Sep 09, 2020

Seagulls and French Fries: Lessons in Horizontal Innovation

Seagulls and French Fries: Lessons in Horizontal Innovation

Let’s be honest, the data protection market is thoroughly boring and relatively static. It’s probably one of the more mundane use cases in tech. But the space is about as crowded and noisy as 100 seagulls on a french fry because data protection is required by every organization. While backup and recovery might not be a use case traditionally associated with value creation, more of us own insurance policies than index funds. Progress in the space has largely been horizontal, extensive innovation consisting of trifling alterations and improvements. Key value drivers have been simplicity and cost avoidance. The backdrop for the recent purpose-built backup appliance (PBBA) craze in data protection was a market category where nothing new had been introduced for 15 years. When one thinks “value creation and digital transformation”, they don’t think “data protection”. IT leadership isn’t lying in bed at night, staring at the ceiling thinking: “If I could only rethink my data protection strategy, I could steer my organization through these uncertain times”.

Ten Thousand Silos and You Don’t Get a Single Lake

Data protection, if done the right way, is one of the few applications that touches and consolidates all of an organization’s data. It’s rarely done the right way. As data becomes increasingly distributed across on-premises data centers, public cloud and SaaS realms, data protection becomes just as siloed as the data it seeks to protect. Span of control and compliance are lost on account of point products deployed within each enclave. For example, there are startups that have taken a quarter billion in funding to address protection alone. A data protection company with a $5B valuation has made a name for itself protecting VMware in the data center at an average customer contract size of ~$2700. On the horizon is a whole crop of “data protection” companies built on the premise of orchestrating snapshots in public cloud and finding cost efficiencies for cloud data protection spend that can represent up to 30% of an entire cloud bill. Side note: there is a Dilbert cartoon in the waiting that captures the moment when a CIO suddenly realizes that the cloud seems a lot less transformational when a significant portion of cloud spend is on data protection.

Cloud Changes Everything. Unless Your Vendor’s Revenue Models are Predicated on Them Selling You a Box or a License for Three Years of Growth

Every major provider in the data protection space has been writing software with the assumption that it would need to run on a finite amount of resources. After all, there are only so many cores that can fit in a server or in a purpose built backup appliance. Few in this space had the luxury of starting with the assumption that the cloud is the computer. Limitless ephemeral compute and object storage capable of massive scale and parallelism create new opportunities to consolidate silos and make use of the data in interesting ways once in its protected space. What if the data being backed up landed in a platform built on infinitely scalable cloud resources? What additional value would you ask of the platform? A time-series, indexed data lake as a byproduct of data protection, maybe? Backup and recovery might actually start to sound, dare I say it, cool.

But the innovator’s dilemma is alive and well in the data protection space. Unfortunately, for the consumer, the traditional players have a CAPEX addiction and run rate expectations that have been defined by a legacy model of slinging physical appliances and software contracts that are sized and sold for several years of forward growth. When these vendors make a move to the cloud, they are going to lose a great deal of appliance revenue to the compute and storage upon which their software rides in the clouds. They’ll also have to shift to just-in-time OPEX consumption models. Beyond that, pivoting a salesforce from selling software and infrastructure to selling SaaS isn’t easy. This is especially true when the sellers don’t get paid on the cloud resources that back-end their virtual appliance software in the cloud. We recently spoke with a company where cloud costs to run a virtual backup appliance rivaled the PBBA software license itself. Neither number was easy to digest. Nobody wins in that scenario. The consumer is faced with paying for a hefty license and running it as inefficient IaaS in-cloud. The PBBA vendor and its sellers don’t capture revenue from the cloud resources required to run the software. While the cloud vendor realizes additional consumption, it’s for tactical, non-sticky creation use cases.

The Network Effect, Platforms, and Why SaaS Multiples are What They are

Traditional backup vendors selling appliances or software need to pivot or they will perish. Revenue models (no matter how healthy and margin rich) that run counter to how organizations want to consume are falling out of favor. This is evidenced by the revenue multiples awarded to software or hardware vendors in the data center infrastructure space. I cite NTNX and NTAP as examples of companies with decent operating margins and strong growth that are penalized for selling software or appliances as CAPEX deals in a world where pay-per-use consumption models are very much in favor. Nutanix did $1.24B in revenue in 2019. As I write this blog post, they are valued at ~$5B. NTAP did $5.4B in revenue 2019. They are valued at ~$10B. Accounting forward revenue, both companies have revenue-to-market cap multiples of under <3x. How will the valuations of appliance focused backup and recovery vendors be any different? The most notable unicorns in the PBBA space have done large private raises and tout lofty valuations. But the public market clearly doesn’t reward CAPEX based consumption models. Any future raises could be down-rounds that would dilute the equity for their employees. Morale will tank. Employees will leave. Innovation will wane.

Why do SaaS companies enjoy far greater multiples on revenue than their software and hardware slinging peers? I’m in the camp that subscribes to the idea that the network effect of platforms matters most.

At its base level, all SaaS providers seek to:

  • Eliminate the need for customers to install and maintain their own technology infrastructure
  • Provide easy to consume services that can be deployed rapidly and embraced at a lower cost of entry
  • Deliver a just-in-time, pay-for-use consumption model
  • Expedite product evolution and improvements given that the SaaS model does not require supporting diverse code bases and generations of hardware scattered about their customer ecosystem
  • Capitalize on the promise of the platform

The most valuable SaaS companies have evolved from product builders to platform providers. See ServiceNow, Workday, Splunk, Salesforce, Snowflake, etc. The platform network effect exponentially accelerates their ability to iterate and innovate. Opening up APIs and allowing third-party developers to build enterprise apps promotes agnostic experimentation and supports rapid innovation. Customers benefit as product gaps get filled with “plug-and-play” solutions – see SalesForce AppExchange, Snowflake DataCloud, etc. Moreover, the SaaS providers capture value and revenue from both sides of their marketplace by charging platform license and listing fees to developers, who in turn attain greater revenue growth than if their platform had not facilitated easy access to enterprise customers. Crossing the chasm from a single-sided application to a multi-sided platform is what takes a SaaS company’s market cap from a 10x multiple to a 30x multiple. It paves the way for faster innovation and additional revenue opportunities. The platform players get the benefit of the network effect and capture a larger market and far greater returns. Their product focused competitors are left with little to gain and compete almost solely on price.

A picture is worth a thousand words:

The above graph compares share prices of SaaS platform companies to data center infrastructure companies.

The chart below compares revenue to market cap for the same companies:


It should come as no surprise that Clumio’s board and investors come from platform focused backgrounds. I remember feverishly taking notes as one of those investors spoke to our team during an all hands call on my first day at Clumio. He is the CEO of a SaaS platform company. Four years post IPO, their company was doing well but its revenue largely came from perpetual software licensing for on-premises use cases. As excited as he was about their core product, he knew they had to pivot. They were too single product focused. They had too few APIs. Their ASP was ~$25K. He realized they needed to become a multi-product platform delivered on cloud. So they rebuilt their plane while it was flying by refactoring their core code and expanding their offerings. In doing so, they moved from hand-to-hand combat on individual product features to a full featured data platform.

His guidance to our team:

Start with a platform model that is as contained as humanly possible.

Focus on a SaaS business model.

Keep your foot on the gas while your product and appliance focused competition endures massive transformation to catch up.

Rest confident that Clumio started with a platform goal and has the backing and a board to achieve their goals.

As you look to partner with a data protection provider, take stock in how they’ve chosen to build. Are they focused on a platform and long term viability or are they seeking to protect a legacy approach to product development and how said product is consumed? It’s a tired old space and it will continue to be so until someone comes along and delivers a scalable platform capable of reducing cost and complexity while flipping to script to supplemental value creation. And that’s exactly what we aim to do here at Clumio. Thanks for reading and keep watching us deliver.