There have a been some posts recently discussing excessive valuation (which I agree is excessive) and the lack of revenue acceleration (which numerically is true) in relation to Cloudflare. I figured I would enumerate the reasons why I am not planning on selling a single share of Cloudflare, but rather looking for a point of further accumulation, in order to bring some non-UPST spice to the board.
Disclaimer: I am not an extremely knowledgeable techie like Muji but I can still follow the numbers and the narrative well enough to form a fairly defensible thesis.


"After a thorough review process, CISA has announced that they have selected Cloudflare and AFS to deliver a joint solution that can be used by departments and agencies of any size within the Federal Civilian Executive Branch."

From the Q2-2021 earnings call:
“We talked earlier about FedRAMP. Again, that’s very US-focused. We’re doing similar efforts, for example, with the German equivalent and others around the world. And again, I think that sets us up very well to be able to satisfy the regulatory and compliance needs of so many large enterprises. And it’s what we keep hearing as the real killer feature that distinguishes workers from any other computing platform.”

I am very excited about this new pipeline. Although they do have other customers within the government currently, I don’t get the impression that they are major contributors. This is a fork in the road that is creating a new path for them: enterprise and government agencies.

From the Q2-2021 earnings call:
“One of the things that’s important about us and one of the reasons why this is important is Cloudflare fundamentally is a foundation on top of which a number of other companies build their products. And so as other companies are looking for companies that they can trust, these sorts of wins we see are resonating across the entire industry.”

Referring to the US Gov win, they believe this is very important for the company’s credibility. Prince believes that this is the drop in the pond and the ripple effects will lead to more and more long-term deals, locally & internationally.


	        Q2-21	Q1-21	Q4-20	Q3-20	Q2-20	Q1-20	Q4-19	Q3-19	Q2-19
Q Revenue	$152.4	$138.1	$125.9	$114.2	$99.7	$91.3	$83.9	$74.2	$74.2
Paying Cust	126,735	119,206	111,183	100,968	96,178	89,223	84,154	80,986	77,626
>$100K ARR	1,088	945	828	736	637	556	526	451	387
D-Based NRR	124%	123%	119%	116%	115%	117%	119%	121%	122%
Revenue QoQ	10.4%	9.7%	10.2%	14.5%	9.2%	8.8%	13.2%	10.1%	9.2%
Revenue YoY	52.9%	51.3%	50.0%	54.0%	48.0%	48.0%	51.0%	48.0%	
Paying Cust QoQ	6.3%	7.2%	10.1%	5.0%	7.8%	6.0%	3.9%	4.3%	5.5%
Paying Cust YoY	31.8%	33.6%	32.1%	24.7%	23.9%	21.3%	15.6%	19.7%	
>$100K QoQ	15.1%	14.1%	12.5%	15.5%	14.6%	5.7%	16.6%	16.5%	15.2%
>$100K YoY	70.8%	70.0%	57.4%	63.2%	64.6%	65.5%	78.9%	72.8%	

What do we learn from the table above?

A) Revenue has actually been accelerating-ish since Q1-20. They mentioned in the Q3-20 earnings call that in Q3-20 they had “a one-time catch-up of $1.9 million related to a customer renewal. Excluding this non-recurring revenue, our year-over-year revenue growth would have been 52%.” I am not as concerned about the lack of accelerating revenue because a consistent 50% YoY growth in revenue is still impressive exponential growth; go plot revenue over time and you’ll see what I mean. If they beat guidance by their average beat (since Q4-19) of 5.6%, then they will have 53% revenue growth YoY. If you remove the one-time catch-up of $1.9 million in Q3-20, then that bumps up to 55.6% YoY. That makes me sufficiently happy as a shareholder.

B) DBNRR has steadily increased since Q2-20. Even at 124%, Matthew Prince is not satisfied. From the Q2-21 earnings call:
“I will say that there might be some noise in the dollar-based net retention number, where we definitely have got – worked hard to get it above 120%, and we think that that’s great. We’re not satisfied with where it is. There may be from quarter-to-quarter a little bit of noise in that metric. But we do think that over time, we can continue to improve that metric.”

