Kaltura, “the video experience cloud,” filed an S1 on March 1, 2021 to support an intended initial public offering (“IPO”). The public offering was delayed earlier this month. Multiple media reports attribute the delay to low demand. There is no public announcement about revised timing.
To keep this post somewhat manageable in size, we limit the analysis of the IPO to Kaltura business model (SaaS deployments running in public cloud), statements of market size, and valuation expectations. These have a fascinating interplay in this situation.
In particular, the SaaS-oriented data points disclosed by Kaltura are a window into a hopeful future where industry participants openly discuss sales-oriented business metrics.
Kaltura stands at the precipice of an IPO after 15 years of operational history and six equity financing rounds raising more than $160 million.
The Kaltura press release announcing its $50 million Series F round from Goldman Sachs (also an underwriter of the IPO) characterized the raise as “pre-IPO.” The date was August 2016; these things take time.
Shareholders now need an exit (first money in was 2007); Kaltura needs access to more capital; and equity markets are positively euphoric. While Kaltura does not consume substantial cash at its current maturity, its obligation over the next year are more than its capacity to pay. That is only a reason to raise some funds. The reason to raise lots of funds now – beyond a shareholder liquidity event – is to substantially increase Kaltura’s sales and marketing spend (more on this below).
The offering is for 23.5M shares: 17.4M from Kaltura and 6.1M from selling shareholders. The division is noted since Kaltura only keeps the funds from the shares it sells.
After the contemplated IPO, approximately 123M shares will remain outstanding. The March 23rd S1 amendment anticipated a pricing of $14.00 to $16.00. At the midpoint ($15.0 per share) and considering the expected pro forma balance sheet, the resulting enterprise value of Kaltura would be $1.6B. To put this amount in context, $1.6B is larger than the enterprise value of Avid or Evertz, and twice that of Brightcove. Brightcove is 60% larger in terms of revenue, Avid and Evertz are both about 3x bigger.
Since Kaltura had annual revenue of $120M in 2020, the implied revenue valuation is then 13.6x. At such valuation levels, there is little purpose in discussing profitability metrics. It is, however, worth noting Kaltura (with a cleaned-up balance sheet post-IPO) would operate at approximately break-even.
Rather than passing judgement about the appropriateness of the valuation level, let us instead consider the company Kaltura would keep at 13x+ revenue multiple. Below is the enterprise value to revenue valuation level of the publicly traded, named competitors in Kaltura’s S1. The chart also notes the prevailing price-to-sales ratio of the NASDAQ composite index.
Below is a graph of the revenue levels of the same peer group.
The point is made without labeling the y-axis.
The second graph is a good reference on how inappropriate it is to use the first graph as a valuation comparison basis since the peer group (again named in Kaltura S1) is not trading based on activities in Kaltura’s addressable markets.
Let us instead use the below comparable group to highlight valuation positioning of Kaltura.
Before you view Kaltura’s management team’s eyes exceeding their income statement, we live in unprecedented times in equity and debt markets. These are the highest equity valuations versus economic activity observed in the historical record.
If you never ask, the answer is always no.
Even more interesting than valuation is the basis for arguing for a valuation.
A good place to start is an excerpt from the letter to investors by Co-founder, Chairman & CEO Ron Yekutiel. It starts on page 111 of the S1. The second-to-final paragraph, under the heading “What’s Next” reads as follows:
As a horizontal platform with a multitude of possible applications and a vast market to which we can cater, this is just the beginning! Many of our customers are growing organically through increased adoption and usage. We are also finally accelerating our investment in sales and marketing on the heels of already establishing market leadership across several large markets, and we are doing so with very attractive unit economics and proven operational efficiency. We recently started commercializing our new and exciting products and solutions from 2020 (Webinars, Virtual Events, Meetings, Virtual Classroom and TV Solution), and plan to follow with much more innovation (e.g., develop advanced video AI tools, as well as new industry solutions for telehealth, retail, government, and smart cities, among others). We are also planning to go down-market and cater to smaller customers including SMEs with new self-serve and low-touch products for companies and developers, and to do so with the support of new channel distribution partners. The opportunities for growth are bountiful!
For those interested in quantifying ‘vast market’ and ‘very attractive unit economics,’ please direct your attention to the right side of the below slide from Kaltura’s S1.
