Fastly - Bull v. Bear

Great article. The ideas on secure sandboxing appear like a whole new way of approaching security. There is a video link about it at the end of the article.

People should watch the video and judge for themselves after hearing her thoughts on FSLY. I sold FSLY after watching it.

I watched the video and was not impressed, it sounded like someone who does not really understand what they are talking about.

I’ve been in tech my entire career and my take is the CDN category has always been a low margin area and has a limited TAM. No matter how good the tech is you can’t increase the TAM. Yes, FSLY gets a bump from the pandemic and that may go on for a while. Long term it’s a tough road. I also believe that this is an area that the big cloud player like AWS won’t allow a third party to dominate – if FSLY tech is that good. Which remains to be seen.

We don’t need third party opinions because we can and do experience the technology every time we download a CDN delivered web page. Non professionals might not know what they are seeing but web developers quickly figure out that the CDNs are not performing well enough. This was not visible or apparent to me before the pandemic but with the increased internet traffic generated by video conferencing it is clearly visible. I’m sorry to say that TMF’s CDN provider is pretty crappy which is why we get so many misses.

This has enormous consequences for investors (if my diagnosis is correct). Fastly created a new CDN architecture that is supposedly more efficient than legacy CDNs. If that is the case and high volume internet traffic is here to stay, then there will be a lot of demand for Fastly at the expense of other CDN providers. Sooner or later they will catch on but will be beset by the “Innovator’s Dilemma.” They would have to reconfigure their networks, hardware and software, an expensive proposition that incumbents don’t like to do. But even if they do it, Fastly has a one or two year window of opportunity.

The above is only addressing the CDN part of their business. IoT needs Edge Computing which only adds to the internet traffic volume. Just what “edge” means is not set in stone. Having been in IT since 1960 I can attest that as technologies evolved (computer and communications) there has been a seesaw from core to edge and back. The current shift is from core to edge. Anything that an IoT device does on it’s own is not, in my estimation as an investor, edge but “local.” “Edge” is the internet node closest to “local.”

This battle reminds me of Apple taking over smartphones from Nokia and Blackberry. Same service, much improved user interface. Will the Nokias and Blackberrys of CDN suffer the same fate?

I have the feeling that the debate over Fastly is ignoring the IoT/Edge bull part of the story. One Fool commented that he sees no new opportunities because all the value has already been discounted by high valuations, a reference, I believe to be to the year 2000. Is 2020 a replay of 2000? I don’t think so, it’s not in the charts

S&P 500 40 year chart: https://invest.kleinnet.com/bmw1/stats40/^GSPC.html

Saul’s 200+% YTD gain is even steeper that the 1995-2000 run but it is only in a very narrow slice of the market, about a dozen or so stocks. Any of them could easily give back 50% on a bad report, a bear raid, or profit taking (which Saul confessed to ;). No growth stock is immune to that eventuality. The question is how to deal with it. My solution is to have a year’s worth spending money in cash and about 1/3 of my securities in covered call income. This way I can let the other 2/3 ride the crazy Wall Street Rollercoaster.


This morning I read some interesting news, interesting not for the human interest side but for the investing implications

The coronavirus pandemic is expanding California’s digital divide
Kevin Frazier
TechCrunch - July 9, 2020

If every California student without an adequate internet connection got together and formed a state, it would contain more residents than Idaho or Hawaii.

A total of 1,529,000 K-12 students in California don’t have the connectivity required for adequate distance learning.

Analysis from Common Sense Media also revealed that students lacking adequate connection commonly lack an adequate device as well. The homework gap that separates those with strong connections from those on the wrong side of the digital divide will become a homework chasm without drastic and immediate intervention.

https://finance.yahoo.com/news/coronavirus-pandemic-expandin…

It’s not just business and social gatherings but the very huge education sector that is driving Internet traffic volume that the current infrastructure is having a hard time coping with. This is a huge boost for providers like Zoom and Fastly that have the better technology in place now.

Denny Schlesinger

61 Likes

I have been reading the Fastly threads, the SoftwareStackInvesting article and Beth Kindig’s interview. We have two overlapping issues that both determine Fastly’s future. One is their superiority as a CDN provider. The other is CDN as a technology versus more advanced edge computing. This post started as notes to myself. If it helps someone else, great. If you already know all this, sorry for the interrupt.

