Tag: saas

Initiated a position in Fastly

Initiated a position in Fastly

Initiated a stake in Fastly recently to make myself follow through on deep diving into this company.

So what piqued my interest with Fastly?

They reported a strong set of Q2 results, with revenue rising 61.7% y-o-y. It was also the first time their EPS became positive, at $0.02 per share. Yet its share price took a hit and declined 33%. Largely due to Donald Trump threatening to ban Tik Tok, which accounts for 12% of Fastly’s revenue.

Even if Tik Tok were to be banned, I see this as a temporary blip to Fastly’s prospect. Fastly mainly serve enterprise clients. Many of the leading digital companies chose Fastly’s edge computing capabilities due to its capabilities, speed and reliability.

Source: Fastly’s presentation

It is worth noting that they have recently bagged Amazon and Shopify as their clients. With both of them enjoying the tailwind of e-commerce acceleration due to COVID-19, Fastly would stand to benefit as well.

Without further ado, and without being overly technical, let’s jump into what Fastly does!

What problems does Fastly address?

Traditionally, when we listen to music on Spotify or stream videos on Netflix, our request would first travel all the way back to the data center (or the cloud). It would be processed at the cloud before the content (along with recommendations for our next song or show) is delivered back to our devices.

This would usually not be a problem. But with increasing devices and applications moving to the cloud, shared bandwidth becomes a problem.

And this is what happens:

Netflix Loading GIF - Netflix Loading Buffering - Discover & Share ...

I hope you did not stare at the above gif for too long. I’m just trying to illustrate that high data usage (i.e. increased load) would increase the latency (i.e. response time).

And we know how detrimental this would be for user’s digital experience.

To give an idea of how much data is running on our networks:

Given how much data is running through our networks, the huge load and distance between user and cloud data centers are bound to increase latency.

Any lag would be detrimental to time-sensitive operations such as self-driving cars, smart hospitals, etc (and of course also to us streaming on Spotify and Netflix).

This is where Fastly’s edge computing capabilities comes in — to solve the latency issue. By bringing data processing closer to the edge of the network where the raw data is generated.

Edge computing sits closer to the users. For a Netflix user in New York, rather than having the request traveling to a remote data center (e.g. AWS data center), getting processed together with millions of other users’ requests. It gets processed at an edge computing server locally, referred to as Points of Presence (PoPs), which is closer to the user.

Apart from ensuring a fast and smooth streaming, edge computing is smart and is able to ensure that the right type of content is served each time by allowing for more on-device computing and analytics. The cloud is able to do these as well. But with millions of requests going to the cloud at the same time will drastically increase latency.

By being strategically distributed in city hubs, edge computing reduces latency by reducing the distance traveled and by distributing the load each server has to process.

Edge computing also helps reduces cost as cloud (e.g. AWS and Azure) is on a pay as you use model. By using edge computing, it helps reduces cloud usage and thus bring down costs.

An example would be the use of security cameras that automatically uploads the videos onto the cloud. Edge computing is smart, users can program it to upload only videos with moving motions onto the cloud for further processing. As opposed to uploading the entire 4k video onto the cloud.

Hence, Fastly essentially solves 2 problems — reducing latency and bringing down costs for its customers.

Developers focused

For enterprise software providers, winning over developers’ hearts is critical.

With a developer-focused usage-based sales model, companies like Datadog, Twilio, and Fastly empower the developers, who are the closest to the products being built to make the purchasing decision.

Fastly is developer-centric with a strong appeal to the developer community because of their advantages in performance, security and programmability.

This is opposed to the traditional model where a single buyer sign a huge contract on behalf of the entire company. An individual developer can proceed to Fastly’s website, register and starting building on their new idea in less than 30 minutes.

Large and growing TAM

Source: Fastly presentation

According to research firms, MarketsandMarkets and IDC, content delivery network (CDN) & Streaming and App Services & Security at the Edge are estimated to be a $35.4 billion by 2022 and is expected to continue growing at a high clip.

