When I started at Hotmail in 1996, we were thirteen people, and frankly, we didn’t know what we were doing… The company had recently launched the first Software as a Service (SaaS) application to the world, and we were struggling to keep it up and running.

Everything about it was just so different than the traditional software we were used to:

  • We had one version of the software that was being used 24/7, as opposed to 5-10 versions.
  • We rolled out new versions every few weeks, versus months or years.
  • The end users were part of the quality assurance process and we were in a constant state of rolling new bug fixes, as opposed to a huge QA staff and a long process.
  • The software engineers had to learn to collaborate across functions with the data center techs, who rolled out the code, and the customer service group that was getting the customer QA data — compared to before, when those organizations were siloed.
  • We could see exactly how the end users used the software — real time — which dramatically increased the innovation cycle. This was very different from past approaches, where you had to do focus groups to figure out how an end user was using the software.

These were major shifts, and we weren’t sure how to organize for success … but somehow, we muddled through it.

As it turns out, though, we weren’t the only ones — just about everyone involved in the early internet era who had come from traditional, waterfall-centric programming environments started inventing new ways of doing things that were more in keeping with the always-on, fast-release iterative cycle that especially suited the world of the nascent SaaS and web-based software environments.

In 2001, a group of developers met to debate and discuss a new set of “lightweight” methods for software development, and the Agile Manifesto was born. While the manifesto wasn’t necessarily a direct outcome of the new web/SaaS/cloud organizational meanderings, it struck a nerve by stressing the need for cross-functional collaboration, communications, and short release cycles. Essentially, it helped to codify the hodgepodge of learnings we independently discovered about what worked at Hotmail, Yahoo, and other first-generation internet companies. These learnings ultimately became the underpinnings for what we now call “DevOps” [adapted from Wikipedia]:

DevOps (a portmanteau of “development” and “operations”) is a software development method that stresses communication, collaboration, and integration between software developers and information technology (IT) professionals. A response to the interdependence of software development and IT operations, DevOps aims to help organizations rapidly produce software products and services — and to improve operations performance.

But DevOps is more than just a methodology. It’s a must-have skill set for the modern programmer — and is increasingly becoming its own department as well (the subject of much debate). The rise of the hyperscale cloud datacenter has made this job much harder as developers have had to hack together tools and complex scripts for pushing code to thousands of pancake servers.

The growth of the DevOps movement coupled with this complex cloud infrastructure has opened up an opportunity for a company to own the entire process and help developers and managers manage it.

It is with this backdrop that I am pleased to announce Andreessen Horowitz is leading a $2.8M Series A in Distelli, an infrastructure automation company that I like to think of as the “DevOps dashboard.” Here’s why we’re so excited about the company:

  • It was founded by Rahul Singh, one of the early engineers on the Amazon Web Services team. Rahul is the epitome of one of our most important criteria — founder/market fit. He has essentially solved a set of problems he encountered during his nine years working on Amazon’s cloud.
  • Rahul has a grand vision for Distelli to become the DevOps platform. Application, database, and middleware deployment are just the first tasks that Distelli automates, and they’ve got a rich roadmap to automate the other things DevOps does.
  • Shortly after the first beta, the product was of such high quality that every proof of concept resulted in a paying customer. The customer references raved about how it was simple, scalable, and consistent across multiple cloud and on-premise installations.
  • When it seems like every smart entrepreneur we meet seems to adopt the same new technology at once — indirectly demonstrating it as part of their pitch — we pay attention. A few of the companies in our portfolio that hold this important distinction include Good Data, Mixpanel, Okta, Optimizely, and Zenefits; and now, Distelli.

Rahul is one of those “10x founders” (as we like to call it) that just get shit done at an astounding rate. (These are the founders to whom you’re about to share feedback you’ve heard or suggest a new feature and they’re like, yeah, we already did it.) Every month I check in with Rahul, it feels like he’s progressed ten times more…

I’m so pumped to be joining the board to help Rahul grow into a powerhouse, and am pleased that Distelli is joining the a16z family. Onward!

There is a perfect storm of three distinct disruptive forces that has the potential to topple nearly every major enterprise software incumbent. And the traditional approach of dealing with technology shifts – through acquisition – looks like it’s headed towards failure. As such, there is an unprecedented opportunity to create many, new multi-billion dollar enterprise franchises that are on the right side of these forces and are willing to go the distance in the face of ridiculously high acquisition offers.

