AWS Re:Invent 2014

AWS is a force. The claim is that they have 5x the capacity of all the other public cloud vendors – combined.  The pace of innovation on the platform, the quality of the services and the sheer scale on which they operate their datacenters is really impressive. James Hamilton who is a VP and Distinguished Engineer needs to write a book on the growth. It would be required reading for anyone doing anything in the cloud.

Think about this – Everyday AWS adds enough server capacity to support all of Amazon’s global infrastructure when it was a $7B business (2007). Every. Single. Day.

AWS has spent a lot of time on thinking about networking (and I think its potentially their biggest existential threat to growth). Hamilton again this year outlined how critical the network is – after all Netflix actually built their own CDN. When you think about it – this is the one area of cloud that AWS does not firmly control its own destiny. They will lay fiber and connect their datacenters but will always be dependent on the larger bandwidth providers. Maybe not – but it highlights how critical bandwidth is ultimately for so many of the applications and services – including very basic services like database replication.

Silicon Angle and theCube pulled me into an interview at the last minute – no prep time. But I shared my thoughts and perspectives. Let me know what you think!

 

 

 

 

OpenStack StartUp Ecosystem – It’s Real!

For the upcoming OpenStack conference in Paris, I thought it would be interesting to take a look at how the ecosystem has evolved since OpenStack started. Storm has been fortunate to be involved early and found some success with SwiftStack and MetaCloud – which was recently sold to Cisco (Blog post here). As most know, I think OpenStack is a big part of the future of infrastructure and have numerous blog posts on the subject. I have invested with some conviction for the past 3 years but I think we are still in the early days – the large companies have really just got on board in the last 12 months with the exception of maybe RedHat. The future is very bright for startups – and the $100 million financing for Mirantis or the new $16 million financing at SwiftStack are just two of the latest data points (which incidentally didn’t make it into the graphic before I took it to print). With some help at Storm (thanks Jordan), we looked companies which are either platinum, gold or corporate sponsors of the OpenStack Foundation. There were 63 that made the list as startups. There is a larger group of supporting companies that would have made the numbers even larger. You can view all the companies here.

Some other interesting statistics: OpenStack isn’t just for newly minted startups – Pactera (1995) and Parallels (1999) were both started long before OpenStack – or the concept of cloud infrastructure for that matter.

The newest company was Stackstorm (2014). I am about to invest in a new company as well – but this is really a moving target. Stackstorm is a great example of a new generation for sure.

More than 180 investors in total have made investments in OpenStack companies. All the data was collected via Crunchbase, OpenStack.org and company websites. Please let me know if you have any comments or thoughts on how to improve – we’ll try to put something together again next year. I think its a trend worth watching.

StormVentures_OpenStack_Infographic_ForwardPushMedia

A MetaCloud Journey

Carnival_CCCC9214-M

Congratulations to the team at MetaCloud. The acquisition by Cisco was announced today. I am obviously very proud to have been involved with the company. It’s something special to have seen the team make it happen since the very first day. It was an incredible journey with the team – as anyone would expect. There is something to be learned from failure, but success at times comes with even more lessons to be learned. And successes certainly are more fun to reflect on and write about for a post.

As a venture investor, I try to constantly learn, improve and do better as a partner. Ultimately, my job is to generate returns for our investors – the endowments, pensions and others that believe in our collective ability at Storm to invest in early stage enterprise companies. My ultimate goal is to make the companies I invest in successful. If I help make our companies successful, good things should and will happen with time. So from that perspective, understanding the how and the what that made MetaCloud a success as a company is an interesting story. Looking back many of the elements that made MetaCloud so successful are universal – like the quality of the team. Sure there are exceptions, but all of these elements make up what every successful company does (it’s certainly my experience) – but each one in a slightly different way. That is what is interesting.

I also want to give a nod to Chris Hipp. I was introduced to one of the founders of MetaCloud, Steve Curry, by our mutual friend Chris. Chris passed away in 2009 unexpectedly – Ashlee Vance (who was at The NYT at the time now Businessweek) did a nice obituary. I wish Chris was still around to see all that was created. It would have made him very happy. Thank you Chris.

