CFLT - Confluent: Can Data In Motion Move Fast Enough To Escape Macro Headwinds?

2023-04-21 12:26:07 ET

Summary

• Confluent is a leader in providing commercial-grade solutions based on the Kafka open-source project.
• Basically, Kafka is a foundation technology for what is called stream processing.
• Stream processing is one of the foundation technologies that allow for the development of digital transformation applications that allow for positive user experiences.
• Confluent is not immune to macro headwinds, and its forecast reflects some caution about both consumption of its cloud solution and the growth of its platform product.
• That said, it is pivoting to profitability and free cash flow generation, both of which are expected in Q4 of this year.

Looking beyond the valley

It has been 18 months now since the investment climate for investing in the high growth IT space has been sunny and warm. It has been a challenge to own a portfolio of such companies over that time, and at least as challenging to write analysis of the state of the software business. That said, there is not some existential correlation between stock prices, the evolution of technology, and the business opportunities in enterprise software. Many software buyers have pulled back from making large commitments, trying to deal with headwinds buffeting their own businesses. Fed policies, besides doing a number on valuations of high growth software companies, have also destroyed a great deal of demand, not just for enterprise software, but for homes, and vehicles and much else besides. But Fed policies, geo-political upsets and a generalized feeling of malaise coupled with some fears of the future have not stopped technology trends even in terms of cloud migration, and that brings me the topic for this article, a discussion of the investment merits of Confluent ( CFLT ), once a highly valued name, now with a much more pedestrian valuation.

Confluent was a new issue back in the summer of 2021 - a very different time in the market. I doubt if an IPO of the shares would even be possible in the current environment in which growth has taken a back seat and investors are far more concerned about many macro factors, and are worried how companies will perform in a recession. Back then, with the shares at $43.50, I was unwilling to make a purchase call based on valuation metrics. I was looking for an entry point 20% below then current valuations. Since then, the shares are down by more than 40% (they are actually 73% below the high they reached in November 2021 although they are up by a bit less than 10% since the start of the year.). Most readers will recognize this to be a fairly typical pattern for high growth IT shares. Confluent hasn't yet reached profitability but - and again like many other high growth IT companies - it is forecasting that it will exit the current year with break-even non-GAAP operating margins with cash flow performance mirroring the improvements expected in non-GAAP operating margins about a year sooner than prior projections.

There are lots of attractive opportunities amongst the casualties in the high growth IT space. Confluent, back when I initially reviewed it, was a leader in a space that its founders had essentially invented. It was a high potential space then, and it really has more potential today than it had in the summer of 2021 as technology trends favor its solutions. The current environment, unlike the summer of 2021, is not easy for enterprise software companies. Deal cycles have elongated, users are trying to find savings, would rather make investments, and many deals are getting "chunked." Confluent is not going to avoid all of this, although it probably will weather the storm better than most.

I am fairly often asked "is something "better" as an investment than something else?" It is of course, a fair question, and an animating factor for me to write articles. Confluent ranks #1 in a specific area of IT, that of providing solutions that facilitate data in motion. Data in motion is a big deal for users wanting to improve the performance of their digital transformation apps. And it can be a big deal for users trying to optimize their usage of the cloud. It means, I believe, that in a less difficult environment, multiple years of hyper growth which I define as growth in the mid 30% and above range. And it also means a business model that supports significant levels of profitability and free cash flow margins.

Will the shares of Confluent achieve greater percentage returns than those of MongoDB ( MDB ) or Snowflake ( SNOW ) or many other alternatives over the coming months/years? I simply lack that kind of specific foresight. What I think I can say is that Confluent will continue its record of consistently beating and raising estimates, both now, and in a future in which the economic environment becomes less negative, and in that environment its future valuation performance will keep pace with a number of peers. As will be discussed, Data in Motion , or the commercial adaptation of the Kafka platform is becoming one of the pillars used by IT to achieve the performance necessary to support the continued migration to the cloud. Deploying such solutions is neither optional, nor replaceable by alternative technologies, to the best of my understanding of the subject. That is really the reason investors ought to consider Confluent shares at this price and at this time.

