State of the Cloud 2023 with Bessemer Venture Partners

Well, it's wonderful to be here with you today. We love joining the Saster community every year for Bessemer's annual State of the Cloud report. And we love going through the trends, benchmarks and metrics that underpin the cloud economy and that are most relevant for the cloud founders and operators in the room. I'm Mary d' Onofrio and today with my colleagues Janelle Tang and Katie Ray, we're going to be going through just that. Janelle is going to be talking about the macro's impact on public cloud software and where we go from here, how public cloud companies have refocused on profitability and efficiency. Katie is then going to talk to us about how you can operationalize some of those lessons for your own business, some of the tactics that you can employ. And then I'm going to go into the topic du jour, AI. But before we do that, we'd love to talk a little bit about where we've been over the past 12 months. As we all know, the past 12 months have been relatively turbulent for cloud founders and many of the cloud companies that likely many of you represent. SBB collapsed market multiples are down and yet we have a platform shift to AI that is exciting everybody and the IPO window seems to be reopening. So what does this all mean for cloud companies? We're going to start by handing it to Janelle who's going to talk about the public markets.

Thank you Mary. As Mary mentioned, it is no secret that cloud stocks took a drubbing in 2022 in a severe market pullback event that we coined the sassacre. And as you can see from the slide, within the year the BVP cloud index had slid over 50%, which was a more marked decline compared to broader market indices. A main reason for the pullback was due to the external shock of hiking interest rates. So within the year we moved out of a hyper low interest rate environment at record breaking speed. And this led to multiple compression and a significant amount of macro uncertainty. By the end of the year, valuation multiples for cloud companies had not just fallen below the trailing ten year average, but had also fallen below the pre pandemic long term averages that a lot of us in the markets we're used to seeing. And beyond multiples, many cloud companies also faced a deluge of headwinds that impacted business fundamentals such as lengthening sales cycles and tightening customer budgets. With such a bleak macro backdrop in place, it came as no surprise that SaaS IPOs came to truly a screeching halt, which was quite a stark contrast to the frenzy of what we saw in 2021. On the M and A front. Similarly, things slowed but remained more robust for several reasons. One, it was a record year for private equity sponsored deals, specifically take private transactions and two, on the strategic side, there were announcements of several blockbuster acquisitions. Most notably, the biggest announcement was from Adobe that announced its acquisition of Figma, which marked the highest valuation multiple offered in an acquisition for any software company of scale. And while that deal has yet to close, what it demonstrates is that best in class cloud companies can derive a valuation premium even in extremely tough times. So suffice to say, as we've gone through, 2022 was certainly a gloomy year for the cloud, but the skies do seem to be partying in 2023. Year to date, the BVP cloud index is already up over 25%, indicating that there may be light at the end of this tunnel. As we talked about, a lot of the uncertainty in 2022 stemmed from changing macro conditions, most notably interest rate hikes from the Fed in an attempt to curb inflation. In 2023 we've seen a lot more stabilization on this front. So for instance, we have slowing inflation and the CBOE volatility index has remained healthier this year. And while we're still quite a ways away from the highs of the valuation multiples that we saw in 2021, we are seeing valuation multiples normalize closer to the range of what we saw for long term pre pandemic averages. Consequently, with more stabilization, the IPO window is also showing more signs of life. Just a few weeks ago, marketing automation cloud leader Klaviyo filed its S1 to a lot of fanfare in the cloud community because this marks the first pure play SaaS company to file to go public after a nearly two year drought in SaaS IPOs. And at Bessemer we very much believe that the the top private cloud companies will shortly follow suit as well. And on the M and A front, a lot of the trends that we saw in 2022 carried on into 2023. Specifically we saw again a significant amount of take private transactions. And on the strategic side we saw a recent acquisition from DataBricks for Mosaic ML for a $1.3 billion transaction, which indicates how AI is becoming a top M and a priority for a lot of companies. You know, at Bessemer we very much fundamentally believe that the cloud model, with its recurring revenue nature, low marginal distribution costs and strong net dollar retention dynamics, is quite possibly the best business model to be invented, and definitely one of the most resilient business models ever to be invented. And in the face of one of the toughest years for the cloud economy, we witnessed many companies remain resilient by adapting quickly from the age of excess. As we moved into an age of efficiency, we witnessed cloud companies start to pivot their focus from just growth at all costs toward driving profitability. And that's visualized very powerfully just within the last couple months in this graph. Today, the top 10 highest valued BVP Cloud Index companies are driving an average of 14% free cash flow. And this is double what the same top 10 cohort had achieved just a year ago. Just as companies adapted, the market has also responded to the paradigm shift. And efforts to drive profitability are not going unnoticed and in fact are being rewarded. So let's put this into a heuristic to contextualize how the public market valuation sentiment has changed at peak bull market exuberance. So this was the end of 2021. A 1% improvement in revenue growth had the same impact on valuations as a 6% improvement in free cash flow margin. So said another way, growth was more important than profitability by a factor of six at the height of bull markets today, this ratio is closer to 2 to 1. Still in favor of growth, but certainly a bit more balanced from what we saw in the bull market years. And so today, a 1% improvement in revenue growth has the same impact on cloud valuations as a 2% improvement in free cash flow margin. And so we've talked a bit about how the public market responded to from the change of the age of access to the age of efficiency. And I'll hand it over to Mary to walk us through what happened on the private side.

