Fastly Q3 Notes

These are my notes. Many are in my own words. I tried to keep opinions to a minimum.

Shareholder Letter: http://d18rn0p25nwr6d.cloudfront.net/CIK-0001517413%20/48325…

  • Revenue $71 million, up 42% YoY

  • Dollar-Based Net Expansion Rate (DBNER) of 147% (!!!), up from 137% in Q2 20201
    – Net Retention Rate (NRR) of 122%, compared to 138% in Q2 20202, and 141% on a last-twelve-month (LTM) basis ← "We believe the LTM Net Retention Rate is supplemental as it removes some of the volatility that is inherent in a usage-based business model. NRR measures the net change in monthly revenue from existing customers in the last month of the period compared to the last month of the same period one year prior." I’m not clear on how much of this is attributed to TikTok nor am I convinced that smoothing the numbers to look better is the right way to view this drop.

  • non-GAAP gross margin 59.8%, up 56.1% YoY.

  • R&D 26% up from 24% and $18M up from $12M

  • S&M 32% DOWN from 35% but $23M UP from $18M. Spending more but not scaling the same pace as revenue.

  • G&A 34% up from 21% or $24M up from $11M. That is a big increase.

  • non-GAAP loss of $4M vs $8M while GAAP loss was $24M up from $12M (This level of finance is not my strong suit)

  • They added 9 enterprise customers. Here are my updated tables that show total customer and average spend are not accelerating. Yes there are some one-time events here, and they say some things pushed in to next quarter, but even ignoring some of last quarter as a rare spike, it would be a stretch to view this as acceleration.


TOTAL CUSTOMERS
Quarter:	Q3'20	Q2'20	Q1'20	Q4'19	Q3'19
Total:  	2,047	1,951	1,837	1,743	1,684
Change:  	  96	 114	  94	  59	  57
Growth:  	4.9%	6.2%	5.4%	3.5%	3.5%
					
AVERAGE SPEND					
Quarter:	Q3'20	Q2'20	Q1'20	Q4'19	Q3'19
Total:  	$753k	$716k	$642k	$607k	$575k
Change: 	  37k	  74k	  35k	  32k	  19k
Growth:  	5.2%	11.5%	 5.8%	 5.6%	 3.4%

  • Compute@Edge has moved out of beta into limited availability, with customers now running production traffic on the environment…creating new and innovative applications ranging from waiting room tokens, dynamic personalization, authentication at the edge, full serverless applications, and more. … Feedback from customers, including Vox Media, HashiCorp, and loveholidays, is that Fastly’s deep investments in Compute@Edge are delivering on the promise of serverless with rock-solid performance and features”

  • acquisition of Signal Sciences complete but integration still underway - called Secure@Edge

  • …in Q3, we did not meet our original third-quarter forecast due to two distinct challenges that impacted our revenue from a few key customers:
    – TikTok (10.8% of revenue) has mostly migrating away by the end of the month “in response to the potential of a prohibition of U.S. companies being able to work with this customer.”
    – “Existing customer timing: In the latter part of Q3, our forecast for the timing of new traffic coming onto our network from a few existing customers did not meet our expectations. A majority of these timing issues have been resolved…”
    – "Aside from these few customers, the business performed as we had forecasted. In addition to these two factors, our Q4 guidance now includes the revenue contribution from Signal Sciences.

  • “Ecommerce - …in ecommerce, one of the largest sportswear and footwear retailers in the U.S. migrated to Fastly during the third-quarter, , leveraging our security portfolio to enhance ecommerce security for its customers, as well as our Image Optimization. They’re also utilizing the ability to build complex application logic at the edge using VCL and deploy in real-time…”

  • “Media - …continuing to win new business in Europe. Our media delivery continues to differentiate by providing consistent, exceptional performance at large scale.”

  • Google Cloud Marketplace mentioned but nothing quantified.

  • “…several significant customer wins, including a global leader in telecommunications, media and entertainment; a leading EdTech software and solutions company widely used by higher education, K-12, business, and government clients around the world; and one of the largest American multinational retail companies.”

Guidance

REDUCED! FY2020 dropped to $288-$292 from $290-$300 and they will be losing more per share ~(0.91) now compared to ~(0.04).

