Description

Long: Datadog (DDOG)

Consumption software names, which are typically SaaS businesses with a usage-based revenue model, saw a significant revenue growth (and valuation) reset after the Covid years. The reset happened as software customers came back to focus on the optimization of spending on cloud services and related developer tools. After a rally in 2023 in these names, fears of persistent inflation and elevated rates have taken over the market narrative this year. Consumption software names have come under pressure again, as seen in their recent performance, including the drawdown this past week.

If you have a longer investment horizon and are looking for a high-quality software name with some idiosyncratic drivers, Datadog is one candidate to consider. Datadog is a platform play on vendor consolidation in observability space. Specifically, it provides a cloud-native and hyperscaler-agnostic full observability stack (with devOps as well as devSecOps tools) for SMB, mid-market, and large enterprises. I underwrite ARR and revenue growth over 2023-28E at a 27%+ CAGR. Given this volatile market, it’s likely not a straight-line path to the upside. I present a framework to think through key drivers whether you are a bull or a bear, so you can juxtapose your worldview over it.

Datadog Overview:

Observability:

• Datadog falls under the umbrella of developer tools dedicated to observability, a set of solutions that helps developers monitor and analyze the performance of their applications/websites and related infrastructure in real-time, ensuring optimal performance and identifying issues before they affect users.

• Essentially, observability is like an insurance policy for the performance of applications delivered on-premise, in the cloud/multi-cloud, or in a hybrid mix of on-premise and cloud.  
• It’s a ‘dev-ops’ productivity improvement tool that brings together silos of developer and operations teams.
• To illustrate an example use case for observability, if you are running Intuit/Turbotax during tax filing season, it is mission-critical to have the servers running. Any server downtime will be met with serious customer backlash and leaving you for competitors. Observability vendors provide you with logs/metrics/traces on distributed servers and applications. The goal is to alert you on potential bottlenecks and proactively resolve issues before they escalate to cause downtime.

Global public cloud services is estimated to be a $560B market in 2023, and per Gartner, is expected to grow at a 20% CAGR 2023-27E. Per Gartner, observability segment is <10% of this market, and Datadog’s 2023 revenue <4% of the observability market.

Observability takes different forms, such as infrastructure monitoring, log management, application performance monitoring (APM), and cloud security. Each of these subsegments can be more important for one use case than others.

• Infrastructure monitoring: if you run a large language model training or inference software platform, back-end support in terms of infrastructure monitoring becomes critical. For example, monitoring logs/metrics/traces across servers, networking, and storage becomes important for ensuring efficient compute resource utilization for training as well as ensuring low query latencies for inference.   
• Application performance monitoring: If you run a business, say Instacart, where a large portion of revenues come from an app, APM mobile monitoring is critical for ensuring high shopping cart conversion rates and customer satisfaction.
• Security monitoring: For a large bank like JPM Chase, fraud alerts and compliance monitoring necessitate robust real-time security solutions that need to be coupled with the rest of the observability stack.

Industry context:

• Software development has undergone significant changes over the last several decades. What started as monolithic architectures have now evolved into modular ones with the advent of containers and microservices. Decades ago, when hardware costs were high, software architectures were designed to fit within one large server or mainframe. This was far from agile --- hard to implement, scale, and debug, often resulting in large downtimes that severely dented business operations.
• Thanks to developments in the semiconductor industry, the cost of compute and storage declined rapidly, per the famous Moore’s Law. As a result, the software industry witnessed the rise of virtual machines (multiple instances on a single physical server) and thereafter, distributed cloud infrastructure. Containers and microservices were the next step in the software industry's evolution to develop modular and scalable solutions tailored to specific use cases. This was necessary given the exponential rise of data generated across multiple domains. An accelerated pace of development and shorter innovation cycles have compounded the complexity of software solutions across multiple access points (cloud and edge), formats (desktop, tablet, mobile), and mediums (text, audio, video), making observability a critical developer tool.

Business/product 101:

Datadog entered the observability market with a solution for back-end infrastructure monitoring, targeting developer teams. The short time to value, frictionless pay-as-you-go subscription model, and product-led growth (developers don’t need to contact sales reps until they reach large spend levels) model --- all synced up for Datadog. The sales motion was smooth – the land and expand were split between two teams, one focused just on land and another on customer success (expand).