As Janelle noted, the public cloud markets are now rewarding efficiency and profitability over the growth at all cost model. So what does this mean for private cloud companies? We can use the Cloud100 as a proxy. The Cloud100 is a list that Bessemer puts together every year in partnership with Forbes and Salesforce Ventures, and it represents the 100 best private cloud companies in the world every year. This year for the first year, the aggregate value of the Cloud 100 actually decreased year over year by 11%, still to a phenomenal $659 billion. But nonetheless, it points to the fact that the macro has impacted the best of private cloud. So what are they doing in that context? Like their public peers, the Cloud 100 companies are focusing on efficiency. 23% of Cloud 100 companies will be profitable by the end of this year 2023 and 2/3 are planning to be profitable by the end of 2024. Even for the companies that are still burning, 60% of them are meaningfully reducing their burn, only 7% are meaningfully increasing their burn. And just to give you some context, that's a market departure from the heights of 2020 2021, where growth at all costs was more valued. And why are they doing this? Like with the public cloud, Cloud 100 growth rates have come down. Macro is impacting the private cloud markets, softness in purchasing power, longer sales cycles, and as a consequence, the average growth rate in a Cloud 100 business in 2022 was 100% and and this year went down to 55%. If these companies did not focus on operational efficiency and right sizing their PNLs, that discrepancy in revenue that they might have been projecting would flow through the bottom line as increased burn and decrease Runway and speaking of Runway, and speaking of the market dynamics, market multiples have gone down even for the best of cloud. In 2021 the average cloud 100 multiple was 34 times, decreasing to 30 times in 20 and 26 times in 2023. This is important because even though it hasn't gone down as much as the growth adjusted public cloud multiples have, it nonetheless reflects the fact that cloud companies can't command the same market multiples that they had previously. And by right sizing your P and L, focusing on efficiency and focusing on operational leverage, you can raise when you feel like your fundamentals are strongest and when you have the most opportunity to get a strong valuation because not from a position of weakness because you have to raise in order to finance your business. But how exactly do you do that? How do you drive efficiency? And what are the tactics that you all in the audience can employ in order to drive more operating leverage in your businesses? To talk about that, I'm going to pass it to my colleague Katie.