 

 

So what attributes have contributed to make MetaCloud a success?

 

  1. Not being wed to the original idea

Sean Lynch and Steve Curry had the original idea of acting like a sort of cloud broker of compute resource. Sean had experienced first hand the excess compute capacity that often sits idle behind large corporate firewalls. If there was a way to enable others to leverage that capacity – and enable a large enterprise with capacity to sell that capacity back to the “grid”, amazing things could happen. Steve Curry and I met many times at the Dutch Goose in Menlo Park to brainstorm about the idea before funding – even pulling in an occasional customer into the discussion. We spent a lot of time going over the idea, talking with potential customers but it quickly became clear to us that the idea of a broker model was just too early for the market. Mostly because if you think about it, the prerequisite for being able to enable this compute brokering is for enterprises to already have effectively what we would call today a private cloud. Those private clouds did not exist then and we are just starting to see them get deployed today. The broker model has merit and likely will exist when the market is ready. You heard it hear first

 

 

  1. Build what customers want

The team was very smart to recognize the broker model was flawed early and adjust the plan – otherwise we would have failed.  I am not sure I would call what happened next a pivot so much as carefully listening to customers. Customers wanted a private cloud. The team built what customers wanted – it sounds obvious but building the broker concept technically was interesting and at the time seemed more disruptive than just standing up and managing private clouds for customers. But customers wanted private clouds and needed help designing and operating them. The team decided to deliver OpenStack as a service. Deliver a fully functional OpenStack cloud behind a corporate firewall where the cloud would be fully managed so that there would be no need to have any internal OpenStack expertise. One thing we learned early – and this was critical – was that we needed to make a promise to the customer, if at any time they want to take over managing the cloud themselves, they could do so without cost. MetaCloud would give them a license to everything. Everything. No cost. Zero. This put the proverbial monkey on our back to deliver a solution that never made that choice relevant. It focused the development efforts on managing scale, deployment and operations as well as contributing code back to the OpenStack community. MetaCloud had little incentive to deliver anything proprietary since customers could in theory take it at any time. Poof.  We struggled to explain what this meant to the market and to customers. Is it a consulting business? – no. Is it a distribution I can buy support for?- no. What is it?  The best answer:  it’s a solution.

 

 

  1. Delivering a solution not a product

OpenStack as a service. We were able to get a lot of traction in an early market by delivering the right solution versus the right product. Customers ultimately want to get their problems solved so they can focus on their own business. Sometimes products are good enough to deliver against that value proposition or customers may be well equipped to buy just products (think copy machine). The MetaCloud team realized early on that many customers needed a solution around OpenStack. Customers either didn’t have the resources to execute an implementation themselves or didn’t view placing internal resources against a problem as the best use of their own teams. MetaCloud enabled those teams to do even more – focus on the harder problems and scale way beyond what they were able to do on their own. MetaCloud in many ways became part of their customer’s team and we treated the customers as if they were part of our team. It paid off.

 

 

  1. Execute. Execute. Execute

The team could execute. If they got the opportunity to sit down with a technical leader, it usually led to a serious sales discussion. I cannot think of a single instance where MetaCloud was the contributor to the issue when standing up a cloud with a customer. We struggled sometimes to get hardware delivered on time or projects got pushed out based on a customer’s new timeline but the team executed. Sean Lynch hired an amazing team of people to help him execute. Customers asked us to get OpenStack running with their NetApp filers? – no problem. Want it to run with EMC arrays? – we can do that. And this was long before EMC, HP and others committed fully to OpenStack. We did a little custom work for sure but we could do it because this team could execute with confidence.

 

 

  1. Team had technical credibility with customers

MetaCloud had a tough problem to solve in most sales situations. Infrastructure is not like a sales tool that someone can just try out, roll it out to the team and if it works great, if not it was just some wasted spends. For MetaCloud’s customers, they were making a huge bet not only on OpenStack but also on MetaCloud’s ability to deliver. The team had some technical credibility with some of the early customers by virtue of having worked with some of those team members side by side in other companies but existing relationships doesn’t scale. Technical credibility was critical to success (and I probably underestimated it at the beginning) simply because it was such a major infrastructure decision. I am biased, but I would guess there is not a better technical team on the planet.

 

 

  1. Financial prudence and sequencing

When we launched OpenStack as a service, we didn’t have customers banging down the door. The team wasn’t well known outside of the companies that they had worked at before and OpenStack was viewed as an immature project with little hope of success. We had to figure out what OpenStack as a service really meant for customers – how much were they willing to pay? what was the right way to charge? – per core, cpu, server or some other metric? OpenStack was early and as a startup we were not going to move the project forward by ourselves – we knew we couldn’t move much faster than the community. As a result, the company was very careful with our spend until first we really understood what it was we were selling. Then with the first sales hire, Bert, we were very careful to figure out how we were going to sell repeatedly before investing more in terms of sales and not until recently have we really started to ramp up the marketing team. The team was also thoughtful about raising equity.They operated on a shoestring for two years from the start when we wrote the seed check in summer 2011 to the A round which was done in June 2013 with Canaan and Maha Ibrahim. She is a great partner at a great firm. The team then raised a Series B financing last spring and it made sense since we knew what to scale and how to scale.

 

 

I am grateful to have worked with such an outstanding team and wish them well in their next chapter of success with Cisco.

 

 

SaaS Metrics

There are others that have spent much more time thinking about metrics in SaaS. I am enjoying watching the community develop better and better ways about thinking about these new businesses. They make up the majority of the models that we are investing in at Storm today.

Dave Kellog, who is CEO at SaaS company Host Analytics, put together a great post titled “The Ultimate SaaS Metric: The Customer Lifetime Value to Customer Acquisition Cost Ratio (LTV/CAC)

I like the fact that Dave is a fan of actually looking at the financial model as well – I think this is critical because any one metric or combination of metrics can be misleading (positive or negative) especially early in a company’s life – but even can cause confusion for public companies. As an entrepreneur (or investor), understanding the model is critical. Its guaranteed to be wrong but it gives you a framework. He references a SaaS metric periodic table from Insight Venture Partners which I thought was creative and generally accurate.

Unlike Dave, I don’t think all four financial statements matter – at least on a regular basis. A statement of retained earnings really doesn’t tell you much about running the business – but cash flow, income and balance sheets do.

He makes the point that one of the best ratios to looks at is lifetime value / CAC. I agree – though with a caveat. The biggest problem with it early on with a business is LTV and CAC are dubious calculations because the datasets are small – and the ratio therefore is just as potentially flawed. And Dave recognizes some of the limitations as well like calculating churn which has another great post on here. But the basic point is right on – a business should value what is willing to pay for a customer (CAC) based in some way to what that customer is worth (LTV).

The CAC ratio captures the cost of acquiring customers. In plain English, the CAC ratio is the multiple you are willing to pay for $1 for annual recurring revenue (ARR). With a CAC ratio of 1.5, you are paying $1.50 for a $1 of ARR, implying an 18 month payback period on a revenue basis and 18-months divided by subscription-GM on a gross margin basis.

Lifetime value (LTV) attempts to calculate what a customer is worth and is typically calculated using gross margin (the profit from a customer after paying the cost of operating the service) as opposed to simply revenue. LTV is calculated first by inverting the annual churn rate (to get the average customer lifetime in years) and then multiplying by subscription-GM.

For example, with a churn rate is 10%, subscription GM of 75%, and a CAC ratio of 1.5, the LTV/CAC ratio is (1/10%) * 0.75 / 1.5 = 5.0.

The general rule of thumb is that LTV/CAC should be 3.0 or higher, with of course, the higher the better.

Happy modeling!

Surviving Technology Extinction: Mobile, Cloud and SaaS Are the IT Asteroids of our Day

Impact_eventJeff Haynie who is the CEO and co-founder of Appcelerator recently put together a great post with Re/code called Surviving Technology Extinction. It is a great read. Like so many things in life, one can view challenges as obstacles or opportunities. The changing IT landscape is no different. I have made the case before that we are in the middle of a complete rebuild of the IT stack from infrastructure to applications. Its the reason I have written other posts such as my thoughts on Openstack and applications like GuideSpark.