Reading between the Lines
Look, there may be nothing between the lines to read here. 42% revenue ain’t bad. BUT having gone through this letter it feels to me that they are trying really hard to paint a rosy picture. Adding extra numbers to make other numbers look better or less important. That DBNER number is insane. I don’t understand how they can put up a number like that and not win more customers at the same time. The general language felt stretched to me. I will listen to the call, but…

I’m going to be selling all of my shares for now. I just don’t like the amount of spin I’m feeling coming off this letter. It is hard to quantify this feeling. I’m curious if anyone else agrees with me about an under-tone that makes me picture a forced-smile that doesn’t include the eyes.

I love what this company is doing. I think their tech, culture and mission are all special. At the same time I’m not all that confidant in the business-side after this one.

57 Likes

Here is the revenue growth with Q4 added in using the top of the guidance range. This quarter in bold, guidance in italics:


Quarter:	*Q4'20*	**Q3'20**	Q2'20	Q1'20
Revenue:	   *84*	   **71**	 74.7	 62.9
YoY Growth:	*42.6%*	**42.6%**	61.7%	37.9%

1 Like

RafesUserName,

That 80-84M Q4 guidance includes 8M from Signal Sciences. Organic revenue is 74M (mid), i.e: 25% YoY Growth.

11 Likes

Note that the Q4 guidance from the press release includes 8m from the Signal acquisition, so we’re looking at organic growth from 71 this Q to 72-76 next Q. Tough comps coming up in Q2 21 as Saul had pointed out too. I maintain that until the edge compute story starts generating noticeable revenue that FSLY is mispriced, not a bad company by any means but hard to see a ton of upside at 70 when compared to some of the alternatives out there

2 Likes

Rafe, in their investor letter they say that $8 million of their estimate for next quarter is revenue from the acquisition and not organic growth. That drops organic revenue down to $76 million, and organic growth rate to just under 29%.

https://investors.fastly.com/files/doc_financials/2020/q3/3Q…

With the successfully completed acquisition of Signal Sciences on October 1, 2020, our Q4 and FY2020 guidance includes the revenue contribution from Signal Sciences, which we expect to be approximately $8 million.

Saul

20 Likes

Hey Rafe,

I had the same feeling. The line about "the second-highest quarter of new customer additions since going public” was funny. They added fewer customers than the Q before. Shouldn’t new customer adds be accelerating? And “… since going public”? They just went public last May. This is not straight talk. Sure I could be overrating some ultimately unimportant spinning, but it could definitely be a red flag about Bixby’s leadership/candor, ability to recognize and confront threat.

My guess is competition is heating up and they’re not comfortable just saying it. And it is still possible TikTok left for reasons beyond just “geopolitical pressure.” Even that feels like an excuse. Ultimately it doesn’t matter why you lost your biggest customer. If you have the most dominant offering, Oracle, or whoever’s running them should stay with you.

Sure seems like a story shaping up of great tech, amazing potential, legit questions about execution.

BD

6 Likes

I’m not selling any shares and will consider adding.

Why? The bad news is already in.

They lost ~10% of their revenue from TikTok this quarter and MIGHT end up getting that revenue back. Additionally, they had some customer delays due to issues completely out of their control. Covid causing data center build-out delays + code-freezes.

Even with this almost worst-case scenario, they still managed 25% organic growth.

But I’m invested for the Compute@Edge and now Secure@Edge which look to be ramping up nicely. They have guidance of roughly 38% YoY growth next quarter which I think they’ll beat nicely.

They cited the strongest new bookings in a quarter this year and free cash flow was $13M up from ($17M) in Q3’19.

Gross margins and DBNER also ticked up. I like what I see and I feel management has been extremely transparent about the TikTok challenges.

Tough quarter, but this is a great company.

66 Likes

Rafe, in their investor letter they say that $8 million of their estimate for next quarter is revenue from the acquisition and not organic growth. That drops organic revenue down to $76 million, and organic growth rate to just under 29%.

Mathematically correct but misleading. We know that Fastly had a 10% politically motivated China accident. I think the right way to make this calculation is to take out Tik-Tok’s prior year’s contribution to revenue and then recalculate organic growth.

If my arithmetic is correct, the recalculated organic growth excluding Tik-Tok is north of 43%.

Contrary to the OP’s negative opinion, I found this earnings report quite satisfactory.

