Fastly (FSLY) Q4 2019 Earnings Results Review

Fastly announced earnings results for Q4 and full year 2019 on February 20, 2020. Results beat expectations on most measures, with strong revenue outperformance for Q4. Bundled with the earnings results was an announced CEO transition. The market reacted negatively the following day, pushing the stock down about 6%. This did coincide with a general market sell-off, which may have over-corrected the reaction. Projections for 2020 were inline and provide room for upside. Let’s take a look at the details and take-aways for investors.

Headline Financial Results (EPS is Non-GAAP)

• Q4 2019 Revenue of $59M vs. consensus estimate of $54.9M, up 44% year/year. Beat of $4.1M or about 7.5%. Note that original guidance represented growth of 33-34%, versus the actual growth rate of 44%.
• Q4 2019 EPS of ($0.10) vs. consensus estimate of ($0.11), representing a beat of $0.01. Compares to ($0.24) for Q4 2018.
• Non-GAAP gross margin of 57.6%, up from 56.8% year/year.
• Non-GAAP operating margin of -15%, compared to -14.6% in Q4 2018.
• Free Cash Flow was -$9M in Q4 2019, versus -$8M in Q4 2018.
• FY 2019 Revenue of $200M, up 39% year/year. Prior estimate was $196M, representing growth of 35.5% year/year.
• FY 2019 EPS of ($0.52) vs. consensus estimate of ($0.52). Compares to ($1.08) for 2018.
• Full year Non-GAAP gross margin of 56.6%, up from 54.9% in 2018.
• Full year Non-GAAP operating margin of -17%, compared to -17.3% in 2018.
• Full year free cash flow was -$50.7M, compared to -$36.6M in 2018.
• Q1 2020 Revenue estimate of $58-60M vs. $58.1M. Raise of $0.9M or 1.5% at the midpoint. Represents year/year growth of 29.4% over Q1 2019.
• Q1 2020 EPS estimate of ($0.12) vs. ($0.11) estimated. Lowered by $0.01.
• FY 2020 Revenue estimate of $255-265M vs. $256.5M estimate. Raise of $3.5M at the midpoint or 1.4%. With FY 2019 Revenue of $200M, this estimate represents a growth target of 30%.

Other Notes

• DBNER of 136%, up from 135% in Q3 2019 and 132% in Q2.
• 288 enterprise customers at the end of Q4 2019, versus 274 at end of Q3 and 227 at the end of 2018. This represents a sequential growth of 5.1% and about 27% for the full year of 2019. Enterprise customers are defined as those who spend over $100k in the prior 12 months.
• Total customer count increased to 1,743, up from 1,684 in the previous quarter and 1,582 at end of 2018, representing growth of 3.5% and 10.2% for the year.
• Average enterprise customer spend increased to $607k, up 5.5% from $575k in Q3 2019 and $556k in Q2. Management attributed this to increased adoption of optimization and security products such as Media Shield, Cloud Optimizer, WAF and the new TLS services.
• Enterprise customer spend represented 87% of trailing 12-month revenue, up from 86% in Q3.
• Annual revenue retention rate increased to 99.3% for the full year 2019, up from 98.9% in 2018. In addition to DBNER, overall revenue retention rate demonstrates the stickiness of the platform as customers build critical infrastructure around Fastly services.
• Ended Q4 2019 with $201M in cash and equivalents. This compares to $207.8M at the end of the prior quarter.

• Product focus for 2020 will continue to be on the Compute@Edge offering, with potential for new security offerings at the edge.
• Other product improvements announced in Q4:
  • Launched new TLS solutions (security protocol) to enable better and easier encryption of routed traffic.
  • Launched Origin Connect, which provides a direct fiber connection between a customer’s origin servers and the edge cloud network. This should decrease response times to refresh cached content and lower transit costs.
  • Launched Cloud Optimizer which streamlines content sourcing across multiple cloud providers.