Shares of Confluent will almost surely be criticized by many because of their valuation. Valuation for high growth companies is a function of… well growth but equally, the ability generate free cash flow while sustaining growth. Looking at historical metrics can be misleading in attempting evaluate companies such as this. Of course investors are aware that Confluent hasn't been profitable. And of course investors are aware that growth in 2023 is going to be constrained because of severe macro headwinds. My bet is that growth will return as the Fed policy of destroying demand through high rates and a compression of its balance sheet gets dialed back. But even beyond that, so-called cloud optimization efforts will gradually become less of a headwind, as cloud spending growth recovers, at least in percentage terms, in 2024 and beyond because of the mission critical nature of the apps supported by usage.

I have projected a 3 year CAGR for Confluent of 39% and much of this article is based on substantiating that kind of expectation. One thing to note is that a substantial component of the company's business comes from consumption. When times are flush, consumption levels can easily surpass estimates. Consumption is hard to forecast - one can ask an enterprise software salesman to project the usage of the solution he sold a year ago and expect an intelligible answer. Just for the record, last quarter usage growth surprised Confluent's management and led in part to the reported upside in revenue growth.

Another issue that often clouds judgement is that of profitability. This company has never been profitable and it hasn't generated cash. It is now projecting that it will reach those milestones at the end of the year. The mechanics for 2023 are pretty simple; product gross margin to remain stable at around 72%, revenue growth of 30% somewhat back-half loaded and opex growth of 15% or less. The question is what comes after that. The company had a free cash flow margin of negative 29% last year, and its projections essentially equate to a free cash flow margin of negative 11% this year. In a recovery scenario, it is easy to project 25% free cash flow margins in 2025 especially because of incremental usage, as more and more of the company's business becomes cloud based. Obviously incremental usage is almost entirely profit, and to the extent usage growth exceeds expectations it will lead to a spike in margins and cash flow generation.

If that happens, and I think it is a probable outcome, then the company will reach a free cash flow run rate of greater than $400 million, a little more than 2 years from now, with prospects for continued growth in that metric. It is that kind of calculus that undergirds the company's valuation but it is often a sticking point for some analysts. Of course such a projection is not written in stone, but I think it is a likely scenario, and one that makes Confluent's valuation far more palatable than some readers might imagine.

That does not mean that I have a particular set of expectations for the quarter the company is scheduled to announce on 5/3/23. So far as it goes, all of Confluent's quarters have been beats in terms of revenue than its prior expectations. What will matter is whether or not the company feels confident in maintaining its current forecast for 30% growth for the full fiscal year, with non-GAAP earnings reaching break-even levels by the end of the period.

Confluent also has an analyst day scheduled for mid-June. These kinds of events typically provide qualitative substantiation that undergirds forecasts. My positive recommendation for the shares is not based on the company's performance in one particular quarter or the evaluation of projections to be made at its analyst day, but rather on its competitive position in a high growth market and its ability to execute successfully to take advantage of that opportunity.

Most recently, a survey undertaken by the analyst team at Morgan Stanley essentially suggested that Confluent is in a position to benefit from the trend toward cloud optimization which is a hotly discussed and debated subject amongst analysts in the enterprise software space. Since the survey confirms my own predilections with regards to user interest in both cloud optimization and the use of Confluent to achieve that goal, it must surely be accurate. I kid. That said, it is really the use cases that are proliferating for Confluent that show why its solution set can be a key part in optimizing cloud usage. Since the concomitant upgrade by the MS team of Confluent shares in the wake of that survey, they have risen by about 5% before essentially retracing that move.

Confluent latest numbers - A pretty strong quarter and outlook all things considered

First, Confluent does use stock based compensation. Stock based comp was 45% of revenues last quarter. That is, no doubt, an elevated ratio, but it is down from 48% of revenues the prior year and from 50% of revenue the prior sequential quarter. For those readers who are concerned about the reported SBC expense ratios, Confluent shares will not be a reasonable investment.

In my opinion, looking at dilution is a better way of actually assessing the cost of SBC to shareholders. Currently, dilution is running at a little bit more than 1%/quarter. The company is well aware of the impact of dilution on shareholder values; during the latest conference call the CFO spoke about actively managing dilution to the 3%-4% range this year. With that in mind, my valuation calculations are based on 299 million weighted average shares. The CFO spoke to a policy of further reducing dilution going forward. If the company can reach a reasonable free cash flow margin, and can grow the business at rates in the mid-30% range, 2% dilution to shareholders works out pretty well, at least in my opinion. That's basically the calculus I use in supporting a purchase recommendation for the shares.

