“The Room Where It Happens” is one of my favorite songs from the excellent musical Hamilton that talks about how Alexander Hamilton, Thomas Jefferson, and James Madison supposedly negotiated in secret to establish what is now the U.S. financial system. And there are a great deal of books and movies that glorify how people in powerful positions set out to decide the future of humanity, wars, and radical changes in secret rooms.
When it comes to business, databases were for a long time ‘the room where it happens’. In a typical IT system, companies build databases with hundreds of variables to capture and organize multiple use cases for one specific application, but with zero to little attention to the orchestration of the interdependencies of the entire IT stack. And as the complexity of the system increases, the business observes diminishing returns with newer datasets, be it by the lack of visibility of the available information to the right teams or the speed and accuracy of data.
In 2014, Confluent was founded to offer an enterprise-grade alternative to how data flows and applications communicate and, since then, the company has experienced rapid growth becoming one of the premier Cloud-Consumption Event Streaming Platforms.
CONFLUENT BUSINESS OVERVIEW
In the late 2000s, LinkedIn was experiencing scalability problems with its core database and the integration among its key systems, including its complex social graph and search engine. The pipelines among these databases (when available) were based on an architecture that transferred data in once-a-day of never-ending ETL (Extract, Transform, Load) data integration work that is ripped for disruption, errors and complexity.
Additionally, the lack of real-time data flow meant that many of the services could only be analyzed and updated after overnight data load, which is not suitable to a platform that was built for real-time data consumption of troves of different types of data.
When scouting the market for a solution that could offer real-time continuous and easily scalable data flows, the LinkedIn infrastructure team couldn’t find a solution and decided to build one from scratch — and named after Kafka because one of the creators (and now CEO and co-founder of Confluent) liked the famous author.
Once data feeds were past delayed ETL processes to become nearly real-time processed, teams started asking that Kafka should be added to new services including alerts, errors logs and performance monitoring of every system.
Overtime, Kafka became a vital component of LinkedIn IT Infrastructure to ingest, process and flow data among the entire organization. As its maturity level increased, LinkedIn open-sourced it to Apache Foundation in 2011 and graduated to production in 2012.
With the successful graduation of the open-source technology the core Kafka developer team left LinkedIn to found Confluent in 2014 and sell it as an enterprise-ready service and built a rapidly growing business over the last seven years.
The company offers cloud-native and self-hosted options, with the latter plan aimed for enterprises with dedicated and experienced resources to work with Kafka. And, just like Snowflake bills their customers on a pay-as-you-use (instead of per-seat), Confluent belongs to the ‘Cloud Consumption’ cohort, a rising trend that offers the following benefits to consumers:
- The ability to pay for customized offerings, avoiding the need to provision resources for peak usage
- Buyers have greater transparency to correlate usage and cost, with added flexibility in trimming or expanding services and use cases based on consumption.
- Allows for a cheaper entry price point to explore the solution before committing to additional services and scale
- Remove contract duration and upfront fees friction when negotiating large contracts
The latter two points are the foundation to which cloud-consumption-based pricing experiences high growth rates and elevated retention rates for long periods of time.
Companies in the consumption-pricing model cohort deliver, on average, 16% higher Net Retention Rates (NRR) then the broader Cloud publicly traded companies (n= 9 and 60, respectively). And this is not different to Confluent, who has experienced 117%+ NRR over the last couple of years, with its service being used to ingest, transfer and analyze larger data sets within large enterprises.
Another data point is the comparison of Average Contract Value over the reporting periods in the prospectus, which will be covered in the Financial section.
Before that, I will cover how the leadership team and board of directors is structured.
The ‘Best Ranked Cloud Computing Companies’ trade at a 6.3x (20.1x vs 13.4x) higher Enterprise Value/Sales multiple than the average cloud firm according to research from Battery Ventures and Glassdoor. This is one indication that there may exist a ‘talent flywheel’ effect in Cloud companies that spins when the founding team defines a culture that incentivizes, for instance, excellence, teamwork, and diversity. This allows for sourcing of new hires that drives performance and new learning and growth opportunities.
Confluent boosts a 4.5-star rating on Glassdoor with a 90%+ CEO approval, placing the firm among the best Cloud-based workplaces.