C) YoY increases in Paying Customers continues to impress. Sure this last quarter went down but seeing as they were at 19.7% Q3-19, I would call that an acceleration. I believe part of the reason is word of mouth between enterprises as they begin to use their products. From the Q2-21 earnings call:
“A Fortune 500 auto parts manufacturing company faces similar threat and turned to us to solve it. They signed a $300,000 annual deal. The CISO said the Cloudflare team was fantastic. We took this threat very seriously. Cloudflare was there for us. You can’t make a car with only 99% of the parts. So, one malicious attack in the supply chain impacts everyone.”
“In that spirit, he referred us to another Fortune 500 manufacturing company to sign a two-year $320,000 contract. The CSO of that second firm referred us to a third Fortune 500 manufacturing firm who signed a two-year $570,000 contract. That’s over $1 million of new contracts because our solutions worked and solved some of the problems that are top of mind for every technology leader right now. It increasingly feels like Cloudflare is becoming the de facto choice for enterprises and they are thinking about the future of their network security.”

D) Customers spending >$100K had been slowing down since Q4-19 until the last two quarters. Hopefully this trend will continue as it clearly shows that either companies are upgrading at a rapid pace (should show up in increase DBNRR) or companies are signing on with larger contracts. I believe this is the latter and should continue as they land more enterprise customers.

One more thing I want to add here. I agree with GauchoRico (…) that I would prefer to see more positive movement in their profitability (FCF, operating margin, etc.) much like we see with Crowdstrike who pretty much has a straight line of increased profitability over the years. While this could become a concern if revenue were to decrease, I think it is just going to be a factor of being a Cloudflare shareholder for some time; there is no perfect company.

I’ll also mention here that they closed the Q2-2021 earnings call by mentioning that even at current prices for Workers, they see the margins as accretive. Flashing sign saying “bright future.”
Prince: “So again, I think that that – even if the prices that we charge for Workers today, it is very margin accretive for us. And so, again, I think it is a competitive product. We think that being very cost effective is important when you’re selling cloud products. We don’t think of ourselves as competing with other niche edge computing platforms. We think of ourselves as competing with the big public cloud platforms.”
Separate analyst than who asked the original question: “And then, just a quick follow-up on Workers, so even at its somewhat low scale today, this is a very margin accretive offering for you?”
Prince: “Yes. It’s – it is margin accretive even at 78% gross margin.”


I love how Cloudflare continually pumps out new products. This to me is why their free user base is a treasure trove. They can throw something out there, have the free users test it with no consequences, and either improve it if it sticks or scrap it if it doesn’t. I listened to an interview with Matthew Prince and Michelle Zatlyn (…) and I would say 1/4 of the interview was talking about how much scrutiny they use when hiring someone. They mentioned how a large part of their (Prince and Zatlyn) jobs is still recruiting. On top of recruiting, they are big on retention as well. I am a mechanical engineer working in the aerospace industry. Another engineer could certainly hire in and do my job but they would struggle for at least a year after being hired because on the job training is critical. This drives home the point to me that retention is huge and if Cloudflare can continue to succeed there, this will greatly contribute to them being the best.

Long quote from Q1-2021 earnings call:
“So, we have 90% of our R&D organization that’s really focused on how do we go deep on any of the products and features that we’ve launched. And they are always thinking about how do we continue to take feedback from customers whether they are current customers or prospects in order to build out the functionality across our platform, so that we can service whatever it is that they need.”
“10% of our team is really dedicated to an organization we call ETI, which is the Emerging Technology and Incubations team. And their job is to invent the future. And we don’t spend – they have incentives around thinking up what is the product or feature, which two years from now you guys are going to say, wow, that was a huge way that Cloudflare expanded TAM. They take lots of shots on goal. Not all of them are successful, but they are the team that came up with Cloudflare for Teams. They’re the team that came up with Cloudflare Workers. And so, they are often – when we do expand the TAM dramatically, they’re the team that is driving that.”

Random side note: during the Q2-2021 earnings call, one of the analysts slipped in a question/comment about Apple Private Relay. The CEO replied: “we have a policy of not talking about any of our customers without their permission. And I’m certainly not going to speculate here on that.”
I don’t believe anything here is formally announced (the analyst knew what he was doing) but there is unconfirmed reports that Cloudflare is one of the networks being used with this feature. This would be a pretty wonderful vote of confidence from a tech giant for little Cloudflare.


Some companies have mission statements because they have to. Cloudflare has a mission statement in order to live by it. Cloudflare’s mission is to “help build a better Internet.” Short, sweet, descriptive, all-encompassing for what they do, and most of all, leaves room for limitless paths (TAM) while still maintaining a sharp focus. It is the first line of their 10-K Business section, it is mentioned at least twice during each earnings call, it is upfront in their investor presentations…they actually mean it and I love it.

This doesn’t just apply to business. They recently announced Project Pangea to help deliver free Internet access to underserved communities around the world (…). Furthermore, they recently announced running their internet with zero carbon emissions and “will remove all historical carbon its global network has emitted since its founding by 2025 (….”) Like I said, all-encompassing.