The SoftwareStackInvesting article
https://softwarestackinvesting.com/fastly-edge-compute-expla…

describes Fastly’s technical advantages as a content delivery network (CDN). A CDN is a way to park content close to the consumer, such that it minimizes the time after the consumer clicks on something that the something arrives on their screen. It stores copies of frequently requested content close to the consumer and provides fast pipes for the rest of the content to come from the “origin” server. So as a CDN, they are out ahead.

The article goes on to say that Fastly is building a truer edge computing platform that allows some of the “origin server” computation and decision-making to happen on Fastly equipment (closer to the consumer), rather than at the origin server. They aren’t just storing content, they are storing software that makes decisions about the request and the response. Decisions made there will optimize the actual request made to the origin server(s) and how the content gets back to the consumer.

I am inventing an example: Your kid wants to watch Donald Duck. An edge computing platform receives your home’s Disney login and launches Disney software to parse the request.
“Subscriber ID? Check. Password? Check. Request? Donald Duck.
Decision: Cartoons are stored in Burbank Disney servers. Send the request there and establish a connection to the subscriber for the video stream.”

Later, when the kid is asleep, you want to watch the Cubs win the 2016 World Series again. Your request is parsed and sent to the ESPN portion of Disney in another location.

This isn’t just caching Disney content closer to the consumer like CDN’s do. This is also caching Disney software that decides what happens with the request for content. It bypasses some of the back-and-forth between customer and multiple origin servers that might otherwise have to talk to each other before your request can be filled. The cached software makes decisions about your request, then routes content to you based on the decisions.

Ms. Kindig’s discussion of Fastly starts around the 27 minute mark:
https://seekingalpha.com/article/4355130-cloud-and-zoom-shif…

She assumes that the listener understands the difference between CDN’s and “edge computing”. Hopefully I just explained that. Kindig says that Fastly is primarily a CDN. CDN’s are a low margin business. They got a boost from the increased streaming demand after the pandemic hit like Zoom did. She questions if CDN’s will still do well when the pandemic and streaming demands recede. That implies a risk to Fastly because they are primarily in the CDN business.

Kindig mentions Fastly’s newer “edge computing” platform that is also discussed in the SoftwareStackInvesting article. Says it is in beta testing and she would like to hear from someone who the beta customers are. She thinks Equanix, Amazon Web Services and Microsoft Azure will dominate edge computing and not leave much for anyone else.

I am not offering an opinion on Fastly’s future. What I did here was organize my notes of what I read and heard in a way that was helpful to me, and hopefully to others.

Thanks and regards to all.

34 Likes

And aside on this discussion…Can you comment on FSTLY vs NET? Fastly does look to provide higher end products (not sure how it can be characterized as antiquated) vs Cloudfare’s more ‘popular’ offerings. Is that only down to 2 different sectors and market?

He linked to his portfolio where he listed his positions. Fastly was 24% of his portfolio while his next largest were 17%, 16%, and 12%. Based on that, I’d say that in spite of his cautions, Fastly is by far his highest conviction position.

Best,

Clearly.

When I read the article and last checked the link his portfolio showed FSLY as his largest position but only at 10%. That would have been on July 2 evidently a lifetime ago.

Cheers

draj

I watched the video and was not impressed, it sounded like someone who does not really understand what they are talking about.

Yeah, as I said a couple weeks ago: https://discussion.fool.com/that-was-a-really-frustrating-listen…

I have the feeling that the debate over Fastly is ignoring the IoT/Edge bull part of the story.

It does seem that the bears are either ignorant of or are discounting Fastly’s upcoming Compute@Edge product, which is in beta now, with at least Shopify’s beta use case being promoted recently.

OTOH, there are at least a few bulls that read Compute@Edge’s 35 micro-second process start-up and equate that to their CDN technology. FWIW, Fastly maintains a live network maps with real-time performance stats: https://www.fastly.com/network-map What I’ve seen are over 15M requests per second, cache hit ratios near 90%, and “hit times” (time to server a piece of content under 1/8 of a milli-second (that’s 1/8th of 1/1000th of a second, or under 125 micro-seconds).

