Fastly and Cloudfare

Investment Analysis Clubs Saul’s Investing Discussions
1

Fastly and Cloudfare are Content Delivery Network companies. Fastly used to be discussed here a lot more until the infamous pre-announcement and thestampede for the exits.

In fact, the discussions got downright heated, culminating in a mostly nasty thread beginning November 6th. Interestingly, since that day, Fastly’s performance is nearly double that of Cloudfare.

On a price per sales basis, Fastly is considerably cheaper but they have lower margins so maybe that’s why?

I continue to be long both companies and am immensely pleased with my results.

32 Likes
2

This thread certainly seems OT, but, I just wanted to go on record and say I sold out of FSLY Oct 15th (the date of pre-announcement referenced by Carver, if I’m not mistaken?) at an average of $91. I put all FSLY proceeds into NET at $57 the same day. The very short term changes since then:

FSLY 91 to 117: 28.5%

NET 57 to 93: 63%

I’m also pleased. But, as noted, it’s very short term.

23 Likes
3

The primary issues, as I remember’em, was the growth in customers and slow progress of Compute@Edge which is now a few years behind Net. Has this changed? What I think would be a great way for someone who is actively watching Fastly to turn this thread around would be to go back and summarize the issues people shared and give us a progress report :wink:

7 Likes
4

I just wanted to go on record and say I sold out of FSLY Oct 15th (the date of pre-announcement referenced by Carver, if I’m not mistaken?) at an average of $91. I put all FSLY proceeds into NET at $57 the same day. The very short term changes since then:

FSLY 91 to 117: 28.5%

NET 57 to 93: 63%

I did exactly the same and got the same prices. I am also quite happy with the results. The reasons people got out were that Fastly was gaining new customers at a tenth to a sixtieth of the rate that Cloudflare was gaining the. Do you know ANY other company we’d be interested in, and would you invest in it, that only grew enterprise customers by 2% in a quarter??? Also Fastly was still not in General Availability for Compute@Edge after two years, and Fastly’s one big quarter was proved to be a one-time fluke due to increased usage due to Covid. To me, investing in Fastly is investing on hope that they will do it, almost a turn-around stock. Cloudflare is already knocking the ball out of the park.

Saul

56 Likes
5

This graph was shared on Twitter by Matthew Prince, couple of days ago

https://w3techs.com/technologies/history_overview/proxy/all/…

Along with the comment

“I dislike the number 16 so can’t wait, assuming current trends continue, to get to 17% next quarter.”

If you actually review the graph and pay attention to his hidden message, you can draw couple of conclusions. Mine are:

  1. “None” from the graph most likely includes all apps and websites that are not protected by reverse proxy and can be vulnerable to network and other security attacks. It also means that the potential for Cloudflare is 5 times what they have captured so far, and that number of websites/apps will continue to expand of course

  2. Also means that their service is 16 times more popular than Amazon CloudFront and 20 than Fastly.

  3. But what I find more interesting here is the hidden message

0.6% increase for 1 month and 1 week, January through Feb 8th, vs 0.9% increase in 3 months prior

So the number of properties protected by CloudFlare is growing rapidly, and Mr. Prince is suggesting pace is not slowing down!

The numbers in the graph only relate to their reverse proxy market share expansion. That does not include up sale of their teams, or edge offering or other premium services. But the rate of expansion is huge, considering that the reverse proxy market is very competitive.

If you also look at Akamai’s earning report, their security offering has experienced over 30% growth - just a hint what we can expect from CloudFlare.

I hope that provides some clarity on the argument of why CloudFlare.

45 Likes
6

Interesting find - thanks, rdgyy.

Another way of looking at it, I suppose, would be to say Fastly has increased their penetration (if that’s the right term) by 266% versus 242% for Cloudfare.

I look forward to reviewing each of their upcoming quarterly reports.

1 Like
7

…their service is 16 times more popular than Amazon CloudFront and 20 than Fastly.