Fastly’s trailing-twelve-month (TTM) revenue is currently $845 million. They are in a good position to continue eating up market share as they continue to innovate and rapidly roll out new product offerings in edge computing.

Metrics to keep an eye on

Revenue: As with any growth companies trading at high valuations, it is important that they are able to sustain a high growth rate. Due to their strong Q2 revenue growth of 61.7% y-o-y, management has guided for Q3 revenue growth of 49.6% y-o-y.

Dollar-Based Net Expansion Rate (DBNER): This represents how much existing customers spent today compared to previous years. With a focus on large companies as opposed to their competitors (e.g. Cloudflare who focuses on SMBs), customer curation allowed Fastly to achieve the best-in-class DBNER.

DBNER was up 133% in Q1 and 132% in Q2, meaning their customers increased their spending by by 33% and 32% respectively y-o-y. As a rule-of-thumb, anything above 120% is excellent.

Number of customers & average spend per customer: We want to see that Fastly is able to keep grabbing market share. Total customers count increased to 1,951 in Q2, representing a 20% annualized growth.

Enterprise customers increased to 304 in Q2, from 262 the previous year. Enterprise customers are defined as accounts which spend in excess of $100,000 over the year.

Average enterprise customer spend grew 29%, from $556,000 in Q2 2019 to $716,000 in Q2 2020.

Gross margin: Gross margin was 56.7% in Q1 2020 and jumped to 61.7%. Management attributed this to improved efficiency in the operating model. As they continue to introduce high-margin products and scale, management expects gross margins to improve to 70%.

Why companies are investing heavily in Cloud

Ultimately, it boils down to its value proposition — creating a win-win situation for all parties. The opportunities here are huge due to the cost savings and increased capabilities it can provide.

Final words

Fastly is still a young company at its early innings with its moat still being developed. Compared to my other investments which form the core of my portfolio, Fastly is expected to experience higher volatility (which is not necessarily a bad thing).

Traditional metrics such as P/E ratio would not work for reasons covered in my previous posts. You may check them out below:

In my future posts, I will be covering Fastly’s risk, management, competitors, and valuation. Subscribe to the newsletter to follow along!

So far I have been able to publish at least 1 article per week and hopefully that has been value-adding to you!

I aim to publish 2 articles over the next month to spend more time catching up my backlog of reading and to do a deep-dive into other companies that have piqued my interests.

Disclaimer: This is not a recommendation to buy or to sell securities. Please conduct your own due diligence or consult your advisor. I merely write to improve my investment and thought process.

How to analyze SaaS companies?

How to analyze SaaS companies?

Software as a Service (SaaS) companies share prices skyrocketed since March 2020 as the market recognized that COVID-19 has caused rapid digitalization. SaaS companies also have a fantastic characteristic – recurring revenue.

Understanding its moat

Despite most of the SaaS companies commanding lofty valuations, not all of them will meet the high expectations set out by investors. We have to first ask ourselves if this company has a moat – e.g. high switching costs.

Is it easy for competitors to woo their customers over?

Consider Adobe and Zoom, both offering their software services for a subscription fee.

Most designers will find it impossible to switch Adobe due to the amount of time spent learning the software. They’re taught this in school and would be very resistant to master another designing software. Doing this would be akin to having to pick up another language for work.

Zoom’s success has been dependent on its user interface — its ease of use. You don’t have to set up an account and it is not cluttered with buttons. However, ease of use is a double edge sword. Users can easily switch away when Zoom stops innovating and delivering added-value to fend off competition.

Comparing Adobe and Zoom, Adobe has a much stronger moat (i.e. switching cost). Its users are locked in and thus Adobe is able to extract value for a longer period (higher lifetime value).

This is important as SaaS companies spend huge capital outlay (i.e. sales and marketing expenses) in acquiring customers. With distribution costs near zero, the optimal strategy is to acquire as many customers as possible.

