Description

Summary: We believe that Couchbase is a Tier 1 database asset that is making a transition to the hosted model that should re-accelerate the company's growth and make it better/more efficient business. The transition has been proven out by MongoDB, ElasticSearch and Hashicorp. If the company executes the transition, the upside is >25% IRR with downside protection given the current valuation (~3.2x Revenues).

Company and Product Background:

Couchbase is an opensource database company. The company’s core technology is a distributed, NoSQL database. NoSQL databases were designed to serve as modern alternatives to the relational databases (“RDBMS”). These legacy databases were designed for a pre-internet world where there was less user interaction and unstructured data. The key benefits of NoSQL databases are that they can… 

• …manage unstructured data (e.g., text, video, and audio files). Relational databases can only manage “structured” data (small data in table form). 

• …support dynamic schema. Traditional relational databases require developers to define their schemas upfront and make it very hard to change. 

• …scale more efficiently. 

Couchbase has three products: Couchbase Server, Couchbase Mobile, and Capella. Couchbase also offers a backend services (e.g., Sync Gateway, Analytics) that help developers more quickly build applications.

• Server is the company’s core database product and Mobile is version of Couchbase’s database product designed for mobile and edge devices. Both the Server and Database products are deployed and managed by the Couchbase’s customers. 

• Capella, launched in 2021, is Couchbase’s newest product and the company’s first managed offering. The managed offering is hosted and managed by Couchbase on behalf of the customer.

Couchbase monetizes these services through Support & Services, Licenses, Professional Services, and Hosting Fees. Server and Lite are primarily monetized via by the sale of term licenses and support fees. Capella is monetized via SaaS fees. 

• Licenses (~12% of revenues / 90%+ Gross Margins): term licenses for the server product.

• Services (~8% of revenues / 12-15% Gross Margins): traditional professional services fees.

• Support & Services (~76% of revenues / 90%+ Gross Margins): For the Server and Mobile products, customers deploy Couchbase on their owned and managed infrastructure (cloud or on-premise). Customers then purchase support services from Couchbase to help with installation, troubleshooting and maintenance of the servers. Support customers also receive the latest security and bug updates. The value proposition of the Support functionality is “peace of mind” and technical support.

• Capella (~2-3% of revenues / ~70% Gross Margins at scale): Capella is the company’s “hosted” offering, which means that Couchbase will host and manage the server product for customers in the public cloud. The core value proposition is (1) speed to deploy, (2) ease of use, and (3) developer experience. Deploying and managing Couchbase has a steep learning curve, the hosted offering obviates that need and pushes the burden onto Couchbase. 

The majority of Couchbase’s business is support and services because until very recently they did not have a hosted product. Even though it’s a high gross-margin business, the support model has a few drawbacks:

• Couchbase has no telemetry or into the usage / consumption patterns with their customers. This makes unlocking expansion or new workloads hard for the company’s GTM team.

• Without a hosted offering, database administrators (“DBAs”) become a bottleneck for new deployments. Instead of developers being able to spin-up instances of Couchbase themselves, customers need to rely on DBAs to provision new servers before developers can start app development. 

Go-to-Market and Competition:

Couchbase and its main competitor, MongoDB, were both founded around the same time (mid/late 2000s). Couchbase optimized its product and go-to-market to target DBAs and large enterprise customers. MongoDB pursued a bottoms-up developer focused strategy. MongoDB did so in part because of their inability to compete with Couchbase up-market. Couchbase doubled down on their seaming advantage and in the mid-2010s, Couchbase made a bet that next leg of growth for the NoSQL market was going to be migrations of traditional RDMBS to NoSQL and companies re-platforming existing applications i.e., rip-and-replace.

Targeting enterprise and the rip-and-replace market had a few advantages:

• Large ACVs ($250K+) and early penetration into the Global 2000 customers. 

• The product is incredibly sticky and can expand meaningfully within the customer (net retention is >125%)

• Couchbase was able to focus on building on an enterprise grade product that was built for enterprise scale. On the otherhand, MongoDB optimized for developer experience and speed to market for companies at the expense of enterprise scalability.  

The disadvantages of Couchbase’s strategy were:

• Sales cycles were incredibly long. Rip-and-replace decisions take years to nurture and close. 

• Incentivized Couchbase to not focus on the developer persona and cloud/hosted opportunity. Enterprise customers and DBAs that Couchbase was making all its money from did not commit to cloud first platforms until the late 2010s and like the “centralized” buying that Couchbase enabled. 

Due to a combination of poor GTM decisions, lack of hosted product, and poor GTM leadership…

• ....Couchbase was not able to capture the rapid growth in the NoSQL database market the way MongoDB did. 