#####(5) MOAT

I know some people might disagree with me but I view Cloudflare as having the largest moat in comparison to all of the other companies discussed here. Not only do they have the technological advantage (like most of our other companies discussed here) from their relentless R&D efforts and free users used as beta testers but they also have the infrastructure. From their most recent earnings presentation, they are in 200+ cities with 9,800 interconnects. That seems like quite the head start if someone wanted to compete head-to-head. This would be need an answer from someone more tech-savvy but I don’t believe anyone could do the type of security and DevOps that Cloudflare is doing if they used someone else’s network instead of building their own.

In terms of competition, see below.

From the Q1-2021 earnings call:
“We have a ton of respect for Fastly and Akamai and Limelight, but we’re in a very different business than they are. We are not usage based by and large, which last year actually felt kind of hard.”
“And so, I think that that just shows that it’s just a very different business. I actually looked up the data and Fastly is a – we just don’t see them in deals because, again, I think that they are very much going after the media space and doing media delivery, and that’s something that’s just very different than what we’ve ever intended to do. So, appreciate you recognizing the difference, but that’s not to say that they don’t have a great business. I think they do, it’s just not one that we really compete with.”

This was a watershed moment for me. I knew that Fastly was floundering and Cloudflare was accelerating, but why? It is because they are going after two different things. Fastly wants to be the fastest and best content delivery network. Cloudflare wants to build a better internet. For Prince to say “we just don’t see them in deals” isn’t a put down, I think he’s being honest here saying they just aren’t going after the same people. I keep hearing that the big cloud service providers are their competition (AWS, Azure, Google Cloud), but I think I hear them as competitors for every company in the SaaS world. Not overly worried there.

In closing, I will say that I have grown more and more fond of Cloudflare recently. There is much more to say and research but this seems long enough. I see them as the future for developers, for security, for the Internet in general. I could be wrong but I at least see a good number of years on the horizon as an investor in this business. You should always do your own research but I hope this was helpful to some (excuse the stream of conscience style of writing). I and others would greatly appreciate and benefit from feedback, both affirmative and/or contrary.

Long NET


One thing that concerned me about Cloudflare as investor (as a small customer, I do like it a lot) is the large number of free customers.

Free customers base lets Cloudflare effectively to sort out “bad” requests (made by bots) and pass only “good” requests to the customers’ websites (the DDoS protection is an important part of Cloudflare business).

So, when a “normal” SaaS company can always cut S&M expenses and immediately become hugely profitable, Cloudflare just can’t do that.

They will have to support that huge free customer base forever.

So I’ve decided to check how much they really spend on free customers.

From S-1 we can get the total cost of bandwidth and allocation:

“Bandwidth and co-location costs for paying customers are recorded as cost of revenue in the consolidated statements of operations and as sales and marketing expense in the consolidated statements of operations for free customers. Such costs totaled $17.7 million, $19.2 million, and $27.5 million for the years ended December 31, 2016, 2017, and 2018, respectively, and $12.7 million and $17.1 million for the six months ended June 30, 2018 and 2019 (unaudited), respectively.”

So, the total bandwidth and co-location expenses were:
2016 - 17.7M, 20% of total revenue of 84,8M, 44% of s&m expenses of 40,1
2017 - 19.2M, 14% of total revenue of 134,9M, 31% of s&m expenses of 61,9
2018 - 27.5M, 14% of total revenue of 192,7M, 29% of s&m expenses of 94,4

All (both for paid and free customers) of their bandwidth and co-colation costs were only 29% of s&m expenses in 2018, the costs associated with free customers should be less than that, and there is a downtrend, so I think there is really nothing to worry about.


Yeah, I tend to look at the free tier as a combination of cost of revenue and sales and marketing expense.

Cloudflare has mentioned that the amount of traffic free users generate is very helpful in securing good co-location deals with ISP’s. That creates a flywheel effect, making the network faster, which encourages more customers (free and paying) to sign up. From the S-1:

Because we manage the execution and prioritization of code running across our network, it means that we are both able to optimize the performance of our highest paying customers and also effectively leverage idle capacity across our network. We have chosen to utilize this idle capacity to create a free tier of service—which has generated substantial global scale for us. In turn, this scale makes us attractive partners for Internet Service Providers (ISPs) globally, which reduces our co-location and bandwidth costs. As our network grows, these dynamics become even more powerful.

Management has also mentioned the value of the free tier as a nice funnel into paying services, which reduces the amount of “formal” sales and marketing spend, and the value of the free tier in terms of testing the value of new services before selling them.

NET S-1:…