Great. Thank you Mary and Janelle for that comprehensive overview of the public and private markets. So as Mary mentioned, the question on many cloud founders minds right now is I understand where I need to get to, but how do I actually get there? It's easy to say you need to transition to profitable and efficient growth, but the devil's in the details. And the question is how do I actually look at my own P and L and understand where there are opportunities to optimize it. So one thing that we did at Bessemer this past year has been speaking at and Speaking to and looking at both public and private cloud companies, as well as companies in the Bessemer portfolio and beyond, and talking to them around what tactics they have used to transition to more of an efficient growth model rather than this growth at all cost model. We broke it down by major area, the P and L. So starting with cogs and gross margin and then going into the major buckets of opex, sales and marketing, G and A and R and D and then to people and pricing and packaging, and we accumulated a set of tactics that can be used by cloud founders to discover potential efficiencies and to actually drive those efficiencies within your P and L. So over the next section of this presentation, I'll go through a case study on each of these major areas and a way that a Bessemer portfolio company has actually applied a tactic in each of those areas. So starting with gross margin. Gross margin is effectively the ceiling for profitability of your business a company. Public companies that have less than 60% gross margins find it structurally basically impossible to drive more than 20% free cash flow margins. So given this and given the correlation that Janelle discussed on free cash flow to valuation, it's an incredibly important area to look at and can have an outsized impact on your valuation as you think about areas of the P and L to optimize. Take SendGrid in our portfolio. One tactic that they use that has been very kind of top of mind for founders recently is optimizing their cloud hosting costs. They realized that their cogs could actually be minimized by doing so. And so what they did was they looked at their entire architecture and they tried to understand and audit what were the most expensive areas of their architecture from a compute, network and storage perspective. And then they worked with their engineers to understand is there a way to re architect those components of our architecture in order to in their case not only just drive better performance, but it actually drove seven figure annual cost savings for their cogs. So lucky for cloud founders in this room, SendGrid did this on their own. But there's actually an emerging ecosystem of cloud coast management management providers like Cloud Zero or Kubecost that can actually be leveraged in order to audit your cloud footprint and understand where there may be efficiencies. So moving on to the first major bucket of opex and the largest sales and marketing for most cloud public companies comes in somewhere between 30 to 40% of revenue. As a result, it's really important to look at and understand whether there may be efficiencies to be driven. Take Gloss Genius in our portfolio. They are a vertical SaaS and payments company for the beauty and wellness industry and they were very lucky. From the perspective or not lucky, they were very good and they built a fantastic product that allowed them to grow completely organically at the start. People who own spas and salons, they love the product and they would speak to others about the product and the company just grew incredibly quickly organically at the start. As the company scaled though, they realized that they needed to layer in inorganic channels for growth. But keeping this eye towards efficiency, given their kind of history in their early days, they came up with a heuristic that we would recommend cloud founders in the room use. They came up with this idea of having a North Star benchmark by channel and using it basically as like a hurdle rate. So the idea was for, for any channels that hit this hurdle rate that would justify continued investment, but for channels that didn't clear this hurdle rate, they would stop investing in those channels and then reallocate those sales and marketing dollars into the over performing channels. And so for Gloss Genius, they decided that their North Star benchmark should be a sub 8 month CAC payback and a 3x LTV to CAC. That was kind of what worked for them and that was the hurdle rate that they held their channels to when they did a channel by channel review. We would encourage cloud founders in the room to think about what might be the right hurdle rate to apply to your own channels and then audit your sales and marketing channels accordingly. Moving on to GNA. So G& A tends to be a little bit overlooked when you're thinking about efficiencies in your P and L because it's actually the smallest area of opex. For most public cloud companies, it comes in somewhere between 10 to 15% of revenue and is thus one of the smallest area of opex. That being said, though, considering that it's purely a cost driver or a cost center rather than a revenue driver in the way that sales and marketing or R and D is, it's generally an area that you can find some quick wins and some potential efficiencies without actually potentially impacting your overall growth rate. You know, recently, given the rise of work from home, looking at whether there's efficiencies within your real estate footprint that you could rationalize is one way to think about it or, or automating more back office functions. One thing that a company in our portfolio, Catmo, did though was they looked at their SaaS sprawl footprint and understand whether there were areas to optimize it there. So catmo is a construction management software company based out of Germany. And they realized that they didn't really have a good sense of what software had been procured over recent years, what was the price per seat, how many seats did they have, et cetera. And so they conducted an audit of that. They found potential efficiencies, they found duplicative instances, et cetera. And they also were able to work with a company, Vertice, in this case, where they could actually get some of the benefits of group purchasing and price transparency to negotiate about a third off of their contract costs of a few key large contracts. So Vertice is just one example. And folks in this room are probably familiar with these companies. They actually have booths here. So you might even, might even be a member of some of these companies. But companies like Vendor Vertice, Tropic, et cetera, they can all help you understand your software footprint and potentially fight some of that SaaS sprawl within your organization. So the last bucket of opex, and probably the most fraught one to actually try to optimize, is research and development. So R and D is tricky because you can maybe find short term gains, but in the long term that could actually degrade your competitive advantage. So the question for cloud founders is my R and D spend is a bit of a black box. How do I actually think about optimizing it? And we think that that's a good place to start because sometimes cloud founders might say, I don't even want to touch R and D. You know, I don't want to have any potential risk of degrading competitive advantage. The good news is, you know, there's companies like Jellyfish and Allstacks that can be used to understand your engineering productivity and actually do benchmarking to get a better sense of that department. And there's. You can also. We encourage heads of R and D to force rank your current R and D projects based on highest perceived ROI to lowest perceived ROI and then cut some percent of the bottom performing and again reallocate that capital to the highest roi. That being said, though, a particularly interesting way that we're seeing people drive greater efficiency in R and D as well as all internal functions, is through the rise of LLMs. As I mentioned, they can be used across all internal functions. But take engineering for example. They've been proving to be really efficient and helpful in actually helping generate code and increasing the productivity of your engineering. At Bessemer, we have a four step framework that we encourage our founders to use as they think about where are potential opportunities to leverage and adopt LLMs both externally for product innovation but also internally for potential efficiencies to be had as we think about this transition towards more efficient growth. So on that note, I'll turn it over to Mary to walk through some examples of how Cloud 100 companies will are leveraging AI and LLMs for their external product innovation.

So as you think about leveraging AI and optimizing your R and D efficiency, it's worthwhile to know how the best companies are doing it and what the speed is with which they're doing it. Cloud 100 companies are doing this exceptionally rapidly. In the past eight months alone, 55% of Cloud 100 companies have launched a generative AI product or features. 70% of them use AI or ML somehow in their products today. So it's worthwhile to take a lesson from these best in class companies that leveraging the new technologies and the ones that can deliver the most differentiated product experiences the quickest is the path to success. But how do you do it? There are three strategies that Cloud 100 companies have exemplified over the past year or so on how they've leveraged AI. One is embedded AI, one's acquired AI and one's native AI. Let's start with Embedded Embedded AI is simply a strategy of using external AI tooling to augment your product offering, supercharging your existing cloud platform and delivering better customer experiences. Cloudinary is a company that did just that. Cloudinary is a cloud solution for managing digital media assets, and it used generative AI functionality to augment its existing suite of programmable media APIs. For example, generative replace used natural language prompts to detect and replace objects such as this black sweatshirt to this white coat. By doing so, Cloudinary opened up its TAM by allowing its platform to be used by non technical users as well as its core technical audience. The second is a story that we've already alluded to earlier, which is Databricks acquisition of Mosaic ML Mosaic was a platform enabling enterprises to create, train and deploy their own generative AI models, and through this acquisition, Databricks is now able to offer a unified platform to its customers to deploy their own AI models and it's number two on this year's Cloud 100 list. While that might not be for everyone, acquiring companies is quite an art. Another approach is just being AI native from the get go. AI native companies think about AI as being core to their platform architectures from their very inception A great example of a company that has done this is DeepL Investima's portfolio which offers real time and AI powered language translation. It uses neural machine translation and deep learning to understand very complex linguistic patterns and nuances such that it has the highest level of accuracy, continues to iterate and to learn and can be used in business context given its accuracy rates. Now we talked about cloudinary Databricks and DeepL, but there are many others we could have talked to. Zapier and plugin ecosystem Intercom with Fin it's GPT4 powered chatbot Canva with magic design and presentations. Cloud 100 companies are leveraging AI across the board and we suggest that that you try to follow suit and think about how to embed AI from a first principles approach. Think about your application, your product workflows, your cross product workflows to think about where AI can drive the most operational leverage through R and D efficiency. And here is the market map of tools that you could use should you choose from an app layer company like a bridge and AI for clinical documentation to on the developer tooling side hugging face or the foundational model layer open AI. As many people have already experienced, the AI landscape is really broad, continuing to innovate and the innovation cycles are getting quicker and quicker. Now you've learned a little bit about how we think about driving operational efficiency across the board and how AI can be used in R and D in particular. I'm going to pass it back to Katie to wrap us up.

Great, thank you so much Mary.

Well.

So first of all, we really appreciate you all taking the time to be here today. For any further information on this research or any of the other research that Bessemer does, you can find us online@bvp.com or scanning this QR code. And so just to summarize what we've discussed today, despite a turbulent last few years in the cloud ecosystem, the cloud ecosystem is now showing some early signs of recovery. As Janelle mentioned, we have Klaviyo and a few others starting to test the public market waters and see this sort of return to normalization. That being said, given that we've actually had this paradigm shift in the way that investors are valuing companies for their efficient growth rather than growth at all costs. We believe that this actually presents a great opportunity for cloud founders to build really durable businesses and leverage some of the tactics that we discussed to go through your P and L and understand where there may be optimizations. The kind of dream is to be a rule of 40 company and we'd encourage you to look at your P and L and understand what is a potential path forward to potentially get there in the long run. We believe that some of these tactics for efficient growth, as well as AI, can be leveraged in order to drive some of this efficiency as well as innovation. So we really appreciate you all being here today. We know that you have companies to build, we know that you need to go work on those businesses, and so we appreciate you being here. And at Bessemer, we're excited to continue to back the next generation of great cloud founders. We are happy to open the floor to any questions that you may have on the case studies we went through on efficient growth, on AI, or anything else that we covered today. So thank you so much.