• ….has had a more inefficient GTM compared to its SaaS/Hosted peers. 

All that being said, the core product and technology is best-in-class. Based on customer and former feedback, it’s clear that Couchbase’s core database product is more scalable and performant than MongoDB and the cloud vendor’s solutions at enterprise scale (e.g., Global 2000 customers). The product advantages and differentiation are because of Couchbase's roots as a merger between Membase and CouchOne, and the early focus on enterprise.

Couchbase has a product that…

• …can handle much higher volume of data and throughput than all of its peers.

• …replicates data across multiple databases and data centers automatically and more efficiently. One of the key value propositions of NoSQL databases is the ability to “shard” data across multiple databases.

• …can support very low latency workloads. This is enabled by the company’s memory first architecture. 

Performance benchmarks from a Altoros, which demonstrates that Couchbase’s…

• …latency (i.e., response time for a request) is 8-10x lower than MongoDB

• …the throughput (i.e., number of transactions processed per second) that Couchbase can candle is meaningfully higher than MongoDB.

When Couchbase is in an evaluation vs. MongoDB, the company has a >50% win-rate. The challenge however, is that Couchbase did not enter nearly as many evaluations as it should have because MongoDB has dominated developer mindshare and missed out getting into net new applications. 

Capella/Hosted Opportunity:

Capella is Couchbase’s attempt to take their high quality NoSQL database and offering it in a hosted and easier to deploy/user environment. When setting up a power database like Couchbase there is a lot of configuration that goes into self-hosting the database and have it perform. Most companies want to offload that responsibility to a trusted 3rd party that understands the database and enable their developers to focus on just the business facing applications. Capella should enable Couchbase to (1) unlock workloads within existing customers more efficiently and faster, (2) make it easier for new customers/developers to adopt their product, and (3) unlock a more efficient GTM motion.

Couchbase has not broken out the revenue mix of Capella vs. Server yet but we estimate that the Capella product contributed ~$1.5-2M of revenues in the most recent quarter ($6-8M ARR). Company said they will start disclosing Capella metrics once they have a better understanding of the consumption patterns. However, the company has provided the following datapoints:

• Capella consumption patterns are trending higher than legacy Server customers i.e., NRR for Capella cohorts is greater than non-Capella customers. This is in-line with the commentary from Couchbase’s peers that have already launched hosted offerings. For example, we estimate that today, MongoDB Atlas NRR is 10 percent points higher than core MongoDB.

• ~15-17% of customers are using Capella, which implies ARR/Customer of ~$75-90K. The ARR/Customer is roughly in-line with MongoDB’s ACVs for outbound customers.

• We know there’s latent demand for a cloud version of Couchbase given that >50% of Couchbase’s Server customers are deploying Couchbase in the cloud themselves. 

Couchbase's open-source peers, MongoDB, Elastic, and Hashicorp have all launched hosted/managed versions of their open source products, which have served as a second leg of growth for the companies. After crossing $10M ARR, all three companies’ hosted products scaled or are on a path to $100M+ ARR in three fiscal years. In our upside case, we believe that Capella can achieve a similar trajectory.

Forecast and Valuation:

In our base-case, we forecast that Capella is only able penetrate ~40% of the customer base with NRR going from 125%+ today down to 120%. In our upside case, we forecast that Capella penetrates ~70% of Capella’s customer base and maintains NRR in the 125% range, which implies that in FY 2028 the company is generating $91M of Capella revenues and ~$300M of server revenues. 

Our valuation assumes that the server revenue stream warrants a 3.5x revenue multiple and the Capella revenue stream warrants a 6x in the base-case and 8x in the upside case, which is a discount to long-term revenue multiples of the company's open-source database peers at similar growth rates. Net-net this implies a return profile of ~16% IRR over ~4 years in the base-case and ~25% in the upside case. 

Risks:

• Execution on the core cloud product. Making the transition to "hosted" offering is not easy by any means and most companies have starts/stops because there is a degree of re-architecting that's required. 
• The key question with Couchbase is not whether Capella represents a meaningful growth opportunity or not but rather whether this management team can execute the strategy. Past performance suggests that the team is not well positioned to execute this transition.
  • For one, the company was incredibly late with the Capella product. The GTM leadership team does not have the muscle to market to developers and sell cloud native products. Couchbase’s CEO, CRO and CMO all come from the “on-premise” world and have very little developer tools experience.
  • The CMO’s leadership experience is primarily from HortonWorks, Cloudera, Docker and MarkLogic, which are all database/infra companies that failed to make the leap to cloud-native world.

Catalyst

• Disclosure of Capella metrics @ analyst day