Last quarter was a notable beat in most revenue and cost metrics for Confluent. The company had forecast quarterly revenue of $162 million; actual revenues were $169 million. The company had forecast a non-GAAP loss operating margin of 28% it was 21.5%. The company doesn't forecast RPO balances, but they rose by 48% year on year and by 12% sequentially. That actually may have been less than the company's target, but still provides the company with a high level of visibility. The company's free cash flow margin was negative 18% last quarter compared to 30% in the prior sequential quarter and compared to 22% for the prior year.

As mentioned, this company is seeing the same weakening demand signals that almost all other companies are seeing in the enterprise software space. Probably the sequential rate of growth in RPO balances was impacted by elongated deal cycles and more chunked deals. In any event, the company is forecasting revenue growth of about 30% this calendar year, some of it to be back half loaded. As mentioned, the company is forecasting opex growth of 15%, down sharply from opex growth in 2022. That is the formula that leads to break-even non-GAAP EPS and free cash flow margin by the end of the current year. I think the ability to avoid layoffs on the road to profitability will be a significant benefit for the company, once there is a Fed pivot, and software demand growth starts to recover.

What does Confluent actually do and why is it a priority purchase for many enterprises?

Confluent is obviously not a name on the lips of many investors. I follow technology companies, and subscribers to my service often ask me for reviews of specific companies. This is one of those. But outside that circle, I recognize that a company whose solutions are based on improving the performance of IT infrastructure is not one of the high profile businesses that gets lots of attention. Data in motion probably doesn't sound all that existential.

The technology for Confluent is based on an open source distro called Apache Kafka. It was initially developed at Linked in, and it was developed by the two founders of Confluent, Jay Kreps and Jun Rao. It is called Kafka… because Kreps liked the man's writing. Not my taste, but the picture at the start of the article is a sculpture of Kafka in front of the Prague City Hall. One of the principal components of the technology is stream processing. Here is a link to a description of stream processing.

Stream processing is an alternative to batch processing. Most users want to do something with the data at the time it is created, rather than wait for some predetermined interval. Data comes from many sources; one fairly typical source is the sensors that are deployed for IoT applications. But of course data from payment processing is another common application based on stream processing.

Use cases are proliferating exponentially at this point, as stream processing-or really the speed in terms of application performance that stream processing enables-becomes a must have. The TAM itself is growing quite rapidly, and Gartner believes it will reach $100 billion by 2025. By far, the largest component of the TAM is within the market described as Application Infrastructure.

Obviously Confluent Cloud pricing is consumption based with costs based on an hourly and a data transfer charge. The Confluent Platform has a typical seat based pricing structure. Unlike some other vendors with consumption based pricing models, there has been no indication that usage trends are turning constrained. Specifically, the sequential growth of Confluent Cloud was actually 19%, and while obviously that kind of growth is likely to diminish this year, there is no indication of users attempting to constrain Confluent usage.

One common use case is that of fraud prevention. Many readers will have tried to buy something for more money than typical when using a specific card, or from a foreign vendor, or in some other way anomalous from historical patterns. The transaction may be declined, and it is the use of stream processing that makes it possible to identify anomalous transactions in real-time.

The Internet of Things often is based on sensors deployed across a physical domain. If an anomaly is detected, stream processing identifies the problem immediately and as part of an application, indicates what needs to get fixed. I recently wrote an article about Samsara ( IOT ) whose applications reply on stream processing as a key to their ability to notify users of faults or anomalies.

Real time personalization is a fairly new and emerging use case. The use of stream processing allows companies to deliver personalized, contextual experience - which outside the realm of needless abstruse syntax - means that if a shopper ads something to a cart, and doesn't buy it immediately, the seller can either make a special offer for that product or offer an equivalent at a lower price.

As mentioned, Confluent was founded by two of the creators of Apache Kafka. The TAM calculation presented by the company is about $60 billion Kafka is a very widely used technology with 80% of the Fortune 500 having deployed the distro.