If I’m trying to find real bear cases for Fastly, I could make these points:

  1. Best technology doesn’t always win in the market. VHS beat Beta, Windows beat MacOS, etc.
  2. No CDN is going to win everything. Companies with content want multiple CDNs to protect them from any individual CDN going down and taking everything they have down. This happened with Cloudflare last year and it wasn’t pretty.
  3. Fastly’s CDN is not the best at everything and not everywhere. As seen in Fastly’s network map, Fastly doesn’t have close servers for every country in the world. It’s security features, while they work well, do not offer the most functionality.
  4. Compute@Edge is in Beta and no-one knows how popular it will be. It won’t be generally available until 2021, so no revenue from it until then. Even if popular, you can bet Microsoft and Amazon won’t stand still - Fastly is playing in big boy territory here.
21 Likes

On Fastly vs CDN debate, to me its focused on developer community - allowing developers to write custom code, integrate that with Fastly infrastructure - was a big big tie breaker.

On this board, we have seen few astounding winners with this strategy… TWLO is on top of the list…

If you read Peter Offringa’s article on Fastly, he clarifies why developers prefer Fastly which not only allows them to customize, but do so very easily and delivers much better performance due to its ultra-fast server start technology
(I am not a techie… so may be paraphrasing something here based on what I understood)…

Shopify and other use cases articulated by Fastly and explained by Peter clearly says that this is not vanilla CDN… it is already, true edge service platform with high degree of programmability…

those (including Beth) saying this is old CDN haven’t take time to understand FSLY (btw, I was in the same boat two months back, so I do relate to the thought process)…

And FSLY management clearly identified that they are premier pricing, they don’t really get the volume based, price sensitive CDN business and their expectations are for gross margins to further improve all the way into 70% range… which is clearly not a CDN play…

Can others catch up and beat FSLY, yes, but it can take a long time and even then, with all the custom code written on FSLY infrastructure, their existing client work loads likely sustained… cost of switching increases with time… that is not vanilla CDN…

I see FSLY as super leveraged play on e-commerce, WFH, online gaming, 5G and digital transformation… If there is a pure play company that benefits from each one of these, it is FSLY… I wouldn’t want to get scared out of CDN comparison at this stage…

nilvest

PS: I have ~4% position in FSLY… and looking to build larger position, though it is hard for me to take money out of my larger positions like LVGO, CRWD and DDOG to put into FSLY!!

34 Likes

“I’ve been in tech my entire career and my take is the CDN category has always been a low margin area and has a limited TAM. No matter how good the tech is you can’t increase the TAM.”

You have me on technological competence, but my edge is that i know that this last phrase is not correct.

I have spent the last 31 years of my life studying leadership. What I know is that companies grow or shrink to the size of their leadership, almost regardless of products and markets. If a leader of a $1 billion company has the vision breadth, depth, executive skill and passion to lead a $100 billion, he will find where the TAM is and bring his company along.

Consider an online bookstore company in the late 1990s with thousands of upstart competitors doing online sales. Some asked, how big can Amazon become by selling books? A few knew he might eventually sell other online retail products. No one saw that Jeff Bezos could expand that company into AWS, kindle, prime, and much more, and become a multi trillion dollar company.

It’s not sufficient to have a wonderful idea or a wonderful product. Google, Microsoft, Amazon, Walmart, Schwab, Netflix, etc would have achieved little without genius at its helm.

This is not the place for a detailed discussion on this subject but i’ll leave with a quote and a link. Someone here mentioned they hold ADI stock. Vincent Roche, the capable CEO of that company a few years ago said, “History is littered with the corpses of brilliant ideas that never found… (its way to market).” This discussion is at 3:30 in the link below

https://www.youtube.com/watch?v=81y3E45YcEs

I haven’t looked closely enough at the new FSLY CEO, but if he has the ability and passion to run a company 10X its current size and complexity, the TAM is out there somewhere.

Among our Saul stocks, my favorites almost since they became public, and led by visionaries and executive leaders big enough to run bigger companies, to include AYX, LVGO, WIX, ROKU, RDFN, SE, SQ, PYPL, PEGA, SNAP, ADBE, CRM. I rely on Bert, Saul, and a few of the brilliant technology people here for understanding of products and potential competition and select companies with the strongest leadership.

Sometimes i’m wrong about the stocks. Sometimes i’m wrong about the leadership, usually not about about their genius, but once in a while about their lack of shareholder friendly fiduciary values, executive acumen, or emotional stability. I look at it as closely as i can. Sometimes i get it wrong and sell as soon as i become aware. Of course there are other factors that can affect a stock such as geopolitical issues. Not all leaders with the requisite potential succeed. But no company without the requisite potential at the top, can grow beyond leadership vision/talent.

JMHO

68 Likes

Draj,

If you scroll down to the bottom of the following SSI web page, you will find FSLY is Peter’s largest position to date at 24%.

https://softwarestackinvesting.com/software-stock-recommenda…

Zoro

2 Likes

I have spent the last 31 years of my life studying leadership. What I know is that companies grow or shrink to the size of their leadership, almost regardless of products and markets. If a leader of a $1 billion company has the vision breadth, depth, executive skill and passion to lead a $100 billion, he will find where the TAM is and bring his company along.

Years ago, one of my courses at Harvard Business School focused on venture capitalists and the criteria they used to evaluate startups. After studying case after case, it became apparent that one of the most important factors in gaining access to venture capital was the quality of the management team. The VC’s looked for strong leadership and a seasoned, compatible management team. The company’s Big Product Idea seemed to carry some weight, but it ranked fourth or fifth among the VC’s evaluation criteria.

As a technical person with a freshly minted PhD in Physics, that made no sense to me. I thought that surely the Big Product Idea was the key to success, and brilliant technical people like me were the superstars that made it all happen!

One day after class I approached the professor to try to make sense of it all. He basically explained that when it came to high-tech startups, one good idea does not guarantee success. Even if it’s a great idea, a strong company needs to have a continuous pipeline of great ideas, not just one. He also pointed out that there are plenty of brilliant technical people available, but relatively few brilliant leaders. That’s why a great CEO will eventually grow and move on from one successful venture to another. The VC’s know who they are and they track them carefully. It’s fairly common for a venture firm to recognize a potentially great company that lacks key leadership, and match that company with the right individual(s) to create a solid management team. In fact, that matchmaking may be the most important role that a venture capital firm provides, even more important than providing seed capital.

As a scientist, I can’t help but perform a deep dive assessment of the underlying technology when evaluating a company to invest in. But I’ve learned that the company’s success is ultimately far more dependent on the vision and execution of the management team. That’s why I shake my head at some of the postings on the Fool boards when we endlessly try to evaluate every nuance of a company’s technology and make endless comparisons of the most miniscule technical features to their competitors. Unless you’re a technologist who works directly for the company or their competitor, you will never know enough to make the kind of technology assessments that I see on these boards.

However, you can do something much more important. You can watch videos of the management team, listen carefully to their earnings calls, and gauge the team’s passion, vision, clarity, honesty, and overall leadership. That will tell you a great deal about their chances for future success. That’s what the venture capitalists do, and you can learn the same critical evaluation skills.

Now go back and read addedupon’s post #69370 six more times :slight_smile:

83 Likes

FSLY is Peter’s largest position to date at 24%.

https://softwarestackinvesting.com/software-stock-recommenda…

Correct. Thank you.

draj

He linked to his portfolio where he listed his positions. Fastly was 24% of his portfolio while his next largest were 17%, 16%, and 12%. Based on that, I’d say that in spite of his cautions, Fastly is by far his highest conviction position.

Clearly. When I read the article and last checked the link his portfolio showed FSLY as his largest position but only at 10%. That would have been on July 2 evidently a lifetime ago. Cheers draj

Hi Draj, I just looked at his Software Stack Investing website again and Fastly IS definitely currently listed as 24% of his portfolio. Perhaps when you looked at it, he hadn’t gotten around to updating his portfolio on the web site in a month or two. As you say, in this market that’s a lifetime. :grinning:

Best,

Saul

3 Likes

I have spent the last 31 years of my life studying leadership. What I know is that companies grow or shrink to the size of their leadership, almost regardless of products and markets. If a leader of a $1 billion company has the vision breadth, depth, executive skill and passion to lead a $100 billion, he will find where the TAM is and bring his company along.

I don’t have the investing acumen of many on this board. I do agree that a great management team is crucial, but I don’t agree that identifying them is as easy as you make it sound.

I work in real estate, and had a tenant who was a former fund manager, and now works managing the finances of a few very rich families.

I asked him how to you spot great management, and he said that it is almost impossible. He told me a story of someone who was a great salesman, and convinced fund managers his company was great, and it turned out he was committing fraud.

You look at Enron, and Bernie Madoff and you can see how hard it is.

Now they are saying that Jack Welch did a bad job, by turning GE into a huge bank.

3 Likes

If you scroll down to the bottom of the following SSI web page, you will find FSLY is Peter’s largest position to date at 24%.

https://softwarestackinvesting.com/software-stock-recommenda…

Replying to the above - FWIW I checked his holdings within the last 10 days and he had 38% FSLY at the time (unless he or I made a typo). So, he may have trimmed a little after the big run.

I personally did not sell any FSLY outright, but I did sell some August 21 $140 covered calls for $6.35 each when the stock was over $100. My thinking was I wanted to lock in a little gain after the big run, and if it does actually get to $140 by August 21 I’m pretty sure I would be selling some anyway at that point. I saw it as a can’t lose opportunity. My cost basis is $41, so I was able to get about 15% of my cost basis out and still leave myself the potential for 40% upside in 6 weeks time. I only did this on 2/3 of my shares so that I still have some if it somehow blasts past $140 after earnings.

1 Like

Just a few thoughts.

FSLY does not have an earth breaking technology invention. What FSLY has are a bunch of significant, smart incremental CDN improvements. Most of the time, this is how invention works. Even Einstein put together pieces he discovered from others.

I just bought FSLY last week and yes I know it is a little late, but better late than never. I needed the technical understanding. I have a problem sometimes as an investor when I think I know the technology so well and poo poo some brazen whippersnapper up and comer. I helped develop one of the first CDNs back in 2001 at Cacheflow. I held off purchasing CRWD as well. I worked cyber security the last 10 years at FireEye before I retired this year. So I have a solid knowledge of malware detection. And I helped develop the Qualys cloud 15 years ago before the term ‘cloud’ was created. Again I did not buy CRWD until recently. sigh…(no worries I have killed it with others. You can’t buy them all you know)

So what I have learned is to deeply understand the technology and the business model. But most importantly recognize the market movers (oomph factor) and do a second deeper dive when I overlook.

Let’s be real, FSLY will hit some significant competition at some point. Akamai, AMZN, and others will not just sit by for long and watch customers go to FSLY. But FSLY has a moat for some time, maybe 1-2 years, enough time to get established and realize some significant growth. (This is what we did at FireEye in 2012-15). The FSLY BGP BIRD with Silverton distributed routing agent running on the low cost Arista Networks switch is not something easily or quickly replicated. And the API user controls are a very sticky and a desirable feature for customers. FSLY is smart with their fewer but heavy POPs and SSD based hosts, but this can be replicated pretty easily I think by competitors. (probably already has been)

I think folks need to realize that not all web hosting companies or applications need accelerated content delivery. The TAM may not be as big as we think. Many times customers simply need what is ‘fast enough’ and this is where pricing pressures come into play. CRWD will see this same effect. Mid tier, low security maturity customers have limited security budgets and many times they do not buy best in breed. They buy what is good enough for what they can afford.

-zane
long a little FSLY and CRWD

46 Likes

Let’s be real, FSLY will hit some significant competition at some point. Akamai, AMZN, and others will not just sit by for long and watch customers go to FSLY. But FSLY has a moat for some time, maybe 1-2 years, enough time to get established and realize some significant growth

One of the features described in recently published explanations of FSLY operations was that they did not use off the shelf equipment for their CDN system but created new software and operation logic,.As well they have lowered latency time to microseconds. My impression is that this gives FSLY a headstart with accelerating velocity vs competition. A recent SSI article says for example,

This software-driven design has paid dividends many times over, giving the team the flexibility to evolve their network design as content delivery use cases change.

I surmise it takes more than a year or two for a company to redesign the system, backtrack, redeploy , get up to speed, catch up and compete.

You seem to be saying it can be done in that time frame. Could you elaborate.?

Cheers

draj

2 Likes

FSLY seems to have a head-start in content retrieval speeds and delivery. Having software control over the POP switches as well as the hosts gives FSLY maximum flexibility to adapt. Make no mistake, they are still adapting to their customers needs. FSLY is not done.

From what I read FSLY is using COTS hardware from Arista Networks and customizing. They could be developing their own line interface cards and custom ASICs for this switch but I have not read this. Pass me a link on their hardware if you have it. Also they are using custom configured BGP edge router software. BIRD BGP is an open-source implementation for routing Internet Protocol packets on Unix-like operating systems. So most of the proprietary software seems to be in the switch and the hosts.

On the host since they say they are using their own file system for fast caching, they have customized the OS. (Let’s hope they did not write their own OS like Cacheflow did in 1998. This gets expensive to maintain.)

If I were a major competitor, I would raise my game to offset the FSLY strengths. Add SSD storage resources to my high volume POPs is easy. And assess their routing and storage/retrieval technologies and see if I can do something similar with my architecture.

I am making a lot of guesses here on the FSLY architecture and on how long it will take for a committed competitor to catch up. My 1-2 years is a simple wild ass guess (SWAG). I would be watching Akamai and Limelight. And I would watch the pricing as competition strengthens.

-zane

9 Likes

Add SSD storage resources to my high volume POPs is easy. And assess their routing and storage/retrieval technologies and see if I can do something similar with my architecture.

Adding SSD storage is likely easy, but custom storage around it is likely not easy and some head start. From the software stack investing article:

They also wrote a custom storage engine for their POP servers to bypass the file system and squeeze every last drop of performance out of the SSDs, cramming as much data on them as possible. They employ various algorithms to keep commonly used data in the 384 GB of RAM, making it even faster. For some assets, such as “Like” or “Share” buttons that never change and are requested millions of times a second, they go even further by serving them out of the L3 cache directly from the processor.

Fastly advantage seems to be as Jeff Bezos once said: “We don’t have one big advantage, so we have to weave together a rope of small advantages.”

My feeling is in the end engineering talent matters. Fastly likely over time will evolve into enablers of edge technology removing barriers and bottleneck in the adoption of edge paradigm.

5 Likes

I have a problem sometimes as an investor when I think I know the technology so well and poo poo some brazen whippersnapper up and comer.

That statement is why I have some resistance to believing totally what a “expert” says and I make attempts to understand things myself using multiple opinions. Someone who holds themselves out as “experts” that has worked in the industry now or in the past do not necessarily have “correct” opinions on all topics in that industry, all the time.

There are some problems 100% believing industry insiders:

  1. Usually people working for a company “drink the Koolaide” or 100% believe their company line on how the industry is supposed to play out and will pooh-pooh automatically differing ideas that might ultimately work out. Sometimes, a competitor’s vision might play out and not the company vision where the “expert” works, which could take a industry in a totally different direction than where the “expert” believes.

  2. Sometimes, depending on what job a person works at, a persons job might be so in the weeds that the person has little comprehension of the “big picture” view. For instance, there is one person that I have seen that when talking about software companies seems to believe that the only people qualified to talk about the software industry are people that can write code. That person seems to have little comprehension that the ability to write code might not exactly qualify the person to speak on many different aspects of the industry/company.

For example, I am a ship engineer. I might be great at changing a piston in a diesel engine but that might not necessarily make me a expert on building a whole ship or running a shipping company.

This also works in the opposite way too. A CEO of a shipping company might not necessarily understand how to change a piston in a diesel engine or be a expert on building a ship. Generally, people that work in any industry develops specific expertise in a given area and it is extremely rare to find people that know everything from top to bottom and from side to side in a given industry. A person working might not have expertise in the things a person they work next to has expertise in.

  1. Especially, beware of the “expertise” originating from people that have worked in a industry and have left that industry because knowledge and ways to do things evolve so rapidly these days that someone that has left a industry in as little as one year might be woefully out of date on how things are done today in that industry today. New technology, new ways of doing things, new ideas, online courses, trade schools and on the job training might have made a retired “experts” knowledge obsolete.

Now, industry experts are not always wrong but I have found over time that industry “experts” are not always “right” in their observations or predictions either. I have learned to treat “expert” opinion as only one data point and I base decisions that I make on multiple views rather than only one “guru” view.

Starrob

8 Likes

You are so right starrob.

-zane