Let’s examine these forces individually:

Software as a service (Saas): Seemingly a little long in the tooth as a disruptor, Saas has finally gone mainstream in the Global 2000. The primary disruptive force of this technology is the speed of innovation. The feedback loop is especially powerful: as opposed to using focus groups and surveys to figure out how users are interacting with the product, Saas companies can see what their customers are doing real-time by capturing and analyzing every click. They quickly extend their products through a “cell division” that continuously builds out and A/B tests the features that are getting the most engagement. On-premise and client (PC) software-based product cycles can’t possibly compete here as new releases are typically pushed 10 times faster at 45-60 days vs 18-24 months. There’s always one version/code base so it’s much easier to support, patch bugs, and roll out new features to all customers at once. The old joke of “How did God create the world in 7 days? He didn’t have an installed base!” certainly applies – but Saas also demands entirely new skills sets associated with running a 24×7 services business. Dev/Ops, customer care centers, network operations and delivering uptime via failover, mirroring and hot backups are all new and essential. It’s easy to see how the early Saas pioneers gained so much ground with this innovation but even they are unprepared and poorly architected to take advantage of the additional disruptors that have hit more recently…

Cloud infrastructure: As I detailed in a prior post, “The Building is the New Server,” the humongous internet powers, Facebook and Google, are literally breaking new ground in re-imagining the design, components and cost of running a hyper-scale data center. The cloud infrastructure they are pioneering has the primary disruptive force of massively driving down cost. Facebook, for instance, is experimenting on the bleeding edge of solving the new cost bottlenecks of power and cooling. I recently read that it actually rained inside one of their datacenters. The cloud service providers (CSPs) are following their lead using commodity components, open source software, data center design and testing software defined storage and networking products to enjoy the same, devastating cost curve. The corporate datacenters (aka “private clouds”) will slowly disappear as Global 2000 companies migrate to these irresistible new cost curves. Don’t be fooled that security and reliability concerns will keep large enterprises away – as the CEO of IronPort, I watched in horror as large enterprises started pointing their treasured Mail Exchange (MX) records to cloud services like Postini – a much superior and vastly cheaper cloud based architecture versus our perimeter appliances. And email is the most sensitive and mission critical of applications…

Mobile: About two years ago, all of our consumer companies went through an “Oh shit!” moment with mobile. One year mobile was 10% of traffic and the next year, when everyone was expecting ~20%, it was 30% on it’s way to 50%. Facebook, for instance, famously bought Instagram for $1B and then continued their pursuit of talent to redesign for mobile. The new mobile operating systems and devices are proliferating an entirely new interaction and design paradigm that has the primary disruptive force of a re-imagined user interface. The innovative use of touch/gestures (e.g. pull down, swipe, pinch etc.) pioneered by the consumer applications will become de rigor for enterprise as well. Although it’s still early, the mobile sensors (e.g. GPS, accelerometer, video etc.) will also become integral and spawn new innovations in the enterprise as they have enabled new consumer franchises like Lyft and Instagram. The number one problem facing so many of the startups I talk to is hiring the design talent (e.g. Mobile app, front-end engineering and user interface) to take advantage of this trend. In addition to being in ridiculously high demand, most of these people are “arteests” who eschew just cash and stock as incentives because they want to work for a purpose and in an environment where design is an overarching priority/core competency – not something that is grafted on afterwards. These environments are hard to find.

So exactly why won’t these big incumbents make it to the other side? There are just too many things changing at once. Beyond the technology changes, there are structural impediments as well. The incumbent sales forces have become farmers instead of hunters. They still sell on relationships (e.g. A round of golf, anyone?) and bundling/discounting instead of product attributes. They sell to the CIO instead of the line of business buyer who is making the decision. The quotas and incentives are too different. The accounting systems don’t speak recurring billing and revenue. Ugh – it’s just too much change…

A handful of exits have been priced based on a NTM revenue average of 11X vs around 4X for the rest of Saas companies. Examples include Workday, Splunk, ServiceNow, Marketo and Tableau. Not to mention the SuccessFactors deal (done at 11X) has officially kicked off the next wave of consolidation. On the private side, companies like New Relic, AppDynamics and ZenDesk have seen private transaction multiples of between 9X and 11X.

There is outright panic going on right now at the large incumbents as they pay ridiculous premiums for the early Saas companies. And so why won’t these acquisitions pan out? Most of the early Saas companies weren’t architected to take advantage of the cloud infrastructure cost advantages AND most completely missed the boat on mobile. It’s hard enough for new, cool enterprise startups to hire the necessary design talent but the large incumbents really have no hope.

Next Up

As I’ve said, there is a perfect storm of three distinct disruptive forces brewing which has the potential to erupt into a new multi-billion dollar wave of enterprise franchises. In particular, there will be at least 30 new enterprise franchises that will go the distance, resist high acquisition offers as they either supply or ride this trio of disruptors to dominance.