I think it’s worth pointing out that Cloudflare has a FREE tier for their CDN service, so it’s not surprising to see a much higher uptake for this service than companies that charge. Does anyone have the stats on the paid uptake of Cloudflare versus the competition in this arena?

“None” from the graph most likely includes all apps and websites that are not protected by reverse proxy and can be vulnerable to network and other security attacks.

Well, there are many other protection techniques, and even Cloudflare says that reverse proxies are used for performance and reliability reasons, not necessarily security. Indeed, while a reverse proxy can be of use to prevent DDOS attacks, it’s mostly used to improve performance, as in what CDNs mostly do (including central and edge caching).

It also means that the potential for Cloudflare is 5 times what they have captured so far, and that number of websites/apps will continue to expand of course

I would not claim a TAM for reverse proxy services equal to all the websites in existence. Not every webpage/website benefits appreciably from a reverse proxy.

19 Likes
8

Hi Smorgasbord1,

The numbers in the graph are for you to interpret the way you feel, this comes straight from their CEO. Considering that their strategy is to land and expand, and that their paid services start as low as $5/month for something like WP edge plugin, and end with Enterprise clients spending millions a year. What is important here is that they are landing aggressively (hence the 0.6% comment), and the expanding is something you can get from their earnings, compare the number of enterprise clients they are converting to the ones Fastly are converting.

CloudFlare was never designed to be a CDN, and really does not promote itself as such service. All of their security features, including WAF are designed around their reverse proxy. Every respectful organization/company/etc needs to deploy a zero trust in order to avoid ransomware, downtime, etc - the CloudFlare One service which is build on top of their reverse proxy + edge offering is the ultimate solution, and I don’t see Fastly or anyone else coming close to that.

5 Likes
9

CloudFlare was never designed to be a CDN, and really does not promote itself as such service.

No, they promote it right here: https://www.cloudflare.com/cdn , which contains:

Highly Customizable CDN
Cloudflare lets you apply advanced control over how content is cached on its network. Page Rules let you specify particular behavior for individual URLs, including what gets cached and for how long. The Cloudflare Cache API lets customers have more control over how content is cached, and improves cache hit ratios with advanced routing and URL canonicalization.

Increased bandwidth savings
Caching content on Cloudflare’s network reduces the number of requests to an origin by serving static content from a Cloudflare data center, minimizing bandwidth consumption.

Cloudflare offers predictable bandwidth pricing, with no surge pricing when you get attacked. Customers hosting their websites with Cloudflare partners can reduce or completely eliminate data egress costs from their hosting provider to Cloudflare when dynamic content is requested.

Granular visibility into your cache
Get actionable insights into the caching of your website for a better cache-hit ratio and further drive down your bandwidth costs.

The right CDN for video
Cloudflare’s global anycast network ensures fast video delivery, with shorter video startup times and reduced buffering, no matter where your visitors are located.

13 Likes
10

I agree with Smorgasbord1 here - no matter how much management tries to spin themselves during earning calls and investor presentations, their customers see them as primarily a CDN provider. I think their executive suite hopes to be valued at a higher level if investors see them as a multi service provider, which is what contributes to the conversation. They are being valued pretty fully already though, so it seems like it has worked. Maybe higher adoption of workers will bring their messaging in line with reality.

4 Likes
11

Here’s what I wrote in my own notes as to why I got out of Fastly:

“That 7000 new quarterly paying customers for Cloudflare is SIXTY-ONE TIMES AS MANY as the 114 total new customers that Fastly was bragging about as a record number of adds!!! How did we miss that number?

In Summary: Cloudflare came out with a staggering output of new technology products, while Fastly seems to be having some kind of existential crisis, has seen its growth rate fall by 20 points, is having some kind of mysterious problem with falling usage which it hasn’t elucidated, is two and a half years behind on Edge Compute and can’t seem to get to General Availability, is signing up less than 2% of the number of paying clients that Cloudflare is in a quarter. (Cloudflare signed up 61 TIMES more new paying clients, that means Fastly signed up less than 2% as many). Cloudflare also signed up 11 TIMES as many enterprise customers (over $100,000) in a quarter (80 compared to 7), etc. I personally got tired of making excuses for Fastly. ”

Now I just finished looking at Datadog’s results and conference call:

Datadog added 1100 paying customers in the quarter. Compare that 1100 to the 114 new customers that Fastly added. That’s what we should expect from one of our companies, not bragging about adding 114 customers.

Datadog also added about 400 customers over $100,000. That’s about 100 per quarter average. Compare that 100 to Fastly’s 7 or 9 per quarter.

Clearly there is something wrong at Fastly, no matter how good their tech is supposed to be. Sure they may eventually get it straightened out, but I’d prefer to invest based on execution now, not hope for the future.

Saul

61 Likes
12

I just started reading the Cloudflare Conference Call, and in the opening statement I read the following astounding figure:

Our paying customer account grew to over a 111,000 up 10% sequentially and our strongest quarterly growth in several years. Just because I couldn’t believe my eyes, I went back to the Sept quarter earnings report and sure enough:

We are just shy of 101,000 paying customers, and added 4800 this quarter [Sept 2020]

Do you realize what this says? They added 10,000 paying customers in the October-November-December quarter! Ten thousand! Compare this with Fastly’s 100 additional paying customers in a quarter [it’s 100 times as many], or even Datadog’s 1100 added paying customers. It’s an incredible figure.

As far as their over $100,000 customers they had this to say: Large customers, those that spend over $100,000 per year with us continue to be our strongest growth area adding 92 new customers in Q4 and bringing our total large customer account to 828. Revenue from these large customers increased sequentially to 49% up from 47% in Q3 as our sales team continues to close larger and larger enterprise accounts. [92 added to 736 means they increased those customers by 12.5% sequentially!]

Best,

Saul

82 Likes
13

Also, they only have 828 enterprise customers, versus 313 enterprise customers at Fastly, last quarter. 828/313 ~ 2.6

marketcap of NET/FSLY ~ 28/12 = 2.3

The other, additional 110,000 customers are a significant part of a relative undervaluation, provided they are profitable. Is their ‘land and expand’ profitable from the get-go?

14

Their presentation today speaks volume on this point

https://cloudflare.net/files/doc_financials/2020/q4/Cloudfla…

6 Likes
15

I love Cloudflare and am trying to stay disciplined by not oversizing my position. I think this is a longterm winner for sure…great founder-led management, amazing growth numbers, growing TAM, great execution and product offering, and part of an essential future trend (edge computing).

1 Like
16

Somebody asked me recently about what I thought of FSLY, and I summed up my points and thought it would be useful to share here as I essentially did a NET comparison.

Forgive the incomplete sentences :slight_smile:

Re/FSLY question from Lee. It’s hard to look at FSLY without naturally drawing comparisons to cloudflare. As they are both CDNs trying to moving computing (applications) to the edge and off of servers. They are really the only two in this area, although there are lots of other CDNs (Akamai and Zscaler are big ones), neither of them are focusing on edge computing.

So, in a world of limited resources I choose to compare the two and invest in the better.

Reasons:

Leadership: Bixby preannounced a downside revenue number in his second quarter as CEO, was flippant in the reasoning, and Prince has consistently done what Wall Street expects, beat and raise.

Product: FSLY’s serverless offering is two years behind NETs, and still in beta. NET just announced 50,000 new developers wrote their first code on their edge network in Q4–doubling from Q3. NET is absolutely dominating in developer awareness. Both are optimized for cloud, but FSLY is not optimized for developer customization. Both secure workloads, but FSLY is perimeter security, NET is zero trust (secures traffic even when inside the “firewall”).

Market: FSLY targets large enterprise with big data demands–their offering is cost/performance superior for video, but not much else. NET initially targeted SMBs but is swiftly moving upmarket. NET increased large customers (>$100k ARR) by 13% sequentially last Q, and added 10,000 new paying customers in the quarter (10% seq growth). This customer growth is almost 100x FSLY new customers (Q3 115 new customers or so). 100x!! I get it’s smaller customer size, but for a new product/technology domain, I want the company that has developer mindshare.