Video presentations offer a relevant perspective on how a person communicates and delivers its message, from leaders who have a knack to inspire audiences with lively anecdotes and changes in voice tone and speed, to others who cater to technical viewers with a linear note. Co-founder and CEO Jay Kreps clearly mastered his skills overtime, from a 2-min pitch talking about his experience at LinkedIn 10 years ago, to a technical Kafka deep dive five years ago to a more recent (and arguably more sophisticated) company presentation.
As a leader in the Business-to-Developer (B2D) space, his style skews heavily to be the face of Kafka to the open-source community while selling Confluent to their customers, who are likewise in engineering and technical roles.
For a deep dive in the founding culture of the company, Jay has shared a piece in 2014 that is well worth the read.
In total, management has 8 members, three of whom are women, including the recent appointment of the former Global CMO of Salesforce Stephanie Buscemi to lead the marketing function. Another recent hire is Steffan Tomlinson, former Google Cloud CFO, who brings a wealth of experience in leading private companies through their IPOs (Palo Alto and Arista) and in cloud consumption models.
The Board of Directors has eight members, being two co-founders, three VCs and three independent members, including the former CEO of MuleSoft (another open-source cloud company based that went public and was subsequently acquired by SalesForce) and Lara Caimi, Chief Customer and Partner Officer at ServiceNow.
This is a balanced board — with operators and investors with prior experience in running and investing in cloud and open-source businesses — that was put in place to help Confluent transition to the public markets. There is, though, space to improve women representation, with only two seats (25%) in the room where it happens, especially when one compares Confluent to other recent high-profile IPOs of Coinbase (28% of women in the board), Snowflake and UIPath (30%) and Compass at 33%. Confluent also lags in women representation when compared to companies in the S&P 500 index that have on average 3.3 women out of 11.2 seats (29%).
The Financials section below will compare some of the metrics that investors analyze to gauge the growth and stickiness of the business.
Even with the headwinds the Covid-19 pandemic brought to all of us, the average publicly traded cloud company posted Last Twelve Month (LTM) revenue growth of 42%, and 62% for the consumption-pricing cohort. Confluent recognized LTM growth in excess of 50%, which would make the company in the top decile of the 60 publicly traded cloud businesses.
The Magic Number (or Sales Efficiency) measures the effectiveness of Sales & Marketing expenses in improving Gross Profit. And to quote Tomasz Tunguz:
“If a startup invests $500k in marketing and sales this quarter and generates $1M in incremental revenue, net of the cost to provide the service, for the next 12 months, the sales efficiency would be 2. The inverse of the sales efficiency is payback period. In the example above, customer revenue reimburses sales and marketing costs in ½ year or about 6 months.”
In the case of Confluent, its Magic Number sits at 0.47 over the last eight quarters, with Payback period (in months) of 11.5 months for the same time span. The following table shows that consumption-based Companies are more effective than the broader cloud cohort in their Sales & Marketing motion, with 45% higher Magic Number and faster Payback Periods (14.1 monts for Consumption-based versus 21.5 for broader Cloud based companies).
These numbers are typically a function of the size of the average contract value (total revenue / number of customers). Consumption-based pricing companies have an ACV of $84k or 16% smaller contracts when compared to the average annual Cloud contract of $100k. In the case of Confluent, the ACV has decreased from $158k in 2019 (revenue of $130m and customer count of 820) to $99.3k in 2020. This 37% decrease in contract value with its top quartile Retention Rate is evidence that Confluent lands small contracts and expand its offering just like LinkedIn experienced with Kafka 10 years ago.
At $263m in LTM revenue, 117% in NRR and with growth rates in excess of 55%, I expect Confluent to start trading at a premium to its NTM EV/Sales in the range of 35x to 45x, implying an enterprise value between $13b to $20b.
Companies — small or large — are constantly searching for ways to remove barriers of data siloes to leverage the benefits of real-time data analysis. Ten years ago, a team at LinkedIn pioneered the Event Streaming space to move from batch to streaming systems to provide services in real-time. And what started as a pain-point for one company became the standard to 70% of Fortune 500 companies (136 are Confluent customers) now connect and enable the constant information flow across their organizations.
It will be interesting to watch how Confluent will perform in the public markets, but as the CFO of the company said when he joined the company:
“We look at an IPO as one step along the journey of creating value and having a durable growth model”. Or as Hamilton would put it:
This article is not intended to be used as financial advice. Go speak with your financial advisor and do your own due diligence. It’s worth it.