Last week I wrote about Rackspace’s pivot from a pure play IaaS vendor to a managed service provider. In the article I questioned how the smaller, niche IaaS providers could compete with the “big three” IaaS companies and whether playing the price war game was a losing choice. I spent this week researching the answer to that question and here is what I found out.
Thoughts on Rackspace
I interviewed top executives from CloudSigma, DigitalOcean, and ProfitBricks and asked them about their thoughts on Rackspace’s announcement. Not one of the three companies was surprised with Rackspace’s move to managed services. In fact, Rackspace started out as an MSP before going head to head with the large IaaS providers and this area has always been their strength. One executive speculated that Rackspace and the OpenStack community’s inability to innovate fast enough to keep up with the big three played a major role in the decision. Robert Jenkins, CEO at CloudSigma, had this to say:
Rackspace comes from a traditional dedicated hosting environment and especially around managed services. This is quite a different business model from public cloud. It has been clear for the last year or so that Rackspace has been struggling to compete and innovate on its product whilst at the same time its prices have stayed high compared to the competition. This reveals two underlying problems which give insight into Rackspace’s decision to withdraw from the public cloud space. Firstly, there is a great deal of innovation in both IaaS and PaaS and this is largely a software development challenge. Rackspace does not have pedigree in this field from its legacy business and whilst it tried to mitigate this with the open sourcing of its stack with OpenStack, it wasn’t enough to address the underlying challenge in their case. Secondly, running IaaS needs a very different organization set-up from a managed services type business model. For that reason I believe other companies such as HP will likewise also struggle to make a dent in the public cloud space with their internal attempts at offering public cloud.
If you look at the sector, you will see the companies that are successfully competing against Amazon, Google and Microsoft are for the vast majority, pure play IaaS providers who focus on innovation and have a company set up from the beginning that is tuned to offering public cloud IaaS services.
In order for Rackspace to compete against the big three, they must do more than simply copy industry leaders. Instead they must differentiate themselves and offer services that you can’t get from the market leaders. For Rackspace, that means going back to their core competency which is managed services.
How do they compete on price?
The next question I asked these companies was how are they able to compete with the big three in the battle to be the low cost provider. Ben Uretsky, CEO of DigitalOcean, stated that his company does not participate in the price wars. In fact, they have not changed their price once in their company’s history. Instead, they took a bottom up approach when building their solution and priced their services for the future. He says the smaller IaaS providers do not have all of the overhead within their organization like the big three do. The smaller providers do one thing and one thing only; build public cloud infrastructure as a service. This focus allows them to minimize operational costs to keep prices lower. In contrast, the big three are not singularly focused and have business units that compete for resources, driving costs up. For example, Amazon is investing in drones, Kindle technology, smartphones, television, etc. Google is investing in self-driving cars, Google Glass, Internet of Things technologies and much more.
Andreas Gauger, CMO and Co-founder of ProfitBricks, declares that there is a fake cloud price war going on. He believes that the big three have huge margins and by no means are they the low cost providers.
ProfitBricks Cloud Computing IaaS platform was designed from the beginning to solve the problems of the first generation of Clouds, many of which the new entrants (Google and Microsoft) also suffer from. These problems, slow and inefficient networks, poor resource provisioning and subsequent slow performance. We also designed the architecture with very modern but standard technologies and optimized big parts of the infrastructure stack heavily to achieve cost effectiveness. The whole ProfitBricks infrastructure has efficiency in its DNA. ProfitBricks dropped it’s pricing by 1/2 in the Summer of 2013 – we were the first to stand up and say that the “Fake Cloud Price War” should come to an end, and IaaS providers should be solving their “Mountain of Margin” problem. It was only in March that AWS and Google started to be honest about the high gross margins – which were about 90% and today are still in the 80% range. Today, ProfitBricks continues to be the price/performance leader and we have designed an IaaS platform that is very efficient and provides enough margin for ProfitBricks to be successful, while also providing the best possible price/performance ratio for our customers.
It is also a myth that bigger size companies like Amazon can achieve higher margins due to the sheer volume of their business. Professional niche cloud providers can negotiate hardware and datacenter space nearly as effective as the big three…..
CloudSigma’s response used a neat analogy of raw materials to explain the pricing.
Unlike other technology businesses, IaaS essentially repackages a raw material, in this case compute resources. In this way we are no different from a gas station. If the price of oil halved every two years you’d expect your gas prices to be going down constantly? Well the situation with IaaS is exactly this. Our raw material is computing hardware and their prices are constantly falling. As such you’d expect significant price reductions over time.
In fact if you track the price reductions of companies like AWS, until very recently they have cut them far less than the underlying cost base reductions. In other words, AWS margins INCREASED over the last few years not decreased. It really is a myth that AWS is ‘cheap’ and this can easily be established by comparing 2008 storage prices for example to their 2013 prices and then comparing hard drive prices for the same two date points.
What Microsoft and Google have done is forced AWS to start cutting prices in line with Moore’s law. Still the prices for compute are by no means low at this point despite recent price reductions. So fundamentally I’d challenge the often repeated premise that there is some sort of price war (which implies loss making prices), that simply isn’t the case today.
In terms of how we specifically compete, we offer a very differentiated approach to offering cloud infrastructure that conveys value to customers on a number of fronts, I’d call this approach a ‘virtual data center approach’. Firstly we offer complete granularity with fully unbundled resources. This means customers can size configure and resize their infrastructure in a totally unique way to themselves which delivers purchasing efficiency in a way you just can’t achieve in other public clouds. Secondly we combine this with a true ‘virtual hardware’ approach which exposes advanced storage, networking, high availability and other type functionality that is partially or completely lacking in mainstay public clouds. For this reason AWS and others have failed to penetrate enterprise computing in a meaningful way. For us, our product delivers to IT departments a powerful toolset which wraps value around the underlying resource.
Finally, you shouldn’t be under any illusions that computing hardware offers massive scale benefits. Average margins on hardware are very low, less than 10% for a hard drive. As such regardless of your size you can imagine that the discounts achieved by an AWS aren’t that significant compared to the prices we achieve at CloudSigma. If computing hardware was high margin then that wouldn’t be the case. Public clouds buy commodity modular hardware, not expensive proprietary hardware set-ups so this is why this is the case.
These smaller IaaS providers are able to offer low prices without needing to adjust them every time the big three announces a reduction. My concern is that I don’t believe price is a key differentiator when it comes to public clouds. As I have written in the past, rich feature sets are what drives adoption. Price has a greater impact when influencing a company’s decision on whether to build their own data centers or outsource to the cloud. When it comes to choosing between public cloud providers, I believe those with the biggest catalog of APIs often get the nod.
How do they differentiate themselves?
Obviously the smaller IaaS providers do not have the resources to build the amount of infrastructure and services that the big three are capable of. To stay relevant in this space they must differentiate themselves. DigitalOcean’s differentiator is simplicity. They claim that building applications on the any of the big three’s clouds have a high degree of complexity. DigitalOcean focuses on the developers and provides an easy to use interface that abstracts away the underlying complexities of the infrastructure, thus allowing developers to get the job done faster. I have personally witnessed a number of startups move to DigitalOcean because of both the simplicity and the low cost. I have not seen any large enterprises evaluate DigitalOcean when comparing cloud vendors. To gain traction within the enterprise, DigitalOcean hopes that developers will become their champions because of the user friendly tools and services they provide. This is similar to how Microsoft has always focused on a building a rich IDE for developers so that developers insist to their management they invest in Microsoft development tools.