Denny Schlesinger

44 Likes

After reading the presentation provided, I wanted to point out couple of things which I have noticed before that to me are major red flags, and some data to proof that.

  1. Overall client growth rate: Previous QoQ was: 6.21%, this QoQ is 4.9%
  • Can’t find YoY total client growth
  1. Enterprise level client growth (generating >$100k / annually).
  • Q3: 9 new clients

  • Q2: 7 new clients

  • YoY growth: Q2 20 / Q2 19: 16.03%, compare that to Q3 20 / Q3 19: 10.95% (313 clients total in 2020)

Now let’s compare that to Cloudflare:

CloudFlare: Q2 20 / Q2 19: 64.60%, 637 enterprise clients total, added 250 new, vs 42 for Fastly in the same period, and 80 new clients for Cloudflare vs 7 new in the same period Q2 2020 (11 times difference!)

From Fastly’s investor presentation: “Enterprise customers generated 88% of our trailing twelve-month total revenue, consistent with Q2 2020”

You can draw your own conclusions I believe on the above.

  1. I don’t understand how TikTok moving to Akamai is really based on geo political pressure only? Akamai is based in Cambridge, MA.

  2. If people still think that everything is geopolitical and that Fastly is untouchable and companies can’t migrate out that easily, please check this (btw this will be another company worth looking for when it becomes IPO):

https://about.gitlab.com/blog/2020/01/16/gitlab-changes-to-c…

8 Likes

ByteDance, owner of TikTok, is in fact building a CDN of their own.
This is a current job posting on their website for a Program Manager at Mountain View, CA: https://job.bytedance.com/en/position/detail/688480895961889…
Among the responsibilities for this role:

  • The US Supply Chain team is looking for a Program Manager to support the rapid rollout of our CDN (Content Delivery Network) and PoP (Point of Presence) sites
  • Manage end-to-end Sourcing projects (including planning, execution, measurement, and optimization of projects) and ensure on-time delivery for the new regions CDN and PoP sites

I exited Fastly because I don’t have faith in the management team. From ER: Based on publicly available information, we believe this global traffic reduction was in response to the potential of a prohibition of U.S. companies being able to work with this customer.
If they have to rely on publicly available information to draw conclusions about their largest client’s exit, they have bigger issues.

7 Likes

Anyone want to weigh in on the implications of this for Cloudflare?

As a technoramus, I didn’t even know you could rapidly roll out a CDN.

If they have to rely on publicly available information to draw conclusions about their largest client’s exit, they have bigger issues.

I think that’s misunderstanding the situation. You don’t make comments about your customers, and they are using the ‘publicly available’ discussion as a fig leaf while talking about their customer.

Let’s say (even hope) that “you can’t roll out your own CDN” and be as good as Akamai, Cloudflare, and Fastly; Fastly would gladly get back in a room with ByteDance to discuss business, again. And that will be less likely if they comment too much about TikTok.

  • Another Rob
    Long FSLY, long NET
11 Likes

Anyone want to weigh in on the implications of this for Cloudflare?

To build a CDN with mediocre to average performance that can do the job is not as complicated, if its for own use only. You have to really be a giant company with lots of traffic like TikTok to make sense to attempt that. To add security features, and run own code on the edge however with similar performance to CloudFlare or Fastly is not feasible to be maintained in house for most companies. Facebook created their own caching technology Varnish, which is what Fastly is using for their caching engine.

No implications for CloudFlare - they have never promoted as a CDN, and never had a true CDN product. Their focus is on “making the Internet better”.

3 Likes

https://discussion.fool.com/4056/notes-from-the-cc-34652913.aspx…

@Fortun8, check out the notes from the CC. Fastly says they helped Tik-Tok shift off Fastly, quickly. I think this shows that they really believe they can get this relationship back, at some point in the future. According to TMFKaren’s notes, this was said in response to an observation/question about it being so easy/fast for Tik-Tok to move.

4 Likes

If I were Tik-Tok, I should think the first order of business is to bifurcate my servers, user data, and my CDN; one for China and a second for the US. So I would choose a CDN provider who can guarantee that my US data is not traversing a China or undesirable CDN POP. Dunno, can FSLY enforce geographical presence of routed and cached data? Building a US based CDN is no small short task and Tic-Tok is is right now paddling short legal strokes. So I see the CDN job post as intriguing but this does not seem to be a prudent short term task. Perhaps they just want a ;person to manage the migration of the CDN.

