VCER that goes below 1x basically means earlier investors are close to being wiped out, which is what happened to WeWork. On the contrary, a VCER that steadily trends upwards indicates high capital efficiency as the business achieves economies of scale, as is the case of Zoom.
  • For both metrics, it is useful to examine both the raw number and the trend over time.

At the end, I want to note these capital efficiency metrics should be used as references to help understand startups. They are useful but do not tell the entire story. There is no replacement for analyzing the business, the people and the product. Personally I had lots of fun tinkering with these numbers and forming the idea in this post. As a parting gift, I would like to leave you with the IPO-round numbers for some of the most successful companies in the tech industry:

  Microsoft Amazon Google Facebook
IPO Date Mar 1986 May 1997 Aug 2004 May 2012
Total raised pre-IPO 1.0M 8.0M 36.1M 2.3B
Pre-money valuation 716.0M 384.0M 21.4B 88.0B
Amount raised in IPO 61.0M 54.0M 1.7B 16.0B
Post-money valuation 777.0M 438.0M 23.1B 104.0B
IPO DR 8.52% 14.06% 7.77% 18.18%
IPO VCER 716.0x 48.0x 593.7x 38.7x
  1. I am using a very loose definition of “startups” in this post: any VC-funded private company that aspires to grow rapidly is a startup. Many such companies should no longer be called “startups” in later stages. 

  2. Here is my explanation for Uber’s underachievement: Uber had no upside

  3. Microsoft grew its business with no outside funding before the IPO. They raised $1 million to get a venture capital company onboard for “some adult advice”, but never spent the money. 

  4. Early stage startups are usually very open to share financial metrics with prospective candidates. Make sure to ask for such metrics if you are interviewing with one. 

  5. The definition of the dilution ratio is adapted from chemistry. According to Wikipedia, the dilution ratio is ratio of solute to solvent. Technically the pre-money valuation should be the solute, because existing shareholders are being diluted, not the new money raised. However, I prefer to treat the new funding as the solute, and the pre-money valuation as the solvent. This way the dilution ratio is usually a lower number, as is shown in the examples that follow. 

  6. If the pre-money valuation is the solute and the new money raised is the solvent, Company A’s dilution ratio will be 10:1. 

  7. If the pre-money valuation is the solute and the new money raised is the solvent, Company B’s dilution ratio will be 100:1. 

  8. The proper definition of VCER uses Enterprise Value, which is again not possible to get without insider knowledge, and hence I use the pre-money valuation instead. 

Share via