Read our earlier articles on marketing and R&D expenses to understand why SaaS companies appear unprofitable at first!

Its returns are very dependent on its ability to retain customers. Bringing us to our first metric – churn rate.

Churn Rate

Churn rate measures the percentage of customers who canceled their subscription over the period. The best way to think about this is how much gross profits were lost as a result of customers canceling their subscription?

The churn rate provides us insights into how sticky the software is and it is a key determinant of profitability for the company.

Lower churn rates equate to longer customer lifespan. This directly translates to higher lifetime value (LTV) of each customer.

The longer a customer stays subscribed, the more value may be extracted from the initial capital outlay – customer acquisition cost (CAC).

Churn rate may be expressed in the formula below:

For example, you have 100 users each paying $100 per month for your software. Taking the typical gross profit margin of 80% for most SaaS companies (GPM ranges between 75% – 80% for mature SaaS companies).

This gives us a gross profit of $8,000 per year (100 x $100 x 80%).

Say 20 users dropped out that year. Resulting in a loss of $2,000 in gross profits ($100 x 20). The churn rate would come up to 25% ($200/$800).

Customer Lifespan is simply the inverse of churn rate. With a 25% churn rate, the projected customer lifetime is 4 years (inverse of 25%).

Understanding the churn rate will allow us to know the Lifetime Value (LTV) of a customer.

Lifetime Value

LTV is how much I can expect to receive from each customer on average over the lifespan of their subscription.

LTV may be expressed in the formula below:

Knowing the LTV gives me an idea of whether the customer acquisition cost (CAC) provided good returns.

It helps answer the question – Did acquiring these customers bring in more profits than what it cost?

Customer Acquisition Cost

CAC is the total sales and marketing cost to acquire each customer. This includes items such as salaries of sales staff, advertisement expenses, and everything else it took to successfully onboard the customer.

CAC may be expressed in the formula below:

Let’s say it costs $1,000 in a year to acquire 100 new paying customers. Our CAC will be $10.

In other words, it costs, on average $10 to bring in one new customer!

By knowing the LTV and CAC, we can understand if the company can make more profit from a customer than the total cost of acquiring them.

Unit economics

Unit economics is simply an evaluation of profitability on a per-customer basis. Copying Buffett’s euphemism, CAC is what you pay and LTV is what you get. SaaS companies are ‘profitable’ if LTV exceeds CAC.

A useful way to evaluate this is to look at the LTV to CAC ratio:

Ideally, this ratio should be between 3 and 5. The company’s customers are providing profits three to five times the cost to acquire them.

If it is below 1, that would mean that the cost spent to acquire a customer is higher than their lifetime value. Some companies start off with below 1 to get a network effect going (e.g. Grab and Uber offering massive discounts initially). However, this is ominous if it persists over a prolonged period.

If it is above 5, the company could be missing out on valuable opportunities. Customers are bringing in huge returns and companies may want to spend more on sales and marketing to quickly scale the business.

Update: Insightful discussion on LTV-CAC ratio with the team from Moneywisesmart below:

The timing of the LTV would affect the Internal Rate of Return (IRR) of customers (time value of money). Also, shorter payback times are advantageous as less working capital is required. This in turn provides companies the ability to scale faster.


It is important to understand the LTV vs CAC trade-off when evaluating businesses. Especially for SaaS companies, this metric must be closely studied to understand its profitability.

Bluegrass Capital put it across succinctly, “If I can find a company that invests $1 in acquiring a customer and they can get back $3 to $5 in gross profits, it doesn’t matter to me if the EBITDA margin is negative or positive.”

For many of the SaaS companies today, they are spending heavily today to scale. These companies will not appear profitable today but would be highly profitable in the future, provided they have a strong moat and sound unit economics.

To end off, I will re-share this gem from Marcelo Lima’s 2Q 2018 letter:

In our future article, we will go through a case study on evaluating a SaaS company.

If you wish to follow along, do subscribe to our newsletter or like our Facebook page for updates!