As with many open source projects, Kafka needed to be re-engineered for use in many enterprises, and for multi-cloud or on premise deployment. At this point, Confluent offers a fully managed cloud based version that is available on all the major cloud providers, i.e. Microsoft Azure (MSFT), AWS ( AMZN ) and Google Cloud Platform (GOOG) (GOOGL). The cloud based version of Confluent has been sustaining triple digit growth and it appears as though it will cross 50% of company revenues by the end of the current year. Most new users are choosing the cloud based service; growth of the on-prem version of Confluent showed growth of 17% last quarter.

Although Confluent is a young company it has some impressive reference accounts. Some household names using Confluent include Citi ( C ), Goldman Sachs (GS), Square (SQ), PayPal ( PYPL ), Ticketmaster, Bosch, Walmart ( WMT ), Dick's Sporting Goods ( DKS ), BMW, Humana and Expedia. Like many other infrastructure software companies, the success of Confluent has been based on a land and expand sales motion.

Confluent announced its acquisition of a company called Immerok early this year. Immerlok has been a leading contributor to Apache Flink which lately has been used to develop stream processing applications for very large analytics workloads. Confluent will be offering Flink as part of its cloud. Flink has become a choice of many developers. This is more of a technology/people acquisition rather than a financial transaction. The company is not forecasting any material revenues from the acquisition this year, and has incorporated the company's cost run rates in its financial projections. The cost of the acquisition was not disclosed, suggesting it was not particularly material as these things go. Immerlok/Flink should provide some competitive advantages for Confluent, and help to keep the net expansion rate at healthy levels.

Competition

There are many competitors of various kinds with offerings in the stream processing space. Most users are going to compare Confluent to what is available from vendors who offer other managed Apache Kafka services. Usually I like to look at 3rd party research to avoid the obvious biases of vendor hype. But there really isn't any list of apples to apples alternatives to Confluent's flavor of Kafka. The link here is to a Confluent commercial as to why its Kafka service is better than those of competitors. Amazon ( AMZN ) is in this market. Amazon often has competitive offerings that it has developed from open source projects in several other areas such as its version of Elasticsearch (ESTC). I am not any form of IT consultant. And what I have linked here is self-evidently a commercial prepared by marketers from Confluent. But for readers wanting a deeper dive than I can provide in an article such as this, I recommend looking at the comparisons and contrasts presented in the analysis, even if it is written from a Confluent point of view.

Confluent maintains that its version of Kafka has been specifically reimagined for the cloud and that it enables developers to reliably and securely build next-gen apps faster. There are plenty of testimonials and references suggesting that the assertions might be accurate. After having looked at the third party analysis of competitive stream processing offerings, I simply found nothing of significance that might help me with an investment conclusion.

Certainly the largest competitor for Confluent is the Apache open source project of Kafka. Again, the 3rd party research reviews weren't particularly helpful in evaluating why users would be choosing one or the other offering. I have linked to one such review here, nonetheless. Probably more helpful, again, if readers are trying to take their own deep dive, is the linked commercial prepared by Confluent comparing the open source version of Kafka with what is specifically offered by Confluent. While of course open source Kafka is free, and Confluent's offering is not, the company has made the point that the other cost factors in running an application led to a 60% reduction in TCO when using Confluent. Of course, like any paid offering, Confluent offers a suite of tools to help developers create applications that essentially can run anywhere.

At some level, it would be surprising if Confluent wasn't offering the most effective, easy to use flavor of Kafka. Its founders created Kafka and also something called Apache Samza which is a stream processing system while they were at LinkedIn. Jay Kreps, Confluent's CEO, was the lead architect for data infrastructure at LinkedIn. Sometimes it is said that someone wrote the book about something. In the case of Confluent's other co-founder, Jun Rao that is literally true. While I don't suggest that investors will actually get much from this link , and I confess to not having read most of it, Jun has written a course, Apache Kafka Architecture and Internals. From what I can glean, Rao is the expert on how to exploit Kafka to best create modern applications.

Certainly not all potential customers will choose their stream processing vendor simply based on product feature/function-or even total cost of ownership factors. Many users will choose to work with Amazon and IBM (IBM) and others simply because they already use those vendors and have a vendor consolidation strategy. Others will go for the open source version of Kafka… because it is supposedly free. But I think the hyper-growth story at Confluent is well supported by the functionality and differentiation of its product and this is an advantage I expect will continue over the course of many years. There seems little real controversy that Confluent offers users a superior stream processing solution and that its version of Kafka seems to resonate with users.