Source: Kaltura S1 Filing
We will get to the ‘LTV: CAC’ ratio (Customer Life-time value / Customer acquisition cost) later. At the bottom-right you will note Kaltura estimates its addressable market at $55 billion. This is the ‘vast market’ noted earlier. It is larger than the best industry estimates for total sell-out revenue of all products and services into the professional media production and distribution use case (see IABM DC Global Market Valuation Report).
One of the primary points of discrepancy is the breadth of market verticals included in the Kaltura figure.
The $55 billion addressable market is a Kaltura management estimate for 2020. $55 billion breaks down as $39 billion for the real-time conference market and $16 billion for all other verticals addressed by Kaltura’s solutions. The $39 billion, which is 70% of the stated addressable market, is included in the estimate because of solutions released by Kaltura, following its acquisition of Newrow in March 2020 for $3.8 million of consideration.
There is no source offered for the $39 billion. The only guidance provided on Kaltura sales profile in this market is “strong traction to date.” For those interested in investing at 13.6x sales, it is important to note “to date” encompasses one calendar year.
One paragraph of explanation is offered for the remaining $16 billion market sizing. The internal process at Kaltura had the following steps:
- Estimate the number of organizations in each vertical, segment by geography.
- Determine typical average selling price for each organization type, segment by geography.
- Sum the products of (1) and (2).
There are merits to the above approach. It is quick, inexpensive, easily explained, and arrives at a flatteringly high figure. The downside is no third party should believe it and no management team should base decisions on it.
Kaltura is not the first and will not be the last to take this approach. Only slightly fewer than every public company in the media technology sector manages to find a means to ‘credibly’ suggest an exceptionally large and growing market.
Vimeo took this approach to an art in its recent S-4 filing (detailing its planned spin-off from IAC) where it arrived at a $40 billion addressable market, growing to $70 billion in 2024. In its risk factors, under the title “Vimeo’s total addressable market may prove to be smaller than it expects,” Vimeo’s anticipates the objections of onlookers. It is included below (the emphasis on the last line is ours).
While Vimeo believes, based upon internal data, that every small and midsized business and every larger enterprise will need an online video presence to succeed, the number of entities that are willing and able to pay fees for software-based video services may not be as large as it expects. Vimeo has not conducted research by a third party to validate its data and thesis.
The above is slightly more authoritative than “because we said so.”
We will spare the reader further criticisms of the market sizing methodology or citing additional examples (they are plentiful).
Alas $55B is as much a testament to the imagination of the authors of the document, as it is the lack of energies expended by investors and analysts in understanding the markets of the companies described by these documents. It is not the culprit than interests us, rather the phenomenon.
Business Model Data Points
Referring to the previous slide, the reader will note that directly above the $55B addressable market is the series “5x > 7x > 11x” for “LTV: CAC in 2018 – 2020.” Customer Life-time value (“LTV”) attempts to anticipate the lifetime ‘value’ of a customer relationship to a supplier. Customer acquisition cost (“CAC”) tries to describe the cost associated with acquiring an incremental customer (or at times customer dollar from an existing customer). The ratio of LTV to CAC would then describe the return of investment on customer acquisition (or rather sales and marketing expense).
The seminal “Understanding SaaS valuation Primer” published by Andreessen Horowitz suggests anything above 3x for the LTV: CAC ratio is “a good sign that the business model is working.” (Important point: Andreessen Horowitz use a slightly different calculation for LTV).
Kaltura cites figures for LTV: CAC of 5x in 2018, 7x in 2019, and 11x in 2020. “We estimate that for the years ended December 31, 2018, 2019 and 2020, the lifetime value of our customers exceeded five, seven and eleven times the cost of acquiring them” reads the S1. Taken in its narrow meaning, every $1 sales and marketing expense incurred by Kaltura generates $11 of value from customers.
This is the attractive ‘unit economics’ mentioned in the letter to investors – and goodness it is.
Even a heavy discount to the $55 billion figure allows for plenty of room for value creation at 11x LTV: CAC. Should you take 11x as representative of the unit economics of Kaltura’s business model over the next five years, then you might begin to view 13.6x price to sales with more amenable eyes.
A brief thought experiment helps illustrate the point. Take 11x as an accurate estimate of the return of the marginal dollar Kaltura invests in sales and marketing. If Kaltura then invested its entire post-IPO cash balance ($265 million) in incremental sales and marketing, it would generate lifetime customer value of nearly $3 billion dollars.
In many aspects of the S1, Kaltura points to this opportunity. ‘Acquire new customers,’ is the first bullet under growth strategy. The plan to expand sales efforts is described best in the following passage: “We intend to grow our base of field sales representatives and customer success managers, which we believe will drive both geographic and vertical expansion. Additionally, we are investing for the first time in inside sales, self-serve offerings, and distribution channels.”
To balance the discussion, let us introduce some skepticism to these metrics and the ability of these metrics to predict a future business reality. Many engineering-minded readers might rightly point out that reducing all company success to investments in sales and marketing is dubious.
Kaltura does make this point as well, stating “In order for us to continue to increase revenue within our customer base, we will need to maintain engineering-level customer support and continue to introduce new products and features as well as innovative new use cases that are tailored to our customers’ needs.”
More targeted cynicism is appropriate for the metrics themselves. Extraordinary claims require extraordinary evidence.
LTV and CAC are not accounting terms. Definitions vary, and these figures are not audited. In no way are we impugning the Management team presentation, we expect some calculation does arrive at these values. However, the question is how representative are these values for the future economics of Kaltura’s growth.
There is little to decide the matter in the S1. The totality of the explanation of the “5x > 7x > 11x” is two sentences in a 140-page document, which is roughly the equivalent space devoted to the discussion of the potential impact of inflation. Here are those two sentences.
For any given period, we define lifetime value of our customers as annualized recurring revenue new bookings (ARR (as defined below) from both new customers from the beginning of the given period and upsells to existing customers), multiplied by the gross margin on our recurring revenue, divided by percent gross churn (the percentage of existing subscription revenue that was either downgraded or cancelled in a given period). To calculate customer acquisition unit costs, we divide our selling and marketing expense with a one quarter lag to the calculation period by the gross profit generated from recurring revenue new bookings for the calculation period.
All clear?
The definitions used by Kaltura clarifies a few matters (but only a few). Kaltura is focused on customer revenue as opposed to a count of customer numbers. In other words, there is no segmenting the inputs to focus on gaining new customers as opposed to gaining more revenue from all customers.
In practice, Kaltura is much more expand, than land in terms of sales success.
81% of 2020 revenue came from customers Kaltura already had using the platform in 2018. If the customer acquisition calculations were restricted to just new customers, then the calculations may look far less flattering. Presumable much of the $55 billion market opportunity is ‘new’ customers so then a discussion of new customer acquisition may prove valuable.
Of the inputs described in these LTV and CAC calculations, Kaltura discloses gross margin for subscription revenues and sales and marketing expense (though not those targeted at new customers). Bookings, quarterly ARR, and gross churn are not available. There are proxies we can use for bookings and quarterly ARR, but nothing available to approximate gross churn. Moreover, there are only two full fiscal year financial disclosures in the S1. Very little is provided about 2018 (nothing for years prior) and quarterly income statement figures are only offered for the past eight quarters.
There is then no means to verify, better understand, or stress test these metrics. We can only approximately a calculation (using Kaltura’s definition) for the CAC ratio in 2020 of 1.1x (anybody interested in this calculation, feel free to contact me; this post is long enough). This would mean it takes Kaltura a bit more than 12 months to recover its sales and marketing investment. A figure, experience tells me, is laudable.
The calculation themselves cause some discomfort. Perhaps we are misreading the calculation, but the denominator described in the document does not align with what is needed to present a LTV: CAC calculation. In this instance the language describes the CAC ratio, which has the incorrect units to derive LTV: CAC (we are welcoming a correction on this point, if appropriate). A more impactful observation is the lack of any discount rate in the lifetime value calculation. Meaning the customer lifetime is presumed in perpetuity (never observed in the wild), and thus inflated. The lack of discount rate makes the figure dubious in the abstract.
The larger issue is the true value of metrics in in allowing for benchmarking, which is to say a consistent application of a calculation across a representative peer group. People understand things in terms of comparison. No benchmarks are offered. In fairness to Kaltura, there is no habit in the media technology industry of communicating in these metrics.
We applaud Kaltura for citing these metrics. There is always a natural asymmetry in IPO disclosures. And this is something quantitative to serve as a basis of a larger qualitative discourse. It supports many useful, next questions, such as:
- How have these metrics varied on a quarterly basis?
- What is the CAC for truly new customers?
- How does current CAC in existing markets compare to new industry verticals?
- How do you account for the variation in usage-based components of revenue?
- If these are the unit economics, why have you not run much, much larger losses in the near-term to make the necessary investments in sales and marketing to take advantage of these favorable circumstances?
Another fascinating claim made by Kaltura is the equating of unit economics across its industry verticals. Kaltura reports revenue and gross profit performance across two industry segments: Media & Telecom (“M&T”) and Enterprise, Education & Technology (“EE&T”). Directly after the two sentences provided on LTV and CAC there is statement on how the M&T industry segment compares to EE&T. “Moreover, while our EE&T and M&T segments have different gross margin profiles, we believe their unit economics are similar.”
Our read of the document finds no disclosed data points to affirm this statement. There could be data points we have overlooked or that are not disclosed. The subsequent table lists several of the data points comparing these vertical markets.
Source: Kaltura S1, Devoncroft Analysis
The contrast in subscription gross margins between the verticals (30 percentage points or 3,000 basis points) is stark. Management attributes the significantly lower gross margins for subscriptions in M&T to a higher proportion of licenses deployed in customer private cloud and on-premise servers. One might expect (as the author would) not having to host the solution would raise the gross margin profile. Since license deployments have a lower gross margin, it would then suggest Kaltura has taken a meaningful margin historically from the resell of the underlying infrastructure.
An analysis of how the sales effort varies by industry is a critical component to evaluating the ability of Kaltura to efficiently grow into its stated $55 billion addressable market. By launching products in new verticals, you do not just increase addressable market, you also increase sales expense and stretch other operating categories. Often out of proportion to the new sales opportunity, since industry verticals have distinct go-to-market channels and partners.
How the business model is set up to scale is the final matter we will choose to discuss from the S1 filing. The reader may recall the letter to investors, highlighted earlier, mentioned operational efficiency. Operational efficiency and proper pricing models go a long way to allowing a business model to scale.
Kaltura’s pricing model includes usage components such that its revenue increases with the customer volumes on its platform. For example, media customers are charged on a per video play basis and telecom customers are charged on a per end-subscriber basis.
Because of the exigent circumstances in the second quarter of 2020, Kaltura experienced an increase in the usage of its platform from both existing and new customers. Certain of its customer agreements, the S1 highlights education, did not limit usage or charge for excess usage. This situation negatively impacted gross margins. The environment of 2020 also caused Kaltura to accelerate its existing plans to transition from its data centers to the public cloud. In the words of Management to “provide required stability, reliability, scalability, and elasticity.”
These factors (unbilled usage and the move to the cloud) contributed to a 300-basis point decrease in gross margin in 2020 (versus 2019). The decrease in gross margin is better appreciated as a transition of financing technical infrastructure through the income statement as opposed to the balance sheet. The near-term negative impact of the transition was accentuated in 2020 by the abandonment of existing data center. It resulted in a one-time nearly $4 million charge to general & administrative (“G&A”) expenses and in turned caused a 49% increase in G&A for the 2020 year.
A financing relationship with a certain public cloud provider is more evident when viewing Kaltura’s contractual, yet off-balance sheet commitments. Kaltura has non-cancelable purchase commitment to unnamed public cloud provider in the amount of $40 million thru December 2024, having minimum guarantees of $7 million in any year.
The cloud provider is not named, but Kaltura’s cloud provider of choice is named as AWS. The relevant excerpt from the S1: “While the great majority of our customers are deployed on our AWS public cloud infrastructure, some customers are hosted on separate and dedicated AWS private cloud environments that are also managed by us.”
To describe the public cloud as a “computer in another room” is to speak nonsense. Language is a poor tool to describe the breathtaking economic scale and cost of capital of public cloud providers. Kaltura’s outstanding term loan has an interest rate of 4.5%; borrowing from AWS is a superior alternative.
We are enthusiastic for future disclosures by industry suppliers on sales metrics. It adds a richness to the dialogue, and it provides a reflected glimpse into the buying process of end-users, where data-driven discussions are absent.
We hope you will join us at the upcoming Devoncroft Virtual Summit where we will explore these topics in much greater depth.
Related Content:
© Devoncroft Partners 2009 – 2020. All Rights Reserved.