Cloudflare (NET) Thesis

Summary:

Cloudflare (NET) as a SHORT presents an interesting risk/reward as we approach Q4’22 earnings. In Q3'22, cracks started to appear in the companies' financials with a relatively weak topline beat and guidance raise, which was met with a -18.4% stock price decline the following day. Revenue began to decelerate and fell below the company's consistent 50% growth in the previous eight quarters, which the sell-side suggested investors were "anchored" on. Dollar-based net retention (DBNRR) continued to fall, falling 300bps to 124% from its peak of 127% in Q1’22; this is significant because management has promised a long-term DBNRR target of +130%. In addition, weakness began showing up in the company's primary growth driver, the large customer (>$100K) segment, with net new $100K+ customer additions down -8% YoY, which is meaningful, especially as net new adds had grown 48% YoY the previous quarter (Q2’22). NET’s growth story is dependent on success in the large customer (>$100K) segment, where topline growth of 50%+ has been responsible for the company's overall 50% revenue growth rate vs. the mid-market and SMB (<$100K) segment, which has been a drag on NET’s overall topline as the segment has grown less than 30% YoY the last 4 quarters.

The beat was relatively weak vs. NET’s historic beat cadence, coming in at $3.4m. The previous eight quarter average beat was $7.9 million, with a $7.5 million beat in Q2'22 and a $6.7 million beat in Q1'22. Revenue guidance was soft, with management raising the FY’22 guide by only $4.5m. The raise was likely well below buyside expectations as management previously raised the FY’22 guide by $13m in Q2’22 and $28m in Q1’22. Furthermore, to account for any seasonality in the guidance, we note that the prior year Q3'21 guidance raise was $16.5m on a much lower revenue base. Management had previously raised guidance by an average of $18.5m over the past seven quarters (excluding Q4).

Typically, investors would simply adjust down their expectations of future beat ’n’ raises; however, we don’t believe this is the case as management decided it was a good idea to issue a LT revenue target of $5bn within the next 5 years on the Q3'22 earnings call, requiring revenue to grow at a 38% CAGR over the next 5 years. Management did not provide a bridge to their new target, so we think investors will assume in their models a beat that will equate to at least the 38% annualized CAGR management is promising. This assumption would require the company to grow faster than their Q2/Q3’22 sequential growth rate over the next 5 years if we assumed linear distribution, leaving no room for a continued beat ‘n’ raise cadence and no margin of safety for further revenue deceleration. It is our view that the $5bn target was a short-term lever management decided to pull in hopes of minimizing the impact on their stock price, but this lever has essentially painted management into a corner. We believe management will no longer be able to deliver the beat ‘n’ raise cadence that is baked into NET's premium multiple. If we are correct, management will likely have to pull their LT guide as revenue growth slows. If management does not withdraw their LT $5 billion guidance, we believe the market will simply lose faith and assume the target is unattainable if revenue continues to decelerate below the required 38% CAGR, putting them back in the same situation and likely no longer justifying NET trading at a premium to best-of-breed cybersecurity peers. 

We also find it curious that the Head of Investor Relations, Chief Revenue Officer, and Head of Finance/Chief Accounting Officer announced their departures in Q4/Q1. In our opinion, there are few roles that have a better understanding of the revenue opportunity than the Head of Investor Relations, Chief Revenue Officer, and Head of Finance/Chief Accounting Officer, so why are they departing now? It is our speculation that their departure could have possibly been partially due to realizing management’s targets were unrealistic and lofty at best.

Our catalyst is management issuing disappointing FY’23 guidance on the Q4 call that will likely fall materially below buyside expectations and potentially lose investor confidence in management’s ability to achieve their 5-year target. We believe this target is unattainable because it would require management to maintain their previous quarter's sequential topline growth rate of 8.25% (38% CAGR) over the next five years just to meet their target, effectively killing the beat 'n' raise story as their revenue growth rate would have to accelerate above the previous quarter's growth rate to achieve a beat. We believe their ability to maintain their current growth rate, let alone accelerate to maintain their beat ‘n’ raise story, is highly unlikely as we enter a more challenging macroeconomic environment for software and as NET becomes more reliant on the still relatively new and unproven Zero Trust offering as their main growth driver due to Application Services maturing.