Jeff references a good article from the NYT about how consumers should deal with the complexity of IT choices today and many of the lessons can be applied to the enterprise. I am so impressed with many of the senior IT executives I meet today because the embrace the idea of empowering users and are leading their companies into the next generation of IT services. I completely agree with the strategy to avoid extinction – although it is easier to state in the abstract without any constraints. Regardless, its a great framework and place to start.

Embrace the user

CIOs must recognize the huge productivity gains of providing employees with apps that truly help them perform their jobs better — those that are simple, complete and conform to their lifestyle versus the other way around.

Favor ecosystems over features

The best modern solutions find themselves at the center of a broad, diverse ecosystem as well — composed of groups like developers, as well as software and services companies

Expect openness

Vendors for whom openness is a part of the DNA offer customers more than just lowered risk of lock-in. There is also the ubiquity of skilled resources, as well as the likelihood of a broader ecosystem. Openness and ecosystems go hand in hand.

When it comes to the enterprise, it’s not about just surviving “technology extinction.” There’s a very real opportunity to thrive in these times of uncertainty, simply by embracing the opportunity while competitors dither.

I couldn’t agree more. This is such a powerful force in IT right now. I am counting on some large incumbent vendor’s products going extinct. This creates enormous opportunity for start ups.

Just like the changes that took place on Earth  65 million years ago and resulted in the extinction of the dinosaurs, those same changes also gave rise to an environment where mammals could thrive. Mobile, cloud and SaaS are the IT asteroids of our day.

 

 

 

The End of the Beginning – State of OpenStack

CarnivalIceSculptureOpenstack2014With the OpenStack Summit in Atlanta approaching next week – it has me thinking about the state of OpenStack and cloud infrastructure. I have been a believer for a long time.  I have made only two related investments in the OpenStack ecosystem – MetaCloud and SwiftStack – but these two investments have given me a field level view that OpenStack is indeed growing its installed base. There are certainly other indicators like the billion dollar pledge by HP with Helion. Infrastructure and architectures change slowly but its happening. We are at the end of the beginning of the new rebuild of the data center infrastructure stack.

VMware seems incredibly well positioned especially with its vCloud hybrid solution. If you haven’t checked it out – you should. This product could really knock the cover off the ball – at least with the installed base. I have to imagine new licenses will slow for VMware over time but their existing installed base is enormous and they are the undisputed leader. Microsoft and Citrix never really offered a worthy challenge.

However, what I am seeing is that many new production, test and development deployments are happening on OpenStack. While its harder to justify moving existing infrastructure that is running in production without problems onto OpenStack – even if there is the potential for savings – new deployments, new initiatives, expansion are a different stories.

And its not just compute virtualization where the battle is raging – storage is changing dramatically as well. Inktank, the commercial entity behind Ceph was acquired by Redhat which was a huge win for SDS (software defined storage). EMC made it very clear this week that they view AWS and public cloud storage as a major threat introducing Elastic cloud and VIPR. They are working incredibly hard to align their storage strategies with the gravitational pull of the software defined datacenter. Its impressive and worth looking at I think.

AWS marches on with 1000+ enterprise focused sales people and a relentless focus on execution. Amazon is truly impressive. Really impressive. However – I continue to believe that the economics of AWS work well for the long tail (which might include big name customers spending small dollars) but at scale  – maybe 1m+ spend a month – private cloud economics dominate. Not to mention all the other benefits. Many believe my thinking is flawed – I haven’t accounted for all the hidden costs of private cloud, the capex is hard to procure etc. I have a lot of data points now suggesting that there is at least something to my theory.

Maybe they are right but I can tell you that MetaCloud and SwiftStack are clearly showing there is a value proposition and enterprises are buying OpenStack and Swift. One of the great things about the enterprise is that markets are rarely “winner take all”  like the wed (Google, Facebook) – in fact, its hard for me to name a company other than Microsoft that has really singularly dominated in IT. Maybe IBM in the 80’s.