• Fastly’s edge compute solution, Compute@Edge, was launched in Beta in November. Since then, the Fastly team has been gathering developer feedback and incorporating this into support for new use cases. Fastly didn’t announce any specific improvements to Compute@Edge on the conference call.
• Highlighted some significant customer engagements for the edge compute solution. One example was Jetblue, which uses edge compute to move application logic closer to customers, specifically personalization of content based on geo-detection, which is an ideal use case. Management also called out an unnamed global online payments company that is “transforming its monolithic edge into a workflow designed for devops, automation and scale” (arguably a bit vague). Finally, cited customers Taboola and Shopify as continuing to expand their use of the platform to move application logic to the edge and leverage real-time analytics to drive deeper insights into content delivery.
• Expanded their physical network and global presence locations. Added new POPs in Dublin and Vienna, and added an additional POP in the U.K. Upgraded 8 existing POPs in various locations in the U.S. and Canada. In total, have 68 POPs online across 53 markets providing access to 74 Tb/sec of global network capacity.
• Made multiple software improvements to the hosting software stack that improved operational efficiency. Specifically, led to a reduction in computing requirements for common platform workloads, allowing for better utilization and capital expenditure efficiency.

CEO Transition

The biggest announcement associated with the quarterly results was that founder and CEO Artur Bergman would be transitioning into a Chief Architect role. Artur presented this change as allowing him to “focus more of my attention on our customers and to work more with the product and technology part of our business to build out our edge cloud platform.”

Replacing Artur as CEO will be Joshua Bixby, the President since May 2017. Joshua joined Fastly as the VP of Marketing in 2015, after being a part-time advisor since 2013. Between 2015 and 2017, Joshua held Senior VP roles over marketing, product and sales. Prior to joining Fastly, Joshua was briefly a VP of Acceleration at Radware (RDWR) in 2013. Before Radware, Joshua co-founded a web application acceleration company called Strangeloop Networks in 2006, which Radware acquired in 2013. From 2002 to 2006, Joshua was founder and CEO of IronPoint Technology, a content management software company. He has a degree in Management from University of Toronto.

I have mixed feelings about this transition. On the positive side, I applaud the transparency and humility of Artur to recognize that he may not be suited for the CEO role. More significantly, this move will allow him to focus all of his time on product and customer adoption, which should generate timely outcomes as Fastly is rapidly evolving the edge computing space and needs short feedback iterations from primary customers. Artur emphasized this benefit on the analyst call, pointing out that it repeats the same customer adoption focus that he exercised at Fastly founding, but now for the Compute@Edge product. He is excited to spend more time in person with engineers and customers, getting back into the field to learn from customers and educate them on what’s possible with Fastly edge solutions.

On the negative side, I prefer companies with technical founders in the leadership role, and that is not Joshua Bixby. However, this is mitigated by Artur’s (presumed) full involvement going forward. The reality is that if Artur isn’t suited for or interested in managing sales, marketing, HR, finance and the other duties of a CEO, then a change now is probably for the better. With that said, his replacement, Joshua Bixby appears to me unproven as the CEO of an enterprise software company. He did found and exit two start-ups in the content delivery space, so has deep domain experience and the ability to build teams. However, I would feel a little better about this side of the transition if Joshua were a more recognized player, with a track record at an enterprise software company. We will have to watch his performance.

Analyst Reactions

I found only one analyst update following FSLY earnings results. FSLY’s analyst coverage is fairly light at this point. Piper Sandler analyst James Fish maintained his Overweight rating and raised the price target from $24 to $27.

Fastly reported a “strong” Q4, with 7.5% upside driven by the advancement of Disney+ and other live streaming events in the quarter, Fish tells investors in a post-earnings research note. He also believes the CEO transition makes sense and is a positive for the company.

James Fish, Piper Sandler, Sourced from Thefly.com

Peer Performance – Cloudflare (NET)

Cloudflare (NET) reported earnings on February 13. Cloudflare is a direct competitor to Fastly in edge computing and also has overlap in CDN and other security offerings. So, looking at NET’s Q4 2019 performance may be enlightening for investors. I won’t cover all details from NET’s earnings report (and may do a deep dive at another time). Here are some highlights from NET for comparison.