-zane

1 Like

Mathematically correct but misleading. We know that Fastly had a 10% politically motivated China accident. I think the right way to make this calculation is to take out Tik-Tok’s prior year’s contribution to revenue and then recalculate organic growth.

If my arithmetic is correct, the recalculated organic growth excluding Tik-Tok is north of 43%.

This way of thinking about organic growth doesn’t sit well with me. It seems to be saying there are certain circumstances (subjective?) where one can overlook a decline in revenue, and the “true” organic growth is to back out the previous year’s contribution for an apples-to-apples comparison. The last time I saw similar adjustments was in a rapidly declining retail REIT where management was trying to justify why comparable sales growth was positive in spite of store closures.

To extend this line of thinking further, if a company were to gain new customers due to political incidents, would we consider that organic growth? Or do we have to remove the newly gained revenue to get “true” organic growth?

IMO, one should not adjust for TikTok’s contribution in the prior year. I’m guessing not everyone will agree.

2 Likes

If my arithmetic is correct, the recalculated organic growth excluding Tik-Tok is north of 43%.

This way of thinking about organic growth doesn’t sit well with me.

IMO, one should not adjust for TikTok’s contribution in the prior year. I’m guessing not everyone will agree.

Context matters. I was taught early on that it is not enough to know what a chart says, one has to know why the chart is saying whatever it is saying. I would apply that lesson to any number we can come up with in our job as investors. In this case specifically, is Tik-Tok’s exit a one-off event or part of a trend?

Naturally not everyone will agree as we gauge Tik-Tok’s exit differently.

Denny Schlesinger

5 Likes

QUESTION:
3) I don’t understand how TikTok moving to Akamai is really based on geo political pressure only? Akamai is based in Cambridge, MA.

RESPONSE:
Oracle and Akamai Technologies have a long-standing business partnership going back to at least the year 2002. https://www.cnet.com/news/oracle-akamai-team-on-caching-tech…

TikTok is a viral Chinese video-sharing social networking service owned by ByteDance.

The United States Government has been fighting a “trade war” with China for the past several years. The “war” includes concerns about: theft of intellectual property and technology used to spy on America which is being manufactured in China and sold in the States. This is especially a concern for the transition to 5G and what may be hidden inside technology that America buys from China. Whether it is true or not, it is a major concern. https://www.cnbc.com/2018/10/04/chinese-spy-chips-are-said-t…

Companies in China are known to be more owned/controlled by the government and more secretive than “publicly owned” companies in the United States. The U.S government became concerned about the explosive use of a Chinese-controlled app (called TikTok) within its borders.

The most valuable commodity today is now data, just look at Snowflake. Unfortunately, data can be used for malicious purposes and as a “weapon,” just look at the elections, Cambridge Analytica, CrowdStrike, Palantir, etc. As a result, the U.S. government issued a deadline for TikTok to transfer everything over to American borders (for the safety and protection of the entire country), or else TikTok use in the States would have to shut down. Apparently TikTok did get banned in India.

Two U.S. companies took over control of TikTok in the States. One was Wal-Mart and the other was Oracle. Under Chinese control, TikTok was using the CDN from Fastly. But Oracle had a long-standing business partnership going back to at least the year 2002 with Akamai. Under Oracle control, TikTok began using the CDN from Akamai.

It’s not unusual for a public U.S. company like Fastly to be reticent about sensitive topics involving competition, customers, and divisive issues like politics. Sometimes competitors can become partners. Sometimes former customers can become return customers. In a tense political (and competitive) environment, is it best for a company to delve into the details of divisive drama? Or is it better to keep it simple, stupid? Fastly updated the public about the news in a simple and straightforward manner: “Due to the impacts of the uncertain geopolitical environment, usage of Fastly’s platform by its previously disclosed largest customer did not meet expectations…”

Does it have to be complicated with wild, unsubstantiated speculation that Fastly did something wrong with TikTok? Or what do they say, the simplest explanation is most likely the right one? Because of geopolitics, TikTok transferred to Oracle and their partner CDN.

FURTHER:
And for all the criticism about the original Fastly CEO becoming Chief Architect, it’s not unusual for the “brains” of the operation to prefer being behind the scenes. One example that comes to mind was a little e-commerce company from a Canadian coder. The CEO almost stepped aside and gave the “public role” to someone else too. He ended up remaining CEO, but since it’s no longer a little company, the CEO tends to let the COO, and now President, deal with a lot of the publicity. It comes to mind because it’s also a company that does business with Fastly. On the homepage, it’s the first company on the homepage under: “WE HELP GOOD COMPANIES DO GREAT THINGS.” It’s the one that starts with the letter S: https://www.fastly.com Don’t mean to speculate here, but maybe folks at these two innovative companies understand each other and work well together.

Thursday night, Fastly CEO Joshua Bixby appeared on Mad Money with Jim Cramer. You can hear the full interview on the Mad Money podcast, or watch a clip on YouTube. Maybe they will post the entire interview on YouTube in the future. It’s like how both Twilio CEO Jeff Lawson and Zscaler CEO Jay Chaudhry came on the show after similar volatility. Were they broken companies or just broken stocks? If you can find a leading-edge company that hasn’t had a stumble or two along the way, then expand your time-frame because it happens to the best of them. Jeff Bezos, Reed Hastings, Jensen Huang, Elon Musk, and even the red-hot Frank Slootman.

ServiceNow hammered by error, pulls other B2B cloud stocks down with it
https://www.bizjournals.com/sanjose/news/2016/01/28/servicen…

Oops! Forecasting Error Hammers ServiceNow as Shares Fall 20 Percent
https://www.vox.com/2016/1/28/11589158/oops-forecasting-erro…

Why ServiceNow Inc. Stock Sank Today
https://www.fool.com/investing/general/2016/01/28/why-servic…

Frank Slootman’s company stock went down 50% from the high ($90), but ever since, it’s climbed from $45 to over $500 in about 5 years. Not bad, even after he left the company, and there’s been 2 different CEOs since.

10 Likes

If you can find a leading-edge company that hasn’t had a stumble or two along the way, then expand your time-frame because it happens to the best of them. Jeff Bezos, Reed Hastings, Jensen Huang, Elon Musk, and even the red-hot Frank Slootman.

Apple Is Worth $1,000,000,000,000. Two Decades Ago, It Was Almost Bankrupt.
By Jack Nicas
Aug. 2, 2018

SAN FRANCISCO — In 1997, Apple was on the ropes. The Silicon Valley pioneer was being decimated by Microsoft and its many partners in the personal-computer market. It had just cut a third of its work force, and it was about 90 days from going broke, Apple’s late co-founder, Steve Jobs, later said.

On Thursday, Apple became the first publicly traded American company to be worth more than $1 trillion when its shares climbed 3 percent to end the day at $207.39. The gains came two days after the company announced the latest in a series of remarkably profitable quarters.

https://www.nytimes.com/2018/08/02/technology/apple-stock-1-…

Jobs’ mistake was to go head to head against IBM in the business market. That’s not how disruptive innovation works and Jobs’ ego almost bankrupted Apple. Jobs learned well in exile at Next and Pixar and his comeback was “insanely great.”

https://www.amazon.com/dp/B006ZA7E6M/ref=dp-kindle-redirect?..

No, I have not read the book but I was an Apple developer when Jobs got himself fired. I’m also a former IBM sales rep.

Denny Schlesinger

6 Likes

Some times special companies doing special things you can give a longer leash. Eg Tesla or Amazon or Zoom or Nvidia or Shopify, truly unique world disruptive companies. Note: I’m using these as examples not specifics to make my point.

This said, what makes Fastly so special, unique, must have? How are they disrupting and changing the world?

I see yet another CDN (yes, maybe they do things better. I don’t know. But switching costs were not much if TikTok is an example) that is planning to do the same on the edge (like so many others are).

So this also said, does a Fastly deserve a long leash when (correct me if I’m wrong) it is a technology provider not much different than multiple others, who is also growing slower (materially so) than say a NET?

Doing so during a period of time when remote and content delivery is skyrocketing at that.

So from this perspective, what makes Fastly special enough to argue so fiercely for it? I’m not saying it sucks (no, it has a good business going on). I’m merely saying it does not appear to be that rare special one of a kind world changing business that justifies a long leash and such passion for it.

Succinctly let me know why I may be wrong here.

Tinker

17 Likes