I can’t wait to see friends next week in Atlanta at the OpenStack summit. This will be my 3rd year going to the summit and speaking. At my first summit, I didn’t see another venture investor but I suspect there will be many there this year. I hear the summit is expecting more than 4,000 people. AWS had something like 8,000 attendees. VMworld had maybe 23,000. OpenStack is real.

But OpenStack is far from perfect but its working, better than ever, and its the best alternative to a public turnkey experience with AWS or costly lock in with VMware. Its going to be a huge success over time – nobody moves production applications without careful consideration so it will take time. As I look back over the last 3 years, the arc of private cloud is clearly pointing to OpenStack. I am hoping to invest more in the OpenStack ecosystem.

(Thanks @stu @wikibon for helping with the conference numbers)

 

Double down on what is working as a Startup

When business starts to accelerate in a startup, before you really hit any scale, its easy to fall into the trap of focusing on new strategies for growth. The new shiny object. We have to invest in X or Y. Sometimes the focus can be new market segments (large enterprise customers or a new vertical), sometimes the focus can be sales strategy which my partner covered in this post on GuideSpark and what they were able to do with outbound sales. None of these decisions are easy – except the one to double down on what is working.

While I guess its possible that as a startup you have exhausted the opportunity, its far more likely that you have just begun to find your groove. Most executive teams and investors like to focus on growth – its the engine that drives companies to success and its natural that the discussion of growth should take into account new things to focus on to feed the engine. Every company is resource constrained or at least should be – I think without some constraints, it is hard to make the best decisions. Assuming some constraints, given you have one more dollar to spend – where would you spend it as a startup?  Spend it first on what is working and what is driving growth if at all possible. It is one of the easiest decisions to make. You may hit the limit, but the risk is very low relative to other choices (again assuming growth in the business). When you have more to invest and revenue is really scaling, then start to think about experiments and other ideas.  The companies that I have been fortunate enough to work with have always found the most success building on that foundation and doubling down on what is working.

Having to sell a business is never easy – One specific example

If you are an entrepreneur or an early stage investor of any kind, it’s very likely that you have failed at some point in your career. I have. I have made a lot of mistakes. If you haven’t, it’s likely that you will or you haven’t pushed hard enough. I think any general post about when to sell in business probably lacks any specific detail that would be helpful for any entrepreneurs and probably wouldn’t give much insight into how I think about the subject. I am not talking about the decision to sell because someone offers you a good (or great) offer for the business but rather the situation where it’s the option of last resort. No one is making money and realizing their vision for the company. Lots of people have different opinions on the matter. I have found that the obvious flags for shutting down a business, while incredibly disappointing for everyone, are not nearly as agonizing as times when it’s not so obvious a decision.  People like to retell stories and make it sound obvious – but that is often just revisionist history. It makes for a better (and shorter) story.

I was involved with a company in the past where I wrote the first check to seed fund the entrepreneur with an idea that we all thought had merit. We made some great progress over the following months including some good initial customer interest that led to a series A with a new co-investor who also had a lot of domain experience. We also had an outside board member who I had and still have a tremendous amount of respect for (who was originally brought in by the founder) and had a lot of experience in similar markets and businesses. In other words, we had the right ingredients for success. So what happened?

Over the next 12 months, we realized that the product/service that we had built really didn’t hunt with customers. The initial “interest” we had poorly interpreted and we missed key functionality in the design that was required for success. We spent the next 12 months trying to “fix” things. The founder for good personal reasons had physically moved and was commuting – working just as hard (or harder) but it compounded the problem. By the time we got the founder some help to refocus the business – and we got some great people to help working full time – it was too late. This is all hindsight – at the time it seemed like a good strategy. We spoke with new potential investors as the cash began to run short and we just didn’t have enough new customer traction to get anyone new to invest. Maybe we just didn’t give the new team enough time to figure things out. At the time, I believed that we could re-trench and rebuild and initially was willing to continue to fund the business regardless of any new investor – concluding that while we missed the target initially we could move forward and build success. I was motivated by the opportunity. In addition to believing in the opportunity, emotion was a big factor for me as I had been involved with the business since the very beginning. I was a believer which sometimes gives a rose colored tint to any situation.  I believed in the new executives and their ability to change course.