• Q4 2019 revenue grew 51%, with FY 2019 revenue growing 49%. This is on a revenue base of $287M for 2019, versus $200M for Fastly. Compound annual growth rate over the past 4 years has been 50% for NET.
• Q4 2019 Non-GAAP gross-margin was 78.7%, with an increase of 1.8% year/year.
• Q4 Non-GAAP loss from operations was $18.3 million, or 21.8% of total revenue, compared to $15.9 million, or 28.7% of total revenue, in the fourth quarter of 2018.
• Q4 Free cash flow was negative $23.5 million, or 28% of total revenue, compared to negative $29.0 million, or 52% of total revenue, in the fourth quarter of 2018.
• FY 2020 Revenue of $391M at the mid-point, representing growth of 36% year/year.
• DBNER of 112%, compared to 136% at Fastly.
• Grew large customers, defined as those with >$100k annualized billings, from 313 to 550 in 2019, representing growth of 75%. This is compared to Fastly’s increase of 27% in 2019 to a total of 288. Management clarified on the analyst call that at least 50% of revenue from large customers is associated with security products.
• NET has 2.6M total customers, most of whom utilize a free offering. Of those, 82,882 paid in 2019, which was growth of 22% over 2018. All of Fastly’s customer are paid, and this growth rate is more inline with Fastly’s growth of 27% in enterprise customers.
• Launched Cloudflare Teams, which is a new security offering to allow corporate employees and their devices to access internal applications securely. It includes Cloudflare Access, which is the equivalent of a VPN, and Cloudflare Gateway, which offers a next generation firewall capability. Cloudflare Teams will be enhanced by the browser isolation technology from S2 Systems, a recent acquisition.

As illustrated in the slide above and on the analyst call, the Cloudflare CEO takes the position that large customers prefer a “unified network” that offers security, performance and reliability solutions from a single platform provider. His perspective is that competitors, which offer “point solutions” (like Fastly, Zscaler, Akamai) in these broad areas will be at a disadvantage to Cloudflare. Personally, I think this is a stretch. At minimum, you would likely have different internal buyers within customer organizations for security solutions versus application performance optimization.

Comparing some key metrics between FSLY and NET is enlightening (Non-GAAP):

Metric		FSLY	NET
2019 Revenue Growth		39%	49%
Q4 Gross Margins		57%	78%
Q4 Operating Margins		-15%	-22%
Q4 DBNER		136%	112%
2019 Large Customer Growth (>$100k)		27%	75%
EV / Revenue		9.1	19.8

As you can see above, NET beats FSLY on several metrics. Also, NET operating margin showed improvement of about 7% year/year, versus being static for FSLY. However, the difference in metrics between the two is reflected in valuation, with NET garnering an EV/Revenue ratio of more than 2x FSLY.

My Take-aways for FSLY

• Ended the quarter and year with strong forward momentum. Q4 Revenue was a blow-out, surging to 44% year/year growth, which was 10% over the estimate. This represented a great end to the year, and this was called out on the earnings call by several analysts. Granted, Q4 is seasonally strong and Fastly can opt into a lot of live events incrementally, but performance was still beyond expectations.
• The 2020 revenue estimates seem conservative, allowing for similar upside as seen in Q4. The CFO even hinted at this on the analyst call. “The fact that we are a usage base, there is some variability that can occur and so I think what you saw there is, we just finished a great year, a great quarter and what you see here is just some appropriate conservatism given that we are just this early in the year.” Also, 2020 should have a number of live events to tap into, associated with the elections and the Olympics, which are marginally included in the guidance.
• I particularly like the high DBNER of 136%, which has been increasing over the past two quarters. With enterprise customers accounting for 87% of revenues, DBNER by itself could drive 2020 revenue targets.