Valuation and Comps:

Sales Estimates vs. Street:

NET Beat/Raise, Guidance Raise/Lower, Next Session Stock Reaction:

Company Overview

Cloudflare is a content delivery network (CDN) that provides over 30 products in four primary categories, including Application Services (CDN, DNS, Load Balancing, DDoS), Zero Trust Security (Secure Web Gateway, VPN, DLP, PAM, Email, CASB), Network Services (WAN, WAF, Smart Routing), and Developer Solutions (Storage, Database, Compute). Cloudflare operates a global edge network that has a presence in over 270+ cities and 100+ countries, and provides services to over 20% of the web and 30% of the Fortune 100. The company has a unique and large customer base comprised of 150k+ paying customers and millions of free users.

The company estimates its TAM at $115bn and anticipates it will grow to $135bn by 2024, based on IDC forecasts of the markets' NET addresses. The company believes their TAM estimate could prove conservative as it does not include newer products such as serverless, database, IoT, and 5G that Cloudflare plans to rollout in the future, which could greatly expand their TAM.

NET primarily generates revenue from subscriptions to their platform and products. Pricing is largely based on customer size and also on the plan or tier, which provides different levels of capabilities and features. Enterprise plans are customizable and provide enhanced services, including SLAs and advanced functionality. Self-service plans are priced based on the number of registered domains and internet properties (domains, websites, and mobile apps). The company’s roots started in SMB, but they have steadily moved upmarket, with large customers (>$100K) accounting for 61% of revenue and the remaining 39% coming from mid-market and SMB customers (<$100K). Large customers (>$100K) typically sign 1 to 3-year contracts and are billed on either a monthly or annual basis. Self-serve customers typically pay via credit card and are primarily on monthly contracts. 80% of total customers are billed monthly, while 20% are billed annually. The company has millions of free customers, who are primarily individual developers and hobbyists that are given access to much of the Cloudflare suite for free as the company leverages their free user base for lead generation and beta testing new products.

Consensus View

Cloudflare is viewed as a disruptor of $115bn TAM, akin to a "fourth cloud", due to the breadth of services the company provides delivered via their global edge network. Bulls believe the company is largely underpenetrated and has only captured <1% of their TAM. By leveraging NET's base of millions of free users and over 150k paying customers, management can drive the company to its target of $5bn within 5 years (38% CAGR). The company has a philosophy of "stacking s-curves," whereby management has labelled certain product categories as "Acts" in its product roadmap. Application Services (Act 1), Zero Trust (Act 2), Workers (Act 3), and MPLS (Act 4), each product offering is supposed to "act" as the next leg of growth in Cloudflare’s journey, thereby "stacking S-curves."

Sell-side analysts and investors have been vocal in their confidence in Zero Trust as Cloudflare's (Act 2) primary go-forward growth driver. Zero Trust is succeeding Applications Services (Act 1), which was largely responsible for propelling the company to $1 billion in annualized revenue and has naturally begun maturing. Cloudflare is also making strides with their Act 3 product, Workers, which is largely comprised of Compute, D1 (database) and R2 (storage) products that allow developers to build applications at the edge. NET bulls strongly believe the company will continue their rapid growth trajectory, driven by continuous product innovation, as management continues to execute towards their $5bn target within 5 years (38% CAGR).

Variant Perception

NET demonstrated incredible success in Application Services (Act 1), as evident in their ability to achieve +47% growth (Q3’22) at scale with over $1bn annualized revenue, which we believe is primarily attributable to Application Services products. However, going forward, management made it clear that Application Services is maturing, and the company will largely be dependent on Zero Trust (Act 2) to drive their outsized growth going forward, highlighted by the following remarks the CEO made at their FY’22 investor day: "Act 2 is really the Zero Trust services. That will be the big driver of revenue for us over the next period of time as a significantly larger market than the application services market."

In our view, the ultimate success and traction of NET's Zero Trust offering is questionable and largely unproven. We are skeptical of management's ability to replicate their success with Application Services in the Zero Trust market for two primary reasons. 

1. Different GTM Motion: Cybersecurity (Zero Trust) companies generally have a significantly different GTM motion than NETs, best illustrated by the fact that NETs largest Zero Trust competitor, Zscaler, generates 93% of its revenue with channel partners vs. 13% for Cloudflare. Management has previously admitted that the channel was a weakness for NET but that they’ve since made significant strides in improving their channel presence, as they’ve highlighted the fact that Cloudflare has signed-up the majority of Zscalers channel partners. However, we believe that bulls fail to fully appreciate that channel partnerships take time to develop and bear fruit, as only a handful of channel partners drive the majority of Zscaler’s revenue. This is best highlighted by Zscaler’s disclosure that their "top five channel partners and their affiliates, in aggregate, represented 28% of our revenue for fiscal 2022, 34% of our revenue for fiscal 2021, and 40% of our revenue for fiscal 2020."

2. Still Unproven: Zero Trust is still a relatively new market for Cloudflare and likely not yet a significant contributor to revenue, as the company disclosed only 10% of contracted customers have adopted Zero Trust compared to 75% for Application Services at their investor day in 2022. Reiterating the CEO's comments from their 2022 investor day, they are largely betting on Zero Trust as the biggest driver of future revenue growth. "Act 2 is really the Zero Trust services. That will be the big driver of revenue for us over the next period of time as a significantly larger market than the application services market." We believe this is a risky bet for investors as Cloudflare’s Zero Trust offering isn’t recognized as a category leader by many of the industry analyst reports for market position or product strength. In addition, recent comments from Zero Trust competitors Zscaler and Palo Alto suggest that they don’t view Cloudflare as a formidable competitor. Obviously, most competitors would not speak highly of one another, but we still believe their comments speak to how they view NET's competitive positioning in Zero Trust.

Workers

We believe bulls continue to overestimate the intermediate impact of Workers on the financials, as management has made it clear that Workers won’t be a material contributor for "3 to 5 years." We highlight the CEO’s remarks on Workers from the Q2’22 earnings call, which shed light on the potential timing of Workers becoming a material contributor.

“Act 3 (Workers) for us, which I think will really start to hit in a material way around revenue in 3 to 5 years is really around Workers.” Matthew Prince, Cloudflare CEO @ Q3 earnings call 08/04/22

Timing

Because Cloudflare's Q3 ended on 9/30/22, whereas peers' quarters ended a month later on 10/31/22, we believe that timing positively impacted NET’s Q3 results, as they didn’t experience the same pressures peers reported feeling near the end of their quarter as they reported a month later, when the environment was less favorable. Below, we’ve provided quotes from the ZS and CRWD conference calls to further elaborate on the weakness they experienced late into their quarter.

We believe it is likely the weakness and elevated headwinds these best-of-breed cybersecurity peers felt at the end of the last quarter, which NET avoided due to timing, could potentially show up in Q4. Additionally, peers have been vocal that they don’t believe the typical Q4 "budget flush" will have the same benefit in Q4 as prior years due to customers tightening up their budgets and requiring the C-suite to sign-off on large deals.

Finally, and most important to our thesis, management will need to issue FY’23 guidance on the Q4 call, and we believe their guidance will underwhelm investors as it is unlikely management can raise their guide beyond current street expectations. The best case scenario we can think of is that management maintains Q3’s sequential growth rate of 8.25% over the next 5 quarters, which would only produce 2.4% upside to FY'23 street estimates. This would assume that the macroeconomic picture doesn’t get worse and the company continues its run-rate growth trajectory. If management does indeed raise their guide and assumes linearity, our modeling suggests that would leave no upside for beats or raises, thus breaking the beat ‘n’ raise story, unless revenue reaccelerates in a slowing economy. Alternatively, if management wants to leave room for upside to estimates in order to keep their beat 'n' raise story going, although muted compared to historic beats, they would not be able to raise guidance above street expectations. Either scenario will, we believe, result in a lower stock price. We think it is noteworthy that NET has sold-off every Q4 print since coming public, which we believe is due to management not raising the forward FY guide above buyside expectations, which is evident in the stock’s next day reactions, including a -9.5% (Q4'21), a -5.8% (Q4'20), and a -1.8% (Q4’19).

Competitor Q3’22 (or equivalent) Comments on EoQ Weakness:

Risks

Increasing Price:

On November 29th, 23 days after the Q3 earnings call, NET Co-Founders Matthew Prince and Michelle Zatlyn, co-authored a post on Cloudflare’s website informing customers of an upcoming 25% price increase for Pay-As-You-Go customers, which account for approximately 15% of revenue and are primarily on a monthly billing plan. The deadline for conversion to annual plans in order to avoid the price increase is January 15th, 2023. Management comments suggest the increase was due to a need to improve cash flow timing by billing annually upfront instead of monthly in order to better align cash collection with the timing of their product investments. We believe improving cash flow is one driver of the price increase, but if one were to take a more cynical view, we think it could also be viewed as management pulling a lever to achieve FY’23 street estimates. We find the timing of the announcement curious, as they did not provide any indication of an upcoming price increase to analysts on the earnings call just 3 weeks earlier. Our math suggests that the ultimate benefit of the price increase from customers acquired before YE’22, could provide an incremental $14.6m benefit to FY’23 revenue. Below, we highlight key items related to the price increase. The risk is if our math proves completely incorrect and the price increase has a more positive impact on FY’23 revenue than we anticipate, thus allowing management to raise their outlook more than we expect. This could be the case if more than the 40% of pay-as-you-go customers we’ve allocated choose to stay with the monthly plan, in which case they don’t get to keep the old pricing and their pricing increases by 25%. We doubt this is the case, as management previously stated they expect most of the customers to convert to annual subscriptions. We’ve accounted for the price increase in our model by significantly increasing ASPs over the course of FY’23 for both existing and new <$100K customers.

• Price Increase Math: FY’22 Revenue Consensus Estimate of $975m x 15% of total revenue is pay-as-you-go revenue ($146.2m) x 40% of pay-as-you-go customers that don’t convert ($58.5) x 25% price increase = $14.6m incremental revenue.

Consolidation:

Management has made it clear they will use their "margin as a weapon" to compete, as there is little incremental gross margin impact to bundling various products together in order to get deals done. Management has said they are willing to bundle products and/or offer certain products for free as add-ons to compete with Zero Trust peers. From what we’ve gleaned from customer reviews and expert transcripts, Zero Trust products from most vendors are expensive, and in an environment of tightening budgets, customers may be more motivated to consolidate their security spending in order to realize cost savings. Intuitively, it would make sense for the low-cost vendor to benefit from this, and some might falsely equate those benefits as being realized by Cloudflare. According to our findings, customers are more likely to consolidate their security spending with their existing network security providers, or CSPs, because they can realize more cost savings by consolidating with the vendor with whom they have the most spend vs. a vendor like Cloudflare, where the average ASPs of existing customers are likely lower than that of a Palo Alto or CSPs, who can likely offer more cost savings on a larger bill. However, if we are wrong and customers decide to consolidate their cybersecurity spend with Cloudflare, there is likely upside to our estimates.

Take-Private:

We don’t really think there is much of a take-private risk with Cloudflare due to a recent CNBC appearance with CEO Matthew Prince. When asked bluntly by a CNBC anchor if he would be open to a sale, Mr. Prince's answer was simply "No." However, as SaaS companies are trading at historic discounts, take-out risks are more broadly elevated.

Financials & Forecast

We drive our forecasts using an Annualized Revenue model, as that is the metric disclosed by management in lieu of the typical ARR disclosure by most cybersecurity peers. Management simply multiplies quarterly revenue x 4 for a run-rate revenue annualized metric. Other provided disclosures include Gross Retention, which is only disclosed as being above +90%. Management also provides Dollar-Based Net Retention, which was 124% (as of Q3’22).

Annualized revenue is more appropriate than billings or RPO because contracted revenue is recognized 80% monthly vs. 20% annually. Additionally, approximately 15% of the business was pay-as-you-go as of Q3’22. We forecast each customer segment separately >$100,000 and <$100,000 to more accurately reflect the economic characteristics of each customer segment. >$100K Customers contributed 61% of Annualized Revenue in Q3’22. We anticipate the percentage of revenue derived from >$100K customers will continue to grow as NET continues to move upmarket and focus its GTM on large enterprise customers. The remaining 39% of Revenue is from <$100K customers, who are primarily mid-market and SMB customers. As mentioned, the company discloses that Gross Retention is above 90%. Management recently highlighted that they experienced some SMB customers downgrading to their free tier; however, SMB represents a relatively small part of their business, so we assume stable churn as the company offsets SMB weakness with continued strength in the enterprise. The company’s DBNRR continues to decelerate even as management reiterated confidence in achieving their longer-term target of 130% DBNRR. We model DBNRR declining from 124% in Q3’22 to 122.8% in Q2’23 before modestly recovering in 2H23 to account for macro headwinds and longer deal cycles. Where we think the weakness will show-up in the financials is Net New Customer Revenue. Our modeling suggestions Net New Customer Annualized Revenue declined Q/Q by approximately -5.4%, this weakness in new logo revenue is consistent with recent comments from management outlined below. We forecast a continued decline in New Customer Revenue before starting a gradual recovery in Q4’23.

New Logo Slowdown:

DBNRR Weakness:

Revenue Model:

Our revenue estimates are slightly above street estimates. Typically, if our modeling is in line with street estimates, we would simply move on. However, Cloudflare’s stock is driven by a beat ‘n’ raise story where the real "buyside bogey" is much higher than street estimates due to management's history of sandbagging guidance.

Annualized Revenue Model:

Our Annualized Revenue Model is a rollup of our sub-Annualized Revenue Model’s for both customer segments >$100K and <$100K, which are included below. We assume DBNRR continues to decline modestly from 124% in Q3’21 to 122.8% Q2’23 due to tightening budgets before gradually starting to recover in 2H’23. See the management quotes above for their commentary on softness in DBNRR trends. Our Gross Retention remains at 90% as SMB customer churn is offset by an increasing mix of large enterprise customers with stronger retention rates as the company continues to move upmarket. This is further supported by recent management commentary on continued strength in gross retention. In addition, our model assumes a continued decline in New Customer Revenue. This figure is calculated and not reported by the company. New Customer Revenue is calculated as (Net New Revenue – Existing Customer Revenue (DBNRR)). Management has called out recent weakness in net new logos and we believe this trend will continue before beginning to gradually recover in Q4’23. See the management comments section above for their thoughts on new logo trends slowing. 

>$100K Customer Annualized Revenue Model:

Net New Annualized Revenue per Existing >$100K Customers continues to modestly decline, driving down >$100K Customer DBNRR from 134% (Q3’22) before bottoming in Q2’23 at 128.7%. Please note, the company only discloses DBNRR at the company-level, so all customer-level DBNRR figures are calculated. We calculated >$100K DBNRR by applying a 10% increase to the company-level DBNRR as management had disclosed that large >$100K customer-level DBNRR is higher than their company-level DBNRR. Additionally, we model continued weakness in Annualized Revenue per New >$100K customer to account for continued new logo weakness and smaller large customer lands. We believe this will stabilize and start recovering in 2H’23.

<$100K Customer Annualized Revenue Model:

NET’s price increase provides an incremental benefit to New New Annualized Revenue Per Existing <$100K customer revenue, thus providing an incremental boost to <$100K Customer DBNRR, going from 111.9% (Q3’22) to 114.4% (Q1’23). Please note, DBNRR is only provided at the company-level, so all customer-level DBNRR metrics are calculated. Annualized Revenue from New >$100K Customers also benefits from the prince increase which we account for in rising ASPs through FY’23.

Customer Count Model:

Net New Customer Adds >$100K began to decline YoY in Q3'22. We model continued lower >$100K customer adds before beginning to gradually recover in Q4’23. For Net New Customer Adds <$100K, we assume continued modest softening before beginning to reaccelerate in Q4’23. Our modeling accounts for continued macroeconomic weakness before a gradually recovering in 2H’23.

Netcraft Web Server Survey Data:

On January 27th, CEO Matthew Prince tweeted “Not slowing down” in reference to a 3P data report from Netcraft called “Web Server Survey” which provides market share and operational tracking metrics. At first, the 3P data report looks overwhelmingly positive, with snippets such as:

Link to Web Server Survey Report: https://news.netcraft.com/archives/2023/01/27/january-2023-web-server-survey.html

The report provides 4 key data points related to Cloudflare, including All Sites, Active Sites, Domains, and the Top Million Busiest Sites market share. We believe the CEO and NET bulls were largely focused on the impressive All Sites metric, which highlighted Cloudflare’s jump from 3rd place to 1st in terms of All Sites market share. This metric appears to be outstanding on the surface, with All Sites growth accelerating to 84% year on year. However, we think this metric is misunderstood because it includes sites that are duplicate or dormant. We believe Active Sites is the more accurate measurement for understanding the underlying business momentum because it excludes inactive sites, which are included in the All Sites metric, from our understanding.

The other 3 metrics Active Sites, Domains and the Top Million Busiest Sites, showed decelerating growth in Q4’22 per our modeling of the data (see below).  Active Sites decelerated in Q4 and averaged 12.5% YoY growth vs. 17.1% YoY growth in Q3. Domains also began decelerating, with domain additions averaging 16.5% YoY growth in Q4 vs. 19.4% YoY growth in Q3. Active Sites did re-accelerate to 22% YoY in (Jan '23) from 10% in (Dec. '22). I believe the re-acceleration in January could have been due to customers buying ahead of the price increase; however, it is hard to say with any certainty. Domains did slightly re-accelerate to 18% YoY (Jan. '23) vs. 17% YoY (Dec. '23). Top Million Busiest Sites experienced YoY deceleration in Q4, growing an average of 14% YoY in Q4 vs. 15.4% YoY in Q3, but significantly slower than the average 20.4% YoY growth in Q1. Below, we highlight in red the Q4 Web Server Survey data.

Netcraft “Web Server Survey” Data:

Additional 3rd Party Data:

Below, we provide additional 3rd party data from W3Techs in order to directionally verify the trends in the Web Server Survey data, and we found the W3Techs data also decelerated in Q4.

Link: https://w3techs.com/technologies/history_overview/web_server

Link: https://w3techs.com/technologies/details/ws-cloudflare