But in the end we sold the business. I agonized about it. We didn’t write another check – no one did. The company was sold for essentially just assets. All employees found a new great place to work and a future. My co-investor deserves credit as well – which is a great argument for having multiple investors – in that they were very clear on shutting the business down. There was no disagreement with management either on the path forward. We had not delivered against the numbers, we had tried to fix things without enough tangible evidence. It was time to sell. This is what happens to many startups. It’s a bitter pill. No one likes it to end this way. But that risk is required in order to achieve the rewards that venture and startups offer.

I guess we will never know if it was the right decision ultimately – but it was the right decision at the time. For everyone. We as venture investors contribute money but anyone working at a startup is giving something just as precious which is their time. I would make the same decision today. I probably will remember more about it it some ways than the successes like Sandforce – a story for another post. It just didn’t make sense to invest more money and time. For me, these are hard decisions investors and entrepreneurs have to make together with less than perfect information. It made sense for employees and founders and for investors. The not so obvious flags looking back were lack of product market fit clarity after working at it for some time against a backdrop of a substantial future commitment in terms of time and investment, a collective effort that had repeatedly missed the mark. It was the right decision. I like to think I don’t give up often – but being thoughtful about when to sell I think is part of being an early stage investor and entrepreneur.

GuideSpark Investment

I recently got some excellent feedback from an entrepreneur that I have worked with in the past and he told me that many of the venture blog posts are full of buzz words and high level strategy and one of the things he was interested in learning more about is how and why I make the investments that I do. Great subject. The company announced the funding round today.  For GuideSpark, the decision was simply (1) revenue momentum and (2) an outstanding team.

GuideSpark is a digital video employee communications company with a SaaS solution that transforms how companies engage and inform their workforce about key topics such as healthcare reform and company benefits. The company was founded in 2008 and today has about 50 employees. I met the team on Valentine’s Day and we closed the investment on April 23rd. Several other venture firms passed. Though I know it felt like an eternity for the team, it was a relatively quick first meeting to close in just over 60 days.

The team at GuideSpark is impressive. I obviously got to know the executive team the best but have had the chance to meet just about everyone and it’s just an impressive group all around. These are people I want to work with and would work for if I had some skill that was still useful in an operating capacity. Just one example, the CEO and founder Keith Kitani is an impressive individual. He struggled with the business for several years before finally figuring out the product market fit that really worked to accelerate the business. Success so far hasn’t come easy for him and his other co-founders but they have persevered and iterated on the product making progress (in terms of revenue and customers) with each turn of the offering. The two other founders – John Wolff (Channels) and Joe Larocque (Customer Success) also deserve a lot of credit for supporting the company through some challenging times and also welcoming an organization around them in more senior positions. They are focused on GuideSpark being successful. I like a team that thinks that way. Many say it, but far fewer are able to act on it.

One of the things that has impressed me the most so far is Keith’s ability to hire a great team. Only really in the last 12 months has the business started to grow fast enough that he could rationalize other executive hires – and convince and recruit them to join. The executive team includes now Bob Benedict (CTO), Shep Maher (VP of Sales), Jennifer Turcotte (VP of Marketing), and Stephanie Copeland Weber (VP of Customer Success). It’s unusual for me to invest in a company with such a high caliber team at the start because usually the companies are earlier in their development. Just after I invested, we recruited Chris Krook as the VP of Finance who I think will also make a huge impact on the company’s success.

The second reason was customer and revenue momentum. The company has more than 100 customers today including names like 7-Eleven, Adobe, SurveyMonkey, DPR Construction, Leap Frog Enterprises and Infoblox. The number of customers has more than doubled since my first meeting in February. MRR has grown almost as fast – just under 100% growth over the same period of time – roughly 6 months.  The numbers are still small but I saw clear momentum with customers and revenue. I am not a benefits expert. I am still educating myself about the market. The sales team at GuideSpark has a good time poking fun at me because I haven’t been able to close a customer for them (yet). I just don’t know F500 HR departments well – which is the general buyer. It’s a long way of saying – while I don’t know the specific HR market as well as I know other markets, I know SaaS and I believe the customer growth was a good proxy for everything else.