Amongst others, the new suppliers are companies like Cumulus Networks, Okta, New Relic and Nimble Storage. The “riders” are awesome trifecta companies like Box, Evernote, Base, Expensify and Tidemark.

Where will these 30 New Franchises come from? A double investment cycle in Saas, as the large incumbents buy the early Saas pioneers and fumble them, will pave the way. Like Lenny from “Of Mice and Men,” they will smother these companies with too much negative attention, mismatched salesforces, and misunderstood business models. Following a short vesting period, the product and management talent – who are used to working at a completely different pace – will ultimately leave the incumbent, resulting in a bevy of entrepreneurs that roll out to start even more of these franchises.

I can’t wait to meet them!


The personal computer is dead. As quickly as we moved from the desktop to the laptop, we are moving to the tablet — never to return. With the death of the PC, an entire ecosystem dies with it. The chipset is ARM based, rather than Intel. The operating system is all iOS and Android, rather than Windows. The applications are hosted cloud apps like Box, Google Apps, and Evernote rather than Sharepoint, Office and Outlook.

This is rocking the industry. Dell is being taken private — closing a curtain to start the dirty work of restructuring. HP, Microsoft, and Intel are all trading well off their peaks when the Dow has recently hit an all time high. IBM, looks like the sole winner, jettisoning its PC business years ago to China-based Lenovo. Well, it’s a good thing all of these companies also play a big part in the $55B* server market — that’s not going away anytime soon, right? The worst days are over and hopefully their collective market caps will recover? Not so fast…

Modern web services, such as Google, Apple and Facebook are pushing the limits of data center scaling to unprecedented levels as they deal with an exponential growth in user traffic. They are playing a massive game of Tetris as they grapple with deploying and operating data centers with tens of thousands of servers versus hundreds. They are all on the bleeding edge of trying to contain costs while cramming as much capacity into a physical building as possible. The result is a complete architectural rethink of datacenter designs and the incumbent server vendors are struggling to stay relevant in this new reality.

The new data center designs use only commodity “vanity free” components procured directly from the original design manufacturers (ODM) — the current incumbent’s suppliers. For easy serviceability, components are velcro’d together versus mounted in a box. All bells and whistles are stripped off and the hardware is purpose built for a specific application and therefore carefully tuned. As compute utilization rates skyrocket from virtualization and parallel processing, the CPUs are running harder and hotter and therefore the new expense bottleneck is all about power and cooling.

Take for instance, Facebook’s Open Compute initiative, which lays out a blueprint for an energy efficient hyper-scale datacenter that is 38 percent more energy efficient and 24 percent less expensive than current datacenters. Locating in cold climates and next to super-cheap hydro power has become de rigor. Power distribution, cooling and building layouts have been redesigned from the ground-up to maximize mechanical performance and electrical efficiency of the datacenter. And unfortunately for Intel, the relentless march of Moore’s Law no longer affords them differentiation, as customer needs have shifted from performance to power efficiency, an area where they lag rival ARM processors.

The evolution of the modern hyper-scale datacenter reflects the hyper-scale needs of the applications that run on them. Modern web 2.0 (and increasingly SaaS) applications need to handle thousands of user requests per second, processing terabytes of information in real-time across hundreds of customers. They are by necessity massively parallel and work in concert to service a user request. This is the modern equivalent of a giant supercomputer — except cobbled together from commodity server components and interconnect fabrics. It’s a profound software and hardware architectural shift that is taking us from a world where datacenters consisted of small number of independent high performance branded servers to a brave new world where the giant, datacenter building is the server.

Meanwhile on the enterprise front, the corporate datacenter is becoming increasingly sedate as on-prem applications give way to their SaaS counterparts.

The new data center architectures, borne of necessity from the giant web service providers, have the potential to massively drive down the cost of providing software as a service, the new winner in enterprise applications. As such, the cloud service providers (CSPs) such as Amazon and Rackspace, are adopting these “scale-out” architectures.

So fast forward: SaaS, will win the enterprise market. Face it — it’s just so much better and now, infinitely cheaper than any of the alternatives. And modern SaaS applications will be delivered through hyper-scale data centers that do not have branded servers from Dell, HP or IBM, but rather highly optimized, scale-out white box servers made by Asian ODMs. In addition, the operators of these massive data centers will be experts in servicing their creations — monitoring, fixing and rapidly swapping out their expected-to-fail components. Therefore, there will no longer be a need for the recurring revenue, high-margin service and maintenance contracts that have been a mainstay of the OEM server industry.

I wonder if Lenovo is in the market for a server business too?

: )

I would like to thank my partner, Ramu Arunachalam for his research, analysis, and material contributions to this blog.

*IDC estimates (2012)