Confluent's business model and its path to profitability

As mentioned, Confluent is not yet profitable, and has yet to generate positive cash flow. Investors these days are less inclined to be patient and tolerant of businesses that are not producing returns, and Confluent is making a pivot with the company forecasting profitability and free cash flow by the end of the end of this year and beyond.

There is plenty of opportunity for the company to improve its opex ratios which are elevated. Fortunately for Confluent, the mechanics of doing so are far less draconian than is the case for other companies with more revenue growth issues. To repeat. The company has forecast revenues growing by 30%, non-GAAP gross margins remaining flat at around 72%, and non-GAAP operating expenses rising 15%. The growth the company is forecasting will be a little back half loaded, and that means that Q4 should be the initial quarter for non-GAAP break-even results. To reach profitability the company will not need to embark on layoffs, but will instead moderate the cadence of new hiring.

Confluent is forecasting that its Cloud revenues will reach 50% of the total for 2023. Because cloud revenues are ratable, this has created a modest headwind to gross margins. Over time, and of course depending on the mix in a particular period, that headwind will gradually reverse and become a tailwind. The CFO, in the latest conference call script, indicated that pricing had not deteriorated, and was not a factor in terms of the trajectory for margins.

Even before the announced pivot toward profitability, Confluent's opex ratios were trending down, albeit from very high levels. Last quarter, non-GAAP research and development expense was 25% of revenue compared to 29% the prior year. Sales and marketing expense was 57% of revenue down from 65% of revenue the prior year and general and administrative expense fell to 13% of revenue down from 16% the prior year. Overall, the non-GAAP opex expense ratio in Q4 fell from 110% to 93%. The sequential improvement was equally significant. Overall, non-GAAP operating margins improved by 2000 bps (20%) for the full year, so the company forecast of a further improvement of 600 bps is probably conservative.

The real question, and the one that most investors are focused on, is not that of the improvements in the company's business model, but the security of the revenue forecast. The CFO answered the question this way:

Sanjit Singh

And then, Steffan, I could just sort of connect some of the dots on the financials. The Confluent Cloud revenue in Q4 was excellent, record quarter for Confluent Cloud revenue. The RPO was certainly weaker. And then when I look into the 2023 guidance, revenue guidance, it only came down, I think, $5 million. You sort of narrowed the range. What gives you confidence that like the revenue is sort of set at sort of the right level, just given some of the dynamics you're seeing out on the macro?

Steffan Tomlinson

Well, we took into consideration our current outlook on the macro, and we really focused on a few things. One is our current RPO exiting Q4 gives us about 60% visibility to our total revenue number in FY '23, which is actually five points higher visibility than we had this time last year. We also have more proportionally sales reps that are fully ramped than are ramping, and we see that growing out throughout the year.

And then lastly, we just came off of a quarter where we saw a very robust growth in $100,000-plus customers and $1 million plus customers. And those cohorts contribute north of 85% of revenues. And so we have the right product for the right market, and we feel like '23 will be a decent setup for us.

Many readers are focused on how the company is going to perform during this time of macro headwinds. I prefer to try to look at long term opportunities as opposed to a focus on specific quarters and the tracking of deal scrutiny trends. But here is a further discussion from the conference call on the subject.

Raimo Lenschow

And then one quick follow-up on more numbers. So if I think about the -- you kind of moved to the profitability goal one year forward, which is kind of a big change and it takes a lot of effort from the organization. Can you talk a little bit about the compromises you have to think about there? Was that certain growth projects you kind of maybe kind of deemphasize? It doesn't sound like it's the sales force getting impacted. Like, just talk a little bit about like the puts and takes you have to kind of go through to get to that because that's quite a big effort. Thank you.

Jay Kreps

Yes. Yes. I mean any change like this is a little bit disruptive. And so I think that's the -- probably the biggest impact for us is just making sure that we get off to a fast start at the beginning of the year. We're not so disrupted that, that impacts execution. It's obviously also just a harder thing to go through. We felt like, look, after a couple of years of very fast growth, where we kind of roughly doubled headcount in that time period, there was opportunities for efficiency, right? And despite being very thoughtful in planning and where we were deploying resources we thought there was opportunities to get more efficient.