• Fast forward to today, Datadog offers 19 products in one integrated platform
• As of Sep 2023, Annual Recurring Revenue (ARR) reached $1B for Infrastructure Monitoring, $500M for log management, and $500M for APM
• Usage-based pricing model that starts with a free tier that has basic features. Pricing is on a per-host basis (~$30-40/host/month for APM), with some products (e.g. logs) priced to data ingested (e.g 10 cents/GB/month)
• About 30% of revenues come from outside the US
• SBC is high at 20%+ of revenues. Total annual dilution related to RSUs/PSUs awarded: 2.5-3.5% of share count
• Generative AI exposure: Datadog has exposure via AI-focused players, including OpenAI. Datadog enables OpenAI users to track token consumption and usage costs associated with OpenAI API

Competitive landscape:

Most of the market relies on ad-hoc solutions with legacy monitoring solutions focused on silos in the observability stack. Some key competitors are:

• Dynatrace: leader in Application Performance Monitoring (APM). Usually targets C-suite, large deals, and annual contract values (ACV)
• New Relic: legacy player that started with APM, taken private by PE players: Francisco Partners and TPG in 2023 after having lost significant share to Datadog
• Splunk: log and security specialist – acquired by Cisco in 2023
• AppDynamics: the original APM player that was acquired by Cisco in 2017 right before its planned IPO. (Cisco tried to acquire Datadog for $7+ Billion in 2019 but Datadog rejected the offer)
• Open-source solutions such as Grafana and Chronosphere are great for niche parts of the stack if you have the technical expertise to deploy these solutions. 

Why it could be interesting:

1)     Datadog is a beneficiary of platform consolidation across its user base. Datadog offers a unified platform/single pane of glass for 19 products including infrastructure monitoring, logs, and APM. It has 700+ vendor-backed turnkey integrations that help quick time to value, test, and adopt. Switching costs are likely high given the comprehensiveness of the product stack, scale of data involved, cross-vendor integrations, and mission-criticality of the underlying use case.

• Datadog has maintained mid-high 90s gross revenue retention despite a dynamic and competitive environment over the years. This comes from a mix of 98-99% for enterprises, high 90s for midmarket, and mid-90s for SMBs.
• Management cited in 2024 Investor Day that of the $1M+ deals closed in 2023, about 55% were consolidations away from competitors. Also, the consolidation deals are cited for 4x higher annualized bookings growth than deals not involving consolidation.
• A supportive data point of the platform consolidation thesis is illustrated below based on the steady increase in multiproduct adoption of Datadog’s product across its customer base, shown below in Fig. 1. As (shown for Q1 2022 and Q4 2023 in Fig. 2), which, as shown in Fig. 3, has rebounded from the cyclical lows of Q2 2023.

Fig. 1: Supplementary data from Datadog. Source: Q4 2023 Filings

 Fig. 2: My analysis of the mix of customers using 2+, 4+, 6+, and 8+ products for the last 2 years provided by Datadog in Fig. 1. Over the last two years, adoption of 4+ products increased from 35% to 47% of customers

Fig. 3: My analysis showing Annualized revenue per customer per estimate of min. avg products adopted. This analysis shows that this metric seemed to have bottomed in Q2 2023. On average for about 3.8 products adopted per customer and $2.6B annualized 4Q 2023 revenue and 27k customers, revenue per customer per product comes to ~$23k per product per customer. To arrive at the weighted average estimate for product adoption, for simplicity, I assume 1, 2.5, 4.5, 6.5, and 8 for the average number of products for the respective customer segments of Fig. 2

 2) Cyclical rebound: There are two layers to the cyclical rebound – first, you have customer optimizations of spend on hyperscaler cloud usage, e.g., AWS and Azure. On top of that, you have optimization of spend on consumption software such as Datadog. While it is hard to predict exactly when the respective usages hit their lows and the lag between the trough levels, consensus estimates are in line with the view that AWS revenue growth hit its low at ~12% y/y growth in Q2 2023. The associated likely pick up in cloud migrations has a positive correlation with observability spend, especially application performance monitoring solutions, which can be a beneficiary.

To juxtapose that context, Datadog is lapping tough comps from Q1 and Q2 2023, which were at the worst of the cost optimization seen post-COVID, as shown below. Management guided for Q1 2024 to be sequentially ~flat.  

Fig. 4: DDOG Quarterly Revenues

 3) Secular opportunity: Of a total addressable market of 500k+ global accounts, Datadog reports 27k customers, which is less than 5% penetrated. This bodes well for a new logo adoption runway.

4) An idiosyncratic opportunity for Datadog stems from adoption dynamics within its existing customer base, without having to rely on new customers per se.

• Only 40% of the customer base has adopted the three main products: core infrastructure monitoring, log management, and APM. The cross-sell opportunity is also significant because customers who adopt all three main products tend to spend 10x that of the rest of the user base.
• Security provides a cross-sell opportunity for the current customer base as the security product penetration is low:
  • Infrastructure Monitoring: 8%
  • Application Performance Monitoring: 14%
  • Log Management: 19%

DevSecOps is a growing field in the software industry. Datadog’s approach to increasing security product usage should involve buy-in from CISO teams and bottom-up adoption from developers, as they are close to security needs for their applications/back-end solutions/logs.

5)     Related to some of the above thesis points, Datadog stands to likely benefit from ongoing changes in the competitive landscape as the logs/SIEM peer Splunk is acquired by Cisco and as New Relic transitions to PE hands. A similar dynamic occurred before when Datadog gained a share at the expense of legacy provider AppDynamics after its acquisition by Cisco in 2017.

• Datadog introduced new products such as flex logs to balance out accessibility and cost (index only when a user opts for it), which essentially give users control and flexibility. This differentiated strategy helps datadog win share in a crowded log management market.

 ARR-based Valuation framework:

There are several ways to slice and dice valuation, including multiples and DCF models, given management’s 2024 Investor Day guidance to long-term EBIT margins of 25%+. At IPO, the company had provided long-term goals for 77-78% gross margins, 22-26% R&D margins, 23-27% S&M. and 6-7% G&A --- which they now condensed to a 25%+ long-term EBIT margin. But you must reckon with SBC at a high 20%+ share of revenues, similar to other high growth software names, there are no two ways about it.

 In 2024 Investor Day, management provided qualitative drivers to think through the growth algorithm for ARR. In Fig. 5 below, I present an ARR build-up based on consensus revenue estimates for 2023-28E. The key drivers include: NRR for the existing customer base (which is composed of new product cross-sell and existing product expansion) and ARR added from new logos (contract value x number of new customers added). There are a lot of moving parts here – the intent is to provide a framework to think through the drivers. Consensus is modeling a 24% revenue CAGR for 2023-28E, which, at current levels of ARR from new customers, bakes in Net Revenue Retention (NRR) in the 110-125%, lower than than 130%+ levels Datadog had seen prior to the post-COVID reset.

Based on our views on company management, a typically conservative CFO, and given the cadence of products in the new growth segment of Security and secular opportunity for adding new logos, it does not seem to be a stretch to build on this model to underwrite a 27%+ (bull case being ~30%) revenue CAGR for the period 2023-28E. If the company were to achieve 30% FCF margins (they posted 28 % FCF margin in CY 2023), that could imply a potential for ~$5+ FCF/share by 2028E.

 Fig. 5: Framework showing ARR-buildup using drivers based on existing customers and new logos specified in Datadog’s 2024 Investor Day. There are a lot of moving parts, but this framework could be useful to impose your views on key drivers. Datadog doesn't report ARR per se but provided historical graphs in the recent Investor Day deck.

Risks:

1)     Elevated interest rates and high-growth software are a recipe for violent drawdowns if expectations are not met. If you have a view on rates, it could help to balance out the exposure via ‘shorts’ of similar high-growth names.

2)     Datadog’s push into security may take longer than expected to scale.

3) If competition from Dynatrace increases as Datadog moves upmarket to larger enterprises, revenue growth could be lumpy, and top-line prints may miss expectations.

Catalyst

Potential Catalysts

1) Increased adoption of Security products across its existing customer base

2) Cyclical: Continued easing of cloud optimization trends of existing base

3) Cyclical: increased pickup of new customers and workloads as the market emerges from cloud cyclical correction

4) Secular: monetization of observability products as developer productivity increases with software industry’s broader adoption of generative AI

5) If Datadog were to continue gaining share in the observability market, e.g. from Splunk and AppDynamics of Cisco, and New Relic that would be positive to our thesis.