• Comparables to NET are troubling on customer count growth. NET seems to be adding $100k customers at a much higher rate than FSLY. Analysts called this out on the FSLY earnings call as well. My perspective is that FSLY is more targeted in its large customer engagements, as reflected by DBNER and average spend growth, but broader TAM grab is happening now.
• Gross margins improved about 1.5% year/year, but are still around 56-57%, which seems relatively low. From the earnings call and financials, we can infer that this includes both fixed costs like network/hosting and variable costs of labor. I think the labor component is the primary opportunity for optimization. Fastly leadership provided some examples of efficiency improvements on the call, which should slow down staffing requirements relative to future revenue growth. Also, as Fastly rolls out more purely software based services, like edge computing and security, we should see margins raise further. For comparison, Cloudflare and Akamai gross margins are in the high 70% range on a non-GAAP basis.
• With Compute@Edge coming out in Beta in November, leadership has set a goal for 2020 to “bring out Compute@Edge at scale.” In theory, this would represent potential for upside, as we can assume not much revenue has been associated with this offering to date and there will be further monetization opportunities.
• I was hoping to have some new feature announcements about Compute@Edge in Q4 earnings, like additional language support. The Shareholder letter really just rehashed what we already know. Granted, the product was just launched to Beta in November, but rapid progress will be key given competitive offerings and the large bet FSLY is placing on this offering. I will be looking for new announcements as Q1 progresses. NET launched a number of new products in their Q4 release.
• Hints at edge product extensions into security are interesting and make sense. On the analyst call, Artur specifically referenced customer conversations about security concerns in standard runtime environments from memory overrun exploits like Specter and Meltdown. The Compute@Edge platform is ideally suited to mitigate these risks, since each request essentially has a dedicated server runtime. That capability seems to be very desirable for customers and Artur hinted at deep conversations with customers about it. This could be a nice competitive differentiator for Fastly.

Items to Watch

• Gross margins and profitability improvements. As mentioned above, gross margins seem low compared to Cloudflare and even Akamai (mid 60% range on a GAAP basis, high 70% range on a non-GAAP basis). Fastly is making improvement here and we will want to continue to monitor improvement going forward. Several analysts asked for clarification on gross margin improvements on the earnings call, so it seems to be a topic of interest.
• Overall path towards profitability needs to demonstrate improvement. Operating margin was flat for both 2018 and 2019 at about -17%. I appreciate that for now, the company is investing in growth, but we should see this ratio tick up. NET showed year/year improvement in operating margin of 7%, although from a lower percentage.
• Given that Compute@Edge is held out as the next big thing for Fastly, I want to hear about new product launches and enhancements. I’d like to see a rapid fire of announcements over 2020 with actual customer use case examples. This will give confidence that Compute@Edge can fuel upside and drive the margin improvements associated with software offerings.
• Additionally, I will be interested to see what hints at new security offerings will represent. This could also be incremental, but there is a lot of competition in security, so it will depend on what type of product niche Fastly addresses.
• Large customer growth seems a little slow, at least when compared to Cloudflare (NET). This is likely due to Fastly’s focus on only certain customer types and use cases, but we will need to watch this going forward.

Investment Plan

As I discussed in previous my detailed analysis on Fastly (FSLY), I am not ready to make a recommendation yet. This is primarily due to the recency of the IPO (May 2019) and I tend to avoid investments within one year of an IPO. With that said, I think Fastly has long term potential, assuming they can execute on their vision of being the leading provider of fast, secure edge compute. With a current enterprise value under $2B and 2019 revenue of $200M, the trailing EV/Revenue ratio is about 9.1. If we assume revenue growth at or over 30% for 2020 and 2021, EV/Revenue would drop below 6 by 2022. Optimistically, in 2 years, we could see 66-100% share appreciation to get back to a 10-12 EV/Revenue range for 30% revenue growth with middling gross margins and (assumed) improving operating margins.

At the same time, there are other companies (like ZEN) with growth in the low 30%’s, but with higher gross margins (ZEN non-GAAP gross margin of 75% for 2019) and profitability (ZEN non-GAAP operating margin of 3.3% in 2019, growing to 5% in 2020), that have similar valuation ratios (ZEN EV/Revenue of 11-12). I think Fastly’s success will depend on its ability significantly outperform revenue targets going forward, like we saw in Q4 (target of 33.5%, actual of 44%). In theory, a DBNER of 136% and revenue retention over 99% would almost guarantee this.

The CEO transition also concerns me. I prefer software stack companies that are led by a technical founder. The replacement CEO is unproven at this level of operation. The rationale for the transition makes sense, though, and Artur’s focus on customer adoption could bear significant fruit.

Given all of this, I will continue to monitor Fastly for another quarter, and may open a small tracking position near term, given that the stock has landed below $21 as a result of the recent market sell-off. I will make a formal recommendation after the IPO anniversary, depending on how Q1 shapes up. I would want to see substantial revenue outperformance, improving gross margin and maintained DBNER, as well as updates on Compute@Edge and the CEO transition.