So for us, it was kind of a question of how do you do that. Are you going to do it more slowly, kind of in place? Are you going to do it more quickly? As we got, I think, a better read on just, what's the environment for '23, what's the environment overall in tech, what makes sense for us, we feel like it made sense to do it more quickly. And that kind of, I think, shows a little bit of what's possible for the business in terms of efficiency or is at least one good step in that direction. And it seemed like in the environment, it just made sense to do that now.

Steffan Tomlinson

And we're also doing it, preserving our ability to drive top line growth, and continue to invest in our innovation engine. And we are able to balance the moves that we made to preserve our long-term sustainable competitive advantage.

No one should imagine that industry trends aren't going to impact the company's performance. But this environment is not likely to persist indefinitely. I get many questions about when the time of Cloud Optimization will be over, or put another way, when IT budgets will start to show improved growth. The simple answer is I don't know and really anything I or anyone else might write on the subject is conjecture. Much of Confluent's conference call was about very detailed questions regarding sales execution, and sequential percentage growth.

Wrapping up - The case to buy Confluent shares

When I last reviewed Confluent as a business and as an investment shortly after the company's IPO my recommendation was a bit cautious because of valuation. Valuation has compressed, and the company has achieved all of the operational performance anyone might have anticipated since that initial article. Its trading pattern is very similar to many other high growth IT companies.

The company is a leader in what is known as the stream processing space. For most readers, the key point about stream processing is that it speeds up performance of cloud based applications very substantially, and the technology has to be incorporated in many applications for users to have an acceptable experience. Over time, use cases have proliferated and continue to do so, and the $60 billion TAM estimate is on its way to $100 billion. The company's founders were two of the project leaders in creating Kafka, an open source project that facilitates stream processing and which is widely deployed. Confluent's version of Kafka comes with the features that are important in most commercial applications and which actually provide users with a lower TCO than the open source distro despite the cost of the software.

Confluent offers users two models for consumption. One is a self-managed subscription model and the other is the Confluent Cloud. The cloud offering has proven to be very popular and this year about half of the company's revenues are coming from the cloud. This makes the company's operational performance even more impressive as the cloud, with a consumption based revenue model, basically defers some revenue recognition compared to the company's Platform offering.

Many investors are concerned about how cloud optimization trends are going to affect software companies with a usage model such as that of Confluent. Last quarter, Confluent usage was stronger than anticipated. The company, in its very modest forecast for cloud revenue growth this quarter, has provided a forecast that accounts for the potential for usage issues in Q1 and beyond. The issue of usage is one that should really not be a focus of long-term investors; usage growth. Usage growth will return simply because using Confluent - and for that matter, many other applications priced on a usage model - is necessary for critical applications to achieve user experiences expected by most.

The company has announced a pivot toward profitability which is designed to yield break even non-GAAP profits, and free cash flow at the end of the current year. The forecast is predicated on the ability of the company to maintain 30% revenue growth; I think the company has made a strong case as to why its forecast will be achieved despite macro headwinds.

Confluent shares, despite a notable valuation compression will not appear as cheap to some. Based on the company's guidance my estimate of forward EV/S is about 8.2X based on the share price of $24.45 as of April 19th, 2023. Besides hyper growth, the pivot to profitability and free cash flow generation makes the shares attractive. For example, I think free cash flow generation can reach a $400 million run rate by the end of next year. That kind of free cash flow generation would support significantly higher valuations, of course depending on the market phase.

As I most often do, I will reiterate that Confluent shares will need a risk-on environment to achieve consistent positive share price trends. I make no claim that I know when that will happen, as that kind of sustained market pivot will be a function of trends regarding inflation, and actions by the Fed board of governors.

That said, Confluent dominates a key space in the cloud, and it is likely to continue to do so indefinitely given its current leadership. It has an improving business model, and its projections are quite conservative. Use cases for stream processing continue to proliferate, and are likely to continue to do so despite macro headwinds.

With that in mind, I think the shares are likely to achieve positive alpha over the next year despite the current environment.

For further details see: