"Their model, as articulated in the essay (“Understanding SaaS: Why the Pundits Have It Wrong” from A16Z) assumes that ‘SaaS’ is delivered as a branded, downstream end customer offering—-with its attendant demand generation ‘ground game’ and high cost of sales (& marketing). What they are describing is really SaaS 1.0. The alternative model—-the SaaS 2.0 ‘passing game’ that I recommend innovators embrace circa 2014—- is best described as a next generation “XaaS” or anything as a service model—-characterized by an upstream partner-advantaged marketing approach." Joseph Bentzel, Software is Eating the Income Statement
All As-a-Service IPOs Are Not Created Equal
For "as a service" enterprise unicorns, it's the best & worst of times in sales & marketing. First, the best of times.
A subset of enterprise tech IPOs are on track to enjoy true landgrab economics and category leadership. These building block players enjoy revenue growth and superior revenue efficiency--burning only about a quarter of GAAP revenue on sales/marketing spend. This breed of unicorn chooses to creatively monetize their API & platform innovation inside the apps, clouds, networks, systems, solutions and experiences of developers and partners.
But others--those "new stack" unicorn IPOs that go to market as direct-to-enterprise SaaS--are burning 60+, 70+, 80+ percent of B2B subscription revenue on sales and marketing spend--the single biggest expense item on the income statement.
And some among the high-burn unicorn IPOs have decided (I believe unwisely) to no longer even call out "subscription" revenue as a separate line item in their public quarterly reporting. Instead, they've chosen to bury this critical metric of the direct-to-customer SaaS model to avoid an inconvenient truth--The direct-to-customer SaaS model's poor revenue efficiency and chronic over-dependence on investor capital.
I know this assessment runs counter to many "thought leaders" in direct-to-enterprise SaaS, but I'm not completely alone in thinking this way. One popular VC has opined that "resource constraints" (i.e. less capital) consistently challenge startups to better execute--and that the opposite condition--lots of excess capital--can render startups "weak and unfocused" in the face of competitive challenges.
Good words to heed, especially for the audience I seek to serve--"the Startup 99%", i.e. founder-led, emerging category startups seeking reduced dependency on venture capital, and laser-focused on operating self-sufficiency--Not the hunt for the next VC check to keep the lights on.
So let's briefly explore these two distinct sales/marketing spend patterns and see where the lessons lie.
Cloudera or Twilio: Which Sales/Marketing Spend Pattern is More Revenue Efficient?
To frame the differences in "As a Service" sales/marketing spend patterns, I've selected 2017 big data IPO Cloudera, and 2016 communications-as-a-service IPO Twilio. Let's start with Cloudera.
If "data is the new oil"--as all those big data "growthhackers" tell the story--and top line business is as good as the Cloudera income statement (below) suggests, you'd think that carving out a sober spend path to profitability would be a slam dunk.
But in the 9 month period ending October 31st, Cloudera burned $236+ million in sales and marketing spend (including stock-based compensation) to generate $216+ million in revenue and $344 million in total operating losses. In the most recent quarter reported, the 3 months ending 10/31/17, Cloudera sales/marketing spend has settled at 68% of total revenue (which includes professional services engagements) or 82% of subscription revenue.
While the 68% of total revenue burned on sales/marketing spend (Note: Reported revenue includes 1 time integration/customization services) is high enough, it's the 82% of subscription revenue that is more interesting. For direct-to-customer SaaS players, I think it's important to get a handle on sales/marketing spend as a percentage of subscription revenue because that subscription number is a "north star metric" for a vanilla "SaaS 1.0" revenue design pattern. When subscription-based business models that IPO don't report on "Subscriptions" in their income statement, that's a red flag. It tells me they may be getting a little fuzzy about the moving parts of the business (even as they improve their overall efficiency). Open is better.
And Cloudera management is fully aware they need to get their current sales/marketing spend number waaaaaayyyy down and are projecting a "long term" goal of 30-34% (of all revenue including PS engagements) in their recent quarterly presentation. Given their current direct-to-enterprise revenue design pattern, this goal does NOT seem very realistic as the company expands into new markets. Here's why.
Despite pitching their partner ecosystem as key to their success--including in new initiatives like IOT and edge computing where big data players see landgrab opportunity--the overall revenue "architecture" that is conveyed by Cloudera's marketing/sales spend numbers is one of a high-burn direct-to-enterprise sales force--multiplied by all those other revenue dependencies--e.g. the increasingly steep cost of direct demand generation and top line funnel creation in incumbent-contested or "high tectonics" enterprise markets.
Now let's turn and look at an alternative sales/marketing spend pattern, i.e. an "As a Service" model that engages developers at the top of the funnel (as does Cloudera) but operates more efficiently burning only a fraction of the marketing and sales spend of Cloudera. I'm talking about the Twilio model. I've provided their most recent quarterly income statement below.
Twilio--in the 9 months ending September 30, 2017--generated $283+ million in revenue and only burned $73+ million in sales/marketing spend. In its most recent quarter, Twilio allocates around 26% to sales/marketing expense.
It is my view that the kind of exceptional revenue efficiency found in the Twilio 10Q--compared to what we see in the Cloudera income statement--is based on Twilio's embedded XaaS building block product model and overall revenue design pattern--which runs through a developer self-service community and various major account upstream partners that embed Twilio CaaS building blocks inside their solutions. As these developer partner accounts scale and win, Twilio scales and wins.
This is what I describe as a partner-first revenue design pattern, as distinct from a "user first" model--I.E. a growth model that strives to leverage pre-existing and incumbent ecosystems, and their associated landscape capital--to grow a digital building blocks business at a lower cost of sales/marketing.
So when we compare their most recent quarters, we can see that there is a 42 point spread between Cloudera's 68% sales/marketing burn, and Twilio's incredibly efficient 26% spend.
I call this 42 point spread "the direct-to-enterprise SaaS revenue tax". And here's another notable observation relative to the Twilio spend pattern reflected on the income statement.
Twilio is 1 of only 2 enterprise IPOs included in this blog post currently investing MORE IN R&D, than sales and marketing. Back when I was pretty and had hair (the tech equivalent of the Bronze Age), this used to be an important determinant of what an actual "tech" company income statement looked like.
So the question is--Relative to other "As a Service" and new stack IPOs in recent years, are others more like Cloudera's model--They pay the steep direct SaaS revenue tax--or are they more like Twilio?
Burnapalooza S/M Spend: Rule or Exception
Many recent unicorn IPOs more closely resemble Cloudera when it comes to high sales/marketing burn and poor revenue efficiency. They are paying the direct SaaS revenue tax, with one notable exception.
I've selected Mulesoft, Okta, MongoDB, New Relic, Box, and SendGrid.
All are category-defining "new stack" technology innovators with either an "as a service" or other cloud-powered revenue model that have IPO'd in recent years.
Mulesoft: 65% of Total Revenue or 82% of Subscription Revenue Goes to Sales/Marketing Spend
Like Cloudera, 2017 IPO Mulesoft--a pioneer in iPaaS or integration platform-as-a-service--burnt 65% of total revenue, and 82% of subscription revenue on sales/marketing spend in its most recent quarter. And like Cloudera, Mulesoft primarily relies on direct-to-enterprise sales to drive their SaaS model. Here's how they put it.
"We sell to organizations worldwide primarily through our direct sales efforts. Although our platform can be adopted by organizations of nearly any size, we focus our sales efforts on the largest global organizations. Our sales efforts are targeted towards CIOs, chief IT architects, and line of business leaders, who are driving digital transformation. We also partner with system integrators (SIs) that enhance our sales leverage by sourcing new prospects and providing systems integration services on implementations of our platform." Mulesoft 10Q
This direct-to-enterprise IT model--supplemented by Tier 1 global SI's and consultancies--is the most expensive kind of revenue design pattern a tech company can build--Engaging intensely contested large enterprise accounts that are subject to lots of "market tectonics"--i.e. multiple competing monoculture incumbents all fighting for their piece of the land/expand enterprise pie.
But Mulesoft is in a good position to reduce its direct SaaS revenue tax and pivot to a more capital efficient revenue model via embedded "As a Service" partners. As the company continues to evangelize its "application networks" vision--Connecting SaaS, cloud, API, and on-premise IT via composable, plug/play software building blocks--Mulesoft is a perfect candidate to reduce its sales/marketing spend via a partner-based model in which software and web OEMs embed Mulesoft functionality inside their apps, clouds, networks, systems, solutions, and experiences.
Okta: 69% of Total Revenue or 79% of Subscription Revenue Go to Sales/Marketing Spend
2017 IPO Okta, a new stack leader in IDaaS or "identity-as-a-service", burned 69% of total revenue in sales/marketing spend in the 9 months ended 10/31/17--and 79% of subscription revenue in its most recent quarter.
As I reviewed the Okta quarterly report, what jumped out at me as I read between the lines is the fact that OKTA--like many category pioneers in the "as a service" economy-- is actually 2 companies in one.
One part of Okta is focused on direct sales to enterprises. The other part targets developers seeking to embed SSO and user authentication into their apps and experiences. Okta even bought an API company to expand this part of the business.
This second part of the Okta business--the part seeking to embed Okta functionality inside the apps, clouds, and systems of partners--can run at a dramatically lower cost of sales/marketing than the enterprise side (Twilio's income statement proves that). But this side of the Okta business receives little management coverage in their reports to the SEC and their investors. This is disappointing.
What is discussed instead of an embedded XaaS or OEM revenue conversation in their quarterly report is the Okta Integration Network--"which provides customers with a pre-integrated set of cloud, mobile and web applications that spans the functionality of our products. As of October 31, 2017, we had over 5,000 integrations with third-party software applications".
I get it--they're doing a creative version of an identity "marketplace" or modified "app exchange" strategy--complete with out of the box integration. But here's my question? Are Okta product managers and marketers focused more on having developers "bake them in" to their apps, or is Okta more focused on pre-integrating its platform with 3rd party apps at its own expense to grow their enterprise business?
Clearly the 10Q shows the focus is on direct-to-enterprise--and for that reason the "pre-integrations" marketing pattern will continue to mirror that of a direct-to-enterprise SaaS company, not an upstream-partner focused building block player like Twilio.
In its next quarterly report, Okta--as a public company still generating over $100 million a year in GaaP losses--would be doing its retail investors a service by beginning to break out the various parts of its business--enterprise vs 3rd party developer. Then show the relative revenue growth rates of each. This will also allow analysts and others to weigh in on the strategy behind its current high-burn investments in sales and marketing.
MongoDB: 70% of Total Revenue & 74% of Subscription Revenue Go to Sales/Marketing Spend
MongoDB, a new stack database provider boasting 30 million free downloads of its open source community edition, burned 70% of total revenue on sales/marketing spend in the 9 months ended 10/31/17, and 74% of subscription revenue in its most recent quarter.
MongoDB has strong developer community roots with its document database innovation--by virtue of the free open source community edition of its database. And an outside observer like me would think going in that it's sales/marketing spend pattern would more closely resemble Twilio's. But that turns out not to be the case. MongoDB is much more like Cloudera than Twilio in terms of high sales/marketing spend as a percentage of revenue.
Is it a "focus" issue? Let's see. MongoDB has multiple products in multiple segments in addition to its "Community" edition. The company has an enterprise edition, an OEM edition, a cloud-hosted edition (Atlas), a new backend API service (Stitch), and also markets its support services "as a product" (Professional).
While some would immediately say "There you go. Too many products for execution focus. Too many products for a company at that stage is what's driving high sales/marketing spend", I'm not one of them. All of Mongo's products are clearly extensions of its core and seem to be pegged to various use cases of their platform.
But I would like to see more discussion in the 10Q on how all the moving parts of the Mongo portfolio fit together, and/or gracefully reinforce each other in some meaningful, i.e. "increasing returns" way. If it turns out they don't have a "multiple moving parts" growth model thought through--e.g. their "As a Service" database vs their "on premise" marketing priorities--then the spend numbers will get worse, not better, in the next few quarters.
What I would suggest in terms of the various product line extensions of MongoDB is to look at their overall demand creation model and apply "marketing separation of concerns" to their different segments. I've found that there is a big spend difference between what is commonly described as "developer marketing"--i.e. long tail developer marketing to large sets of unknown developers (LSUD)--vs a true "3rd foothold" sell-through revenue design pattern (RDP) to "small sets of known developers" (SSKD), e.g. incumbent vendor partners with a pre-existing installed base and partner ecosystem.
Simply put, "developer marketing" often manifests as an overly expensive misapplication of B2C user marketing to B2B and B2D. It's too focused on the long tail of unknown developers, and not focused enough on landscape incumbents (e.g. "known" developers) and piggybacking on their ecosystems.
Twilio has a 26% sales/marketing spend rate--But also has free account churn in the high 90's percent. That kind of churn freaks out direct-to-customer SaaS players but is not a factor in what I call "landscape based marketing" or LBM. A Partner1st landscape based marketer assumes that developers are critical gatekeepers in a deal, but may not be the decision makers. So LBM practitioners prioritize "targets of asymmetric opportunity", i.e. product and "line of business leaders" at landscape incumbents.
We also believe that a sticky upstream partner buying decision is not always based on platform--but on "platformula"--the composite economic win for the incumbent partner who decides to embed your building block value prop inside their offering.
New Relic: 61% of Revenue Goes to Sales/Marketing Spend; "Subscriptions" Data Not Called Out Separately in 10Q
New Relic, a pioneer in enterprise application performance monitoring--and host of the FutureStack conference--burns 61% of GaaP revenue on Sales/Marketing spend.
New Relic went public at the end of 2014, so there is a lot of data available to assess their overall revenue efficiency. And while New Relic targets developers and DevOps professionals with its digital intelligence platform and suite of app/cloud performance monitoring solutions, it's income statement reflects a direct-to-enterprise SaaS player, not an embedded XaaS building block revenue design pattern as does Twilio's. This is also manifested in their focus on enterprise use cases like cloud migration, digital transformation, and digital experience delivery. But New Relic has recently (Nov 2017) staffed up and expanded their "channel reseller" initiative.
"Channel" is a common evolution of direct SaaS players, i.e. engage a set of downstream sales/services partners with domain expertise in a give focus area (vertical, regional, technology-specific). This kind of partner approach can work to grow top line revenue, but also can be increasingly expensive to stand up--if things like "account protection", MDF (market development funds), and sales rep compensation issues are on the table to "tax" your revenue efficiency. This is tougher to do when you're still burning about $60 million a year in GaaP losses 3 years after going public, as is New Relic. By the way, this 3 years public and still burning big dollars on marketing and sales is why "marketing automation" player Marketo ended up going private again. Hard for shareholders to justify the burn forever, especially if it's perceived to be baked in to the direct SaaS approach.
One of the main reasons I evangelize upstream partnering, e.g. OEM, white label, APIs, MSPs, major cloud marketplace programs--vs legacy "channel" partnering--is that this ingredient growth model is accelerated with every partner deal you do--progressively driving down your overall sales/marketing spend.
When a digital "ingredient" (aka building block) provider finds a home inside the apps, clouds, networks, systems, solutions and experiences of its sell-through partners, this triggers what I call a "partner DNA inheritance" that can drive your next deal.
This partner DNA inheritance also triggers landgrab economics, as it did for Microsoft with every new "cloner" deal it did, and with Google as it made its search box an ingredient on every portal or ISP homepage.
I see this flavor of Partner1st "path dependency"--i.e. tapping into and monetizing the pre-existing market power of incumbents--as optimal for the Startup 99%. This is especially true given the investment priorities of Tier 1 VCs for oversized exits for only a very small subset of their portfolios. Translation. In the future, there may be dramatically less tolerance for subsidizing high-burn direct-to-customer SaaS by investors.
Box: 61% of Revenue Goes to Sales/Marketing Spend; "Subscriptions" Data Not Called Out Separately in 10Q
2015 IPO Box burns 61% of GaaP revenue on sales/marketing spend, and does not call out "Subscription" revenue as a separate line item on their income statement.
As I said above in relation to Okta, Box is two "As a Service" companies in one. On the one hand, Box is their original user-based freemium marketing model competing with companies like Dropbox. This B2C approach requires lots of capital just to survive, let alone get to something resembling operational self-sufficiency and a category landgrab.
On the other hand, Box is providing platform building blocks for the app and content experience developer that embeds & monetizes Box services inside its apps and solutions. So Box is a company in transition from a high burn freemium user monetization model to a lower burn platform building block model like Twilio.
I describe this monetization pivot from end customers to upstream partners as XaaSification in my essay XaaSify Your Startup. This is the kind of transformation many direct-to-customer SaaS startups will have to go through to avoid running out money in a different and more challenging VC funding environment.
In addition to their developer offering, I think the kinds of co-marketing relationships Box is building--e.g. with IBM, Google, and Apple--are excellent examples of how to "get grounded" on the landscape topography of incumbents that already have extensive ecosystems and installed customers. This is what I call "landscape attach".
It will take Box a while to get to the kind of revenue efficiency exhibited by Twilio--But they can get close over time if they keep the focus on attaching to pre-existing landscape networks, and pushing the developer platform model.
If it were my call, I'd also consider selling off the freemium user business to an upstream partner and drive the spend way down. As the ECM and WCM markets go cloud, API-based and headless, Box should be in a good place to grow even further, and replace some of the old guard ECM players.
SendGrid: 26% of Revenue Goes to Sales/Marketing Spend, Mirroring the Twilio Sales/Marketing Spend Pattern
Cloud-powered email platform provider SendGrid burned 27% of GaaP revenue on Sales/Marketing in 2016, and spent 26% of revenue in the first half of fiscal 2017, according to the company's S1 IPO filing. Yes. Twilio's model is not a fluke, it's a "pattern" for the next generation of cloud-native landscape based marketers.
Like Twilio, SendGrid is an As-a-Service building block innovator, and markets its high volume email platform as an API consumed by its ecommerce partners, OEMs, and digital agencies--inside their apps, solutions, and experiences.
SendGrid is a company that has become well "grounded" on the landscape topography of incumbent ecosystems--Helping high volume email marketers (We call them "basehackers" ) serve personalized messages to their customers and members. It provides what I call a must-have or "ticket to the game" value proposition.
And SendGrid, like Twilio, is growing with less capital dependency, and dramatically higher revenue efficiency--than those paying the direct-to-enterprise SaaS growth tax. They do the kind of things that every XaaS startup on the planet should be doing-- e.g. piggybacking their business on incumbent cloud marketplaces in addition to hosting their own.
So, let's close with two takeaways.
Takeaway 1: The Startup 99% Need to Stop Paying the Direct-to-Enterprise SaaS Tax
I see the 6 "As a Service" IPOs (Box, New Relic, Okta, Cloudera, Mulesoft, and MongoDB) that currently experience high sales/marketing spend--60+, 70+, 80+ percent of subscription revenue--as examples of the victory of "strategic finance" over "strategic marketing" in tech land. One VC succinctly described this high burn approach as a "steroid startup" model in which growth and adoption are "subsidized".
The Startup 99% can not afford to bank (no pun intended) on a VC-subsidized growth model. They need to prioritize Partner1st self-sufficiency--at the start--and internalize the lower burn marketing lessons of building block XaaS winners like Twilio and SendGrid.
Takeaway 2: Direct-to-Enterprise SaaS Players Need to "XaaSify" Their Startups
For example, the Box content-as-a-service platform business can enjoy lower sales/marketing spend as a digital building block XaaS business--than as a freemium SaaS user business. And the Box business can become much more valuable to its ecosystem stakeholders over time, as those stakeholders monetize Box value inside their own applications and experiences. That is what is clear when you study the evolution of Box, post-IPO.
Joseph Bentzel is the founder and senior consultant at Platformula1.
If you'd like to learn more about Landscape Based Marketing (LBM) and build a Partner1st startup or enterprise digital initiative that capitalizes on non-stop landscape tectonics, contact him at Joe@platformula1.com or follow us on Twitter @Platformula1