Price model: FSLY usage based, NET subscription based. So, harder to gauge FSLY revenue growth.

Fastly also publishes a very misleading dollar-based net expansion rate mostly to grab headlines, (its like 147%) that excludes customers that churn. Seems kinda fishy. They do publish a comparable net retention rate (121%) that is slightly better than NETs (119%). Just dislike metrics that are misleading.

Valuation: only advantage FSLY really has. Similar growth next year, about 40% each, although I expect FSLY to actually decelerate faster given usage-based model (less WFH), and NET to maintain. So forward P/S of 30x FSLY and 45x NET.

So, lots in there. FSLY not bad generally; if I owned 50 stocks I’d probably own FSLY, but owning only eight stocks, no room for them.

47 Likes
17

there are lots of other CDNs (Akamai and Zscaler are big ones)

ZScaler is not a CDN. It does not cache copies of client data for faster access. ZScaler is a pure security solution. I don’t even believe ZScaler uses a Reverse Proxy, but instead uses a Man In the Middle approach to redirect all traffic to its servers for validation.

FSLY’s serverless offering is two years behind NETs, and still in beta. … Both are optimized for cloud, but FSLY is not optimized for developer customization.

FSLY’s Compute@Edge is now out of Beta an in Limited Availability. Yes, Cloudflare’s Workers has a big head start in terms of availability.

Here’s a FSLY blog post: https://www.fastly.com/blog/cli-functionality-terraform-api-…

FSLY is very developer centric, and its startup time is the fastest bar none. Cloudflare’s Workers cleverly hides its cold start time inside the HTTPS (security) handshaking time, but I believe still only for top-level “root” domain pages.

What supporting evidence do you have for a claim that Workers offers more “developer customization?”

Both secure workloads, but FSLY is perimeter security, NET is zero trust (secures traffic even when inside the “firewall”).

Let’s be careful not to conflate Cloudflare’s product offerings. Cloudflare For Teams is the ZScaler competitor. Does anyone know if Cloudflare for Teams can be combined with its edge computing offering Cloudflare Workers? Remember that edge computing is designed around bringing compute resources closer to the end user, whereas the ZScaler/Teams style security utilizes a man-in-the-middle clearing house server for authentication/validation, which would seem to go directly against the benefits of edge computing.

Overall, we have to be cognizant that between these two companies there are at least 3 mostly separate offerings:

  1. CDN - Content Delivery Network (content caching)
  2. Edge Computing - (bringing compute closer to end users, via serverless APIs)
  3. Security - And there are various flavors of security. For instance, Cloudflare WAF is a solution that is similar to FSLY’s Signal Sciences acquisition, but Cloudflare For Teams is completely different.

There are ways in which these different offerings can be combined for certain applications, but also there are combinations that don’t make sense.

38 Likes
18

Hi Saul

Maybe not an apples-to-apples comparison re: paying customer adds.

FSLYs enterprise customers are worth nearly 3x ($216k/quarter) enterprise customers for CloudFlare ($74k/quarter), with enterprise customers making up 88% of FSLYs Q3 revenue, vs 49% of CloudFlares Q4 revenue.

Also interestingly, FSLY spent $75m in Non-GAAP TTM Sales and Marketing expense whereas NET spent about $200m, so again roughly 3x.

The other point-of-difference that I found interesting is that NET has much better gross margins, and so generates roughly 2x the Gross Profit $ than FSLY.

I thought those points were interesting :slight_smile:

cheers
Greg

24 Likes
19

Another customer comparison, I’m an individual customer of NET using their Warp 1.1.1.1. VPN. It cost $4.99/mo. for all my devices including my laptop. It works fine and sometimes it annoyingly slows down my speed. I leave it on 99% of the time even when home. I signed up early last year, cancelled it and re-upped when I started traveling again. I’m sure there are lots like me. No land and expand. Unless they up the fees, no additional revenue from my group of clients. If they do I suppose I’d look at alternatives.

I’m long NET and FSLY.

Mike

3 Likes