Enterprise software: not dead, but different
One of my portfolio companies that I have been very involved with is ownCloud which develops (both open source and commercially licensed) enterprise software; it runs within an enterprise’s perimeter (in data-centers or in private/hybrid clouds). The technology allows you to access files regardless of where they sit on the enterprise network without having to replicate them to a cloud storage location as would be the case with Box, Dropbox, Google Drive, Office 365, or one of the myriad vendors that operate in this market. For example, unstructured data residing on your Sharepoint servers can be shared with co-workers or customers as simply and elegantly as sharing a file on your laptop via Dropbox can be. But ownCloud does this directly with the data at source, with all the controls and compliance constraints that a mature IT organization needs. This business has been growing, and fast!
But are we SaaS-y?
Growth begets investor attention and one thread of my discussions with them has been about the prospects for software that is not SaaS.
Isn't everything going to the cloud? Why invest in any enterprise software but SaaS?
This merits a broader discussion and hence this post. Take a look at this data from RightScale which embodies the perception that prompts the questions above. If 89% of enterprises are using the “Public Cloud” it must be surmised that the future is bleak for infrastructure and software vendors providing on-premise solutions.
Then again, take a look at this data-centric view of cloud adoption from EMC.
Indeed, most companies are dipping their toes (or perhaps their ankles) in what public clouds offer and the rate of growth within that niche is high, but as of today, less than 10% of enterprise data resides in the public cloud or in SaaS hosted silos. The rest of the data either continues to live where it has for decades or is increasingly moving into private/hybrid virtualized environments. This data is still within the control of their IT departments and subject to constraints of legacy, compliance and regulations.
The state of the Enterprise Cloud
Although as investors, operators and entrepreneurs, we have all been skating towards the idiomatic future position of the puck (an all-SaaS enterprise), there continues to be a need for software, services, systems management and IT infrastructure to meet the evolving needs of the on-prem enterprise. A few secular trends create a case for enduring enterprise technology residing within the firewall or with hooks into the “on prem” data even if the compute layer is hosted in the cloud.
Core computing : Greenfield applications, and insular use cases like CRM, HRMS, expense management, etc., are increasingly moving to cloud centric and SaaS compute models. But mission-critical applications that transact core business functions (storage and compute) continue to stay largely on-prem in most large and large-ish enterprises. A lot of these applications embody workflows where data moves from silo to silo within an organization. The data these applications consume and generate continues to live on-prem and grow. As mainframes, AS-400s and a legion of legacy apps have become more aware of the broader Internet, they need a modern layer of infrastructure and software to manage and derive value from.
Data Privacy: Nations around the world (the EU in particular) have led the way in establishing data privacy rules. Until recently, Safe Harbor treaty provisions allowed US based SaaS vendors to operate on foreign citizens’ personal data (including notionally innocuous elements like names, email addresses, etc.). Those protections have now been revoked but in my experience many large overseas enterprises, and governments were already wary of SaaS solutions as there is a perception of fewer privacy protections. Increasingly, SaaS applications punt on this by shifting the compliance onus back to customers. For example, MailChimp, the popular bulk mailing SaaS vendor lists recently sent me a notice about changes to its privacy statement:
COMPLIANCE
14. We Operate in the United States
Our servers and offices are located in the United States, so your information may be transferred to, stored, or processed in the United States. While the data protection, privacy, and other laws of the United States might not be as comprehensive as those in your country, we take many steps to protect your privacy, including offering a data processing agreement. By using our Websites, you understand and consent to the collection, storage, processing, and transfer of your information to our facilities in the United States and those third parties with whom we share it as described in this policy.
Data Residency: Where the data resides (especially in a SaaS providers’ data-centers), can have unexpected consequences for a business. Increasingly, courts and regulators have developed a view that access to a server over the Internet makes data on it “fair game” regardless of privacy protections offered by the country where it sits. The most (in)famous example is one where Microsoft is being asked by a US Court to turn over data housed in an Exchange server in Dublin, because it can be accessed from the US (remotely). The court order ignores the stronger privacy laws afforded by the data server residing in Ireland. So there is a fair amount of paranoia about where business information lives and who can access it [more on this topic here].
The logical reaction from CIOs to the privacy and residency concern has been to cling to control over their data assets. And yet their users are asking for more modern and consumer grade access to the information therein.
Cost: As many cloud native startups realize, costs of cloud computing begin to become substantial fairly quickly. At ownCloud, the amount of data we move via on-prem integrations would be cost-prohibitive for our customers to replicate and share on a public cloud storage system; we leverage existing infrastructure and optimize bandwidth usage. CIOs are waking up to the history of their cost increases with IaaS vendors and beginning to compare their TCO over 3 to 5 years of capex planning with what they would incur with public cloud vendors (more on this in a separate post if there is enough interest in the comments below).
The opportunity
As the venn diagram above indicates, enterprises have been moving more towards hybrid and private clouds. This allows them access to all the benefits of cloud computing while deriving other benefits not addressed by pure public cloud solutions (recounted below).
- Leveraging the size, scale and existing investments in legacy data stores
- Working within the scope of the silo’d nature of enterprise data
- Support for applications that are structured as workflows across these data silos
- Cost-efficiency of the private/hybrid clouds for certain workloads
- Maintaining their regulatory compliance, data privacy and data residency footprint
These factors point to the existence of an enduring enterprise software opportunity to provide modern full-stack functionality, analytics, APIs, mobility, and consumer grade UI/UX layered on top of the mountain of existing enterprise data.
There are a few different models for success (if valuations are a measure) that one can point to within the “Unicorn” club. Note that my categorization below and my choices are subjective.
- Gathering and analyzing enterprise data: Palantir, Domo, Jasper (now part of Cisco), Mu Sigma
- Application integration, virtualization, management: Tanium, Mulesoft, Simplivity, AppDynamics, Docker
- Databases: MongoDB, Cloudera
- Data storage and infrastructure: Pure Storage, Nutanix, Actifio, Datto
- Security and threat management: Illumio, Forescout, Zscaler, Okta
These are a subset of the many enterprise focused businesses of course and I have consciously excluded Slack, Github, etc., which do not necessarily operate on legacy data in the enterprise.
At ~10% of the Unicorn-club, I believe the “behind the firewall” enterprise infrastructure club is punching above its weight, especially if you consider that the list is otherwise dominated by players in the much larger global consumer opportunity.
Parting thoughts
Enterprise software is not an easy path to take: it takes a lot more capital to prove early viability and there are real hurdles like incumbent technologies and long sales cycles. However, unlike the “home run” ethos of consumer tech, it is not winner-take-all (usually), takes a lot less capital overall to build a business, and has yielded “singles” and “doubles” more readily for entrepreneurs and investors. SaaS with cloud hosted data will likely prevail eventually, but there is a lot of value to be created based on what lies